included topics:
Player Assist Tendencies
Adjusting Assists
82games: Potential Assists
Value of Assists
Immediate shots created, aka "The Steve Nash measure
Are assists on 3-pt FG worth more than assists for 2pt FG ?
Is it time to add second NBA assist?
Assists in PER - strange decisionmaking by Hollinger?
Assist attempts
Bird vs Dirk
Individual Efficiency Ratings (BoP) from PBP
Assisted % - Creating Your Own Shot?
passing statistics
Michael Redd - all-time leader in turnover rate
Author Message
MattB
Joined: 22 Jun 2006
Posts: 38
Location: Lowell
PostPosted: Mon Dec 06, 2010 8:40 pm Post subject: Player Assist Tendencies Reply with quote
I recently put together a visualization for who is assisting who over on my blog. It got me thinking about considering the percentage of available assists from player a to player b. This looks at whether player b is on the court when player a gets an assist so:
Asst% = (# Assts From A to B) / (# Assts for Player A while B is on the floor)
Interesting to see who players tend look for when on the floor together.
I published a spreadsheet with the results. Feel free to take a look. It's the 09/10 data sorted by greatest to least percentage. I also set a threshhold of 10 assts from player A to B.
https://spreadsheets.google.com/pub?key ... utput=html
The Interactive visualization is here http://hoopism.com/?p=434
Thought some of the pairings were interesting. Hopefully I have these right. I am going to look at a better way of making sense of this information. Not sure what we're going to do with it yet. Wanted to put it out there if anyone had ideas.
Back to top
View user's profile Send private message Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 608
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Dec 07, 2010 8:52 am Post subject: Reply with quote
Another percentage that could be intriguing would be:
ASTD% = (FGM for Player A while B is on the floor)/(Assists from B to A)
So, perhaps Amare was at 50% of his made shots were assisted by Nash when both were on the floor together.
_________________
GodismyJudgeOK.com/DStats
Back to top
View user's profile Send private message Send e-mail Visit poster's website
MattB
Joined: 22 Jun 2006
Posts: 38
Location: Lowell
PostPosted: Tue Dec 07, 2010 10:41 am Post subject: Reply with quote
DSMok1:
That would be interesting. What % of his total shots were assisted by B while sharing the floor. I will work on that.
This is one of the many times I wish I had assisted shots missed. Really needs to be tracked.
Back to top
View user's profile Send private message Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 608
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Dec 07, 2010 10:49 am Post subject: Reply with quote
MattB wrote:
DSMok1:
That would be interesting. What % of his total shots were assisted by B while sharing the floor. I will work on that.
This is one of the many times I wish I had assisted shots missed. Really needs to be tracked.
Absolutely! That would finally enable us to quantify the value of an assist, and what's more, to differentiate the value of each player's assists in terms of FG% added.
_________________
GodismyJudgeOK.com/DStats
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3586
Location: Hendersonville, NC
PostPosted: Tue Dec 07, 2010 10:53 am Post subject: Reply with quote
You are referring to 'potential assists', as they have been called here.
There have been attempts to assess the value of these 'good passes' which may or may not be converted.
However, the (counted) Assist also incorporates the value of the potential assist -- on average. The better the pass, the more likely it's converted into a FG -- and thereby, into an Assist.
If I throw you a pass that is difficult to handle, takes you too far under the basket or into a crowd, I don't especially deserve an assist. If you amazingly convert it anyway, I still get one.
Depending on your teammates, you may get more or fewer than the assists you deserve. Part of your job as a passer is to know who is good at what, where and when they want the ball, etc.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
MattB
Joined: 22 Jun 2006
Posts: 38
Location: Lowell
PostPosted: Tue Dec 07, 2010 11:38 am Post subject: Reply with quote
Mike G wrote:
You are referring to 'potential assists', as they have been called here.
There have been attempts to assess the value of these 'good passes' which may or may not be converted.
However, the (counted) Assist also incorporates the value of the potential assist -- on average. The better the pass, the more likely it's converted into a FG -- and thereby, into an Assist.
If I throw you a pass that is difficult to handle, takes you too far under the basket or into a crowd, I don't especially deserve an assist. If you amazingly convert it anyway, I still get one.
Depending on your teammates, you may get more or fewer than the assists you deserve. Part of your job as a passer is to know who is good at what, where and when they want the ball, etc.
As captured by the large amount of assists Kidd makes while dampier is on the court, and the small percentage directed to Dampier (10%).
Back to top
View user's profile Send private message Visit poster's website
EvanZ
Joined: 22 Nov 2010
Posts: 276
PostPosted: Tue Dec 07, 2010 5:00 pm Post subject: Reply with quote
I think you can also judge the value of an assist by looking at the %AFG of the recipient. If 80% of a guy's shots are assisted, that probably means the assist has a lot more value to him, because he is not as good at creating his own shot. Therefore, your more valuable assists go to players with high %AFG and less valuable assists to players with low %AFG who could create their own shot. A "good" assist would go to a guy with a high %AFG and high TS%. A "bad" assist goes to a guy with low %AFG with low TS%.
(BTW, isn't this pretty much what DeanO said?)
_________________
http://www.thecity2.com
More recovered assist threads
More recovered assist threads
Last edited by Crow on Thu Jun 09, 2011 1:25 am, edited 1 time in total.
Re: More recovered assist threads
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Mar 26, 2010 1:30 pm Post subject: Adjusting Assists Reply with quote
Tom Haberstroh had an excellent article over at Hardwood Paroxysm a few days ago about "Dismantling the Assist". The basic concept was to look at each player's assists and to value them based on how much better an assisted shot from that range was than an unassisted shot.
This struck me as the wrong way to approach it. I believe the better way to attack it is to compare the value of the shot from that location to a generic unassisted shot. For dunks in particular, this makes a huge difference.
The groundwork for all this comes from 82games.com's "The Value of a Good Pass". From there it is possible to see how much value an assist at each location is worth. Unfortunately, these bins don't match the HoopData location tables for assists. I therefore had to do some math to approximate how many assists of each type were represented in the HoopData table. Approximately: 33% of "At Rim" assists were dunks (the rest were "Close Shots"), and 50% of "<10 feet" assists were "Close Shots" (the rest were 2pt Jumpers in 82Games.com parlance).
An unassisted shot has an EFG% of approximately 44% (from what I calculated). To value the assists, then: compare the potential assists' EFG% to the 44% baseline.
The new point value for potential assists to each area:
Dunk: 0.960 pts
Close Shots: 0.346 pts
2 Pt. Jumper: 0.036 pts
3 Pt. Jumper: 0.257 pts
I then summed the points added and normalized so the new, tweaked "assists" still sum to the same total across the NBA. Basically, 1 point added = 2 assists.
So here are the new, adjusted assist leaders (Ast+ is a HoopData metric; my new metric is the "Adj. Assists") (stats per 40 min, NOT pace adjusted):
Code:
# Player Assists Ast+ Total Pts Adj. Assists Change (from Assists)
1 Steve Nash 13.4 15.0 7.0 14.0 0.6
2 Deron Williams 11.2 12.2 5.6 11.1 -0.1
3 Chris Paul 11.4 13.0 5.5 11.0 -0.4
4 Rajon Rondo 10.6 11.8 5.4 10.8 0.2
5 LeBron James 8.8 10.1 5.0 10.0 1.2
6 Baron Davis 9.4 10.2 4.7 9.4 0.0
7 Jameer Nelson 7.4 8.9 4.5 8.9 1.5
8 Russell Westbrook 9.1 10.1 4.4 8.7 -0.4
9 Jason Kidd 10.2 10.9 4.3 8.5 -1.7
10 Jose Calderon 8.6 9.6 4.1 8.2 -0.4
11 Darren Collison 8.3 9.4 4.0 7.9 -0.4
12 Devin Harris 7.7 8.3 3.9 7.8 0.1
13 Chris Duhon 7.4 8.5 3.9 7.8 0.4
14 Raymond Felton 6.5 7.4 3.8 7.5 1.0
15 Brandon Jennings 7.2 8.3 3.7 7.4 0.2
16 Gilbert Arenas 7.8 8.5 3.7 7.3 -0.5
17 Dwyane Wade 7.3 8.4 3.6 7.2 -0.1
18 Earl Watson 6.6 7.8 3.5 7.0 0.4
19 Tony Parker 7.2 8.4 3.5 7.0 -0.2
20 Manu Ginobili 6.9 7.6 3.5 7.0 0.1
21 Mike Conley 6.7 7.3 3.5 6.9 0.2
22 Andre Miller 7.1 8.0 3.4 6.8 -0.3
23 Mo Williams 5.9 6.4 3.4 6.7 0.8
24 Will Bynum 7.2 8.2 3.4 6.7 -0.5
25 Chauncey Billups 6.8 7.6 3.3 6.6 -0.2
26 T.J. Ford 6.1 7.1 3.2 6.4 0.3
27 Jarrett Jack 7.1 7.8 3.2 6.4 -0.7
28 Stephen Curry 6.2 6.9 3.2 6.3 0.1
29 Aaron Brooks 5.7 6.5 3.1 6.1 0.4
30 Larry Hughes 5.3 5.9 3.0 6.0 0.7
EDIT: I just found that Tom had an older post that looked at things more similarly to this (except without using the 44% baseline or using actual ASSISTED FG% for each location). Here that is: http://www.hardwoodparoxysm.com/2010/03 ... w-assists/.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
habetw4
Joined: 12 Nov 2009
Posts: 22
Location: CT
PostPosted: Fri Mar 26, 2010 2:09 pm Post subject: Reply with quote
I don't mean to reinvent the assist statistic because I don't think we have enough granular data, at least publicly, to do the makeover. The two articles were more of an exploration of various ways to record an assist and how that reflects the player's value as a ball distributor. There are many ways to look at the value of the assist. I hope the 82games study and two pieces gets us closer to finding the answer.
The issue essentially boils down to the chicken or the egg problem. Does the passer create the open shots for the offense or does the offense create open shots for the passer? More to the point, should we discount, says, Jason Kidd's assists because his personnel/system environment lends itself to mid range shots, which, in general, are the least efficient shots in the game? Of course not. But then again, we'd like to reward players for opening up opportunities for teammates to get higher percentage shots that they wouldn't have normally received. This is where point expectancy comes in. That, I think, is the ultimate piece to the puzzle.
Like in all player evaluation, we want to assign credit where credit is due. The contextual factor of assists creates a host of misattribution errors that makes player evaluation more difficult. At this point, I think the answer is a blend of different metrics (on/off court, player pairs, type of assist breakdowns, etc) before we are able to delve into point expectancy.
Sometimes, I think assist as a statistic is so flawed it isn't worth the effort and time. But then I feel as though I'm missing an opportunity to find inefficiencies if I ignore them all together.
_________________
I'm a twitterererer: @tomhaberstroh.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Mar 26, 2010 2:27 pm Post subject: Reply with quote
habetw4 wrote:
I don't mean to reinvent the assist statistic because I don't think we have enough granular data, at least publicly, to do the makeover. The two articles were more of an exploration of various ways to record an assist and how that reflects the player's value as a ball distributor. There are many ways to look at the value of the assist. I hope the 82games study and two pieces gets us closer to finding the answer.
The issue essentially boils down to the chicken or the egg problem. Does the passer create the open shots for the offense or does the offense create open shots for the passer? More to the point, should we discount, says, Jason Kidd's assists because his personnel/system environment lends itself to mid range shots, which, in general, are the least efficient shots in the game? Of course not. But then again, we'd like to reward players for opening up opportunities for teammates to get higher percentage shots that they wouldn't have normally received. This is where point expectancy comes in. That, I think, is the ultimate piece to the puzzle.
Like in all player evaluation, we want to assign credit where credit is due. The contextual factor of assists creates a host of misattribution errors that makes player evaluation more difficult. At this point, I think the answer is a blend of different metrics (on/off court, player pairs, type of assist breakdowns, etc) before we are able to delve into point expectancy.
Sometimes, I think assist as a statistic is so flawed it isn't worth the effort and time. But then I feel as though I'm missing an opportunity to find inefficiencies if I ignore them all together.
I see what you were trying to do--more exploration than anything.
I would say that the assist is definitely a valid measure; there is a reason it is quite significant in the SPM regression. It does measure a player's impact in a certain dimension.
As for my version of "assists" and Jason Kidd: it appears that the best valuation would be somewhere in between pure assists (a success metric) vs. my adjusted assist (an efficiency metric). Similar to the difference between yards/carry in football and FirstDowns/carry. One measures "success"; the other "efficiency". True value is a synthesis of the two.
I suspect that using an adjusted version of assists such as this would allow a stronger correlation when doing an SPM-on-APM regression.
A few notable players that increased or decreased on this list:
Code:
Player Assists Adj. Ast. Change (from Assists)
Jameer Nelson 7.4 8.9 1.5
LeBron James 8.8 10.0 1.2
Raymond Felton 6.5 7.5 1.0
Mo Williams 5.9 6.7 0.8
Vince Carter 3.9 4.7 0.8
Larry Hughes 5.3 6.0 0.7
Stephen Jackson 3.7 4.4 0.7
Carmelo Anthony 3.6 4.3 0.7
Kevin Garnett 3.6 4.3 0.7
Steve Nash 13.4 14.0 0.6
Player Assists Adj. Ast. Change (from Assists)
Jason Kidd 10.2 8.5 -1.7
Kirk Hinrich 5.5 4.3 -1.2
Tayshaun Prince 3.4 2.2 -1.2
Derrick Rose 6.4 5.5 -0.9
Rafer Alston 5.4 4.6 -0.8
Rodney Stuckey 5.6 4.9 -0.7
Jarrett Jack 7.1 6.4 -0.7
Mike Miller 4.5 3.8 -0.7
Rafer Alston 4.4 3.8 -0.6
John Salmons 2.9 2.3 -0.6
Interesting!
Back to top
View user's profile Send private message Send e-mail Visit poster's website
erivera7
Joined: 19 Jan 2009
Posts: 178
Location: Chicago, IL
PostPosted: Sat Mar 27, 2010 4:28 pm Post subject: Reply with quote
I like the idea of the "adjusted assist" ... that's certainly one metric, in theory, that could be effectively used to find out who really is a good passer vis-a-vis the shots he creates for his teammates. The assist is a flawed statistic but I don't see why it can't be improved upon so it better captures what people are looking for.
I like what I see from the list provided above, so far.
_________________
@erivera7
I cover the Orlando Magic - Magic Basketball
Back to top
View user's profile Send private message
John Hollinger
Joined: 14 Feb 2005
Posts: 175
PostPosted: Tue Mar 30, 2010 11:09 am Post subject: Reply with quote
I think Tom just hit upon something very important which is why I've restrained myself from doing more in ratings to reflect 'assist quality' -- Jameer Nelson has about a hundred times more opportunity to make a pass leading to a dunk than Jason Kidd, simply because of who he plays with. (ok, slight exaggeration). If Jameer played with Caron Butler and Dirk while Kidd playe dwiht Dwight Howard I suspect the numbers would flip.
An interesting case study for the chicken-egg proposition would be to look at Nash in the two MVP years -- one with Amare, one without. Suspect he was setting up a lot more long Js the second year.
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Tue Mar 30, 2010 11:13 am Post subject: Reply with quote
I think Dean makes a good argument for how to value assist quality in BoP. This is one of the many things I lose sleep over (think impact on usage), but in general I think it's a good way of attacking the problem.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
habetw4
Joined: 12 Nov 2009
Posts: 22
Location: CT
PostPosted: Tue Mar 30, 2010 11:35 am Post subject: Reply with quote
Quote:
I think Dean makes a good argument for how to value assist quality in BoP. This is one of the many things I lose sleep over (think impact on usage), but in general I think it's a good way of attacking the problem.
His discussion in BoP was the inspiration for my assist analysis. IIRC, at its core it is a point expectancy argument. The greater the change in point expectancy, the more credit we assign to the passer. We don't have the luxury of within-play point expectancy so a lot of assumptions would have to be made.
_________________
I'm a twitterererer: @tomhaberstroh.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Mar 30, 2010 11:55 am Post subject: Reply with quote
Good discussion, guys.
I thought my initial evaluation wasn't granular enough, so I went back and calculated an approximate unassisted and assisted FG% for each team at each location range. Estimating the overall unassisted FG% for each team gave me a baseline for each assist to each bin for each team. Now, Jason Kidd gets more credit for midrange assists; Nelson no longer gets overwhelming credit for dishing the ball to Howard.
A bye-product of this approach is that good passers on bad teams get adjusted upwards because of all the potential assists that were missed by their poor-shooting teammates. Devin Harris and Rip Hamilton are the biggest beneficiaries of this adjustment--both of their teams are about 41% estimated unassisted EFG%.
Until we know true unassisted FG% and assisted FG% for each range for each team, though, this will still just be an estimate. Perhaps Kidd's dishes to the midrange make a huge difference in Dirk's FG% for that range... maybe...
The new top 20:
Code:
Player Assists Ast+ Total Pts Adj. Ast. Change (from Assists)
Steve Nash 13.4 15.0 5.3 13.5 0.1
Chris Paul 11.4 13.0 4.4 11.1 -0.3
Deron Williams 11.2 12.2 4.1 10.4 -0.8
LeBron James 8.8 10.1 3.9 10.0 1.2
Rajon Rondo 10.6 11.8 3.8 9.7 -0.9
Baron Davis 9.4 10.2 3.8 9.6 0.2
Jason Kidd 10.2 10.9 3.7 9.4 -0.8
Devin Harris 7.7 8.3 3.5 8.8 1.1
Russell Westbrook 9.1 10.1 3.4 8.6 -0.5
Gilbert Arenas 7.8 8.5 3.4 8.5 0.7
Brandon Jennings 7.2 8.3 3.3 8.3 1.1
Jose Calderon 8.6 9.6 3.3 8.3 -0.3
Darren Collison 8.3 9.4 3.2 8.0 -0.3
Jameer Nelson 7.4 8.9 3.0 7.7 0.3
Dwyane Wade 7.3 8.4 3.0 7.6 0.3
Raymond Felton 6.5 7.4 2.9 7.3 0.8
Andre Miller 7.1 8.0 2.8 7.0 -0.1
Earl Watson 6.6 7.8 2.7 7.0 0.4
Manu Ginobili 6.9 7.6 2.7 6.9 0.0
Mike Conley 6.7 7.3 2.7 6.8 0.1
And the new "biggest movers":
Code:
Player Assists Ast+ Total Pts Adj. Ast. Change (from Assists)
LeBron James 8.8 10.1 3.9 10.0 1.2
Richard Hamilton 5.1 5.5 2.5 6.2 1.1
Brandon Jennings 7.2 8.3 3.3 8.3 1.1
Devin Harris 7.7 8.3 3.5 8.8 1.1
Raymond Felton 6.5 7.4 2.9 7.3 0.8
Carmelo Anthony 3.6 4.0 1.7 4.4 0.8
Andrew Bogut 2.4 2.8 1.3 3.2 0.8
Stephen Jackson 3.7 4.0 1.8 4.5 0.8
Mo Williams 5.9 6.4 2.6 6.7 0.8
Lou Williams 5.3 5.8 2.4 6.1 0.8
Kevin Garnett 3.6 4.1 1.7 4.3 0.7
Gilbert Arenas 7.8 8.5 3.4 8.5 0.7
Andray Blatche 2.5 2.8 1.2 3.2 0.7
Josh Smith 4.9 5.6 2.2 5.5 0.6
Andre Iguodala 5.8 6.3 2.5 6.4 0.6
Player Assists Ast+ Total Pts Adj. Ast. Change (from Assists)
Chauncey Billups 6.8 7.6 2.4 6.2 -0.6
Mehmet Okur 2.4 2.5 0.7 1.8 -0.6
Boris Diaw 4.3 4.9 1.4 3.7 -0.6
Beno Udrih 5.7 6.3 2.0 5.0 -0.7
Derrick Rose 6.4 7.1 2.3 5.7 -0.7
Kirk Hinrich 5.5 5.8 1.9 4.8 -0.7
Tony Parker 7.2 8.4 2.5 6.4 -0.8
Tyreke Evans 6.0 6.9 2.1 5.2 -0.8
Jarrett Jack 7.1 7.8 2.5 6.3 -0.8
Jason Kidd 10.2 10.9 3.7 9.4 -0.8
George Hill 3.8 4.3 1.2 3.0 -0.8
Deron Williams 11.2 12.2 4.1 10.4 -0.8
Rajon Rondo 10.6 11.8 3.8 9.7 -0.9
Paul Pierce 3.7 4.0 1.1 2.8 -0.9
Tayshaun Prince 3.4 3.8 0.9 2.3 -1.1
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Tue Mar 30, 2010 12:37 pm Post subject: Reply with quote
habetw4 wrote:
Quote:
I think Dean makes a good argument for how to value assist quality in BoP. This is one of the many things I lose sleep over (think impact on usage), but in general I think it's a good way of attacking the problem.
His discussion in BoP was the inspiration for my assist analysis. IIRC, at its core it is a point expectancy argument. The greater the change in point expectancy, the more credit we assign to the passer. We don't have the luxury of within-play point expectancy so a lot of assumptions would have to be made.
Yeah I knew you were doing this, just wanted to point that out to JH Cool
Ultimately he's right in that assist totals aren't that useful, even if you do adjust them.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
PostPosted: Thu Apr 01, 2010 6:12 pm Post subject: Reply with quote
John Hollinger wrote:
An interesting case study for the chicken-egg proposition would be to look at Nash in the two MVP years -- one with Amare, one without. Suspect he was setting up a lot more long Js the second year.
um, yes:
05-06:
222 assisted 3's
307 assisted 2pt Jump shots
209 assisted close shots (not dunks)
88 assisted dunks
04-05:
284 assisted 3's
169 assisted 2pt Jump shots
206 assisted close shots (not dunks)
202 assisted dunks
...
Back to top
View user's profile Send private message Visit poster's website
bchaikin
Joined: 27 Jan 2005
Posts: 686
Location: cleveland, ohio
PostPosted: Thu Apr 01, 2010 8:48 pm Post subject: Reply with quote
A bye-product of this approach is that good passers on bad teams get adjusted upwards because of all the potential assists that were missed by their poor-shooting teammates.
fwiw you might want to rethink this line of approach - if you look at say players up and through the age of 30 (just to even out the field as many do not play past that age) over the last 15 years of the nba, a player for example like brevin knight threw for assists at a better ast/min rate than did players like jason kidd, steve nash, andre miller, and a host of others. his ast/min rate was bettered by only chris paul and deron williams...
yet the teams he played on averaged only about 30 wins a season. do you really think adjusting his assists upwards would make him in the minds of some the best "passer" in the game over the last decade and a half?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
BobboFitos
Joined: 21 Feb 2009
Posts: 192
Location: Cambridge, MA
PostPosted: Thu Apr 01, 2010 11:35 pm Post subject: Reply with quote
bchaikin wrote:
A bye-product of this approach is that good passers on bad teams get adjusted upwards because of all the potential assists that were missed by their poor-shooting teammates.
fwiw you might want to rethink this line of approach - if you look at say players up and through the age of 30 (just to even out the field as many do not play past that age) over the last 15 years of the nba, a player for example like brevin knight threw for assists at a better ast/min rate than did players like jason kidd, steve nash, andre miller, and a host of others. his ast/min rate was bettered by only chris paul and deron williams...
yet the teams he played on averaged only about 30 wins a season. do you really think adjusting his assists upwards would make him in the minds of some the best "passer" in the game over the last decade and a half?...
why is it hard to believe brevin knight *could* be one of, if not the best, pure passers in the entire game? brevin was 5'10 (optimistically speaking) without a lot of athletic ability and 0 shooting ability. He had unbelievably quick hands, but even with a great STL%, wasn't ever considered a positive on defense. (And possibly was a detriment) But he had an 11 year career playing over 18,000 minutes. Clearly he was doing *something* well, right?
_________________
-Rob
Back to top
View user's profile Send private message AIM Address
bchaikin
Joined: 27 Jan 2005
Posts: 686
Location: cleveland, ohio
PostPosted: Fri Apr 02, 2010 1:16 am Post subject: Reply with quote
why is it hard to believe brevin knight *could* be one of, if not the best, pure passers in the entire game?... Clearly he was doing *something* well, right?
well then it depends on what your definition of what a "pure passer" is. if its a player that throws a ton of passes and gets alot of assists simply because he throws a ton of passes, then i'd say he could be...
the fact is that in today's game 56%-57% of all FGM have an associated assist. over the past 15 years that percentage goes up to close to 60%. if you assume say just 1/10 of all FGM are clearly on one-on-one plays where a basket is scored and there was no assist (and i'm guessing on that 1/10 number), then close to 2/3 of all of the other FGM had an associated assist - i.e. something like 2/3 of all other FGM preceded by a pass had that pass recorded as an assist...
so the question becomes what proportion of those assists on those other 2/3 of FGM were due more to the passing ability of the passer rather than the scoring ability of the player receiving the pass? or...
The issue essentially boils down to the chicken or the egg problem. Does the passer create the open shots for the offense or does the offense create open shots for the passer?
however if your definition of what a "pure passer" is is not simply just someone who gets alot of assists because he throws alot of passes, but someone who either (1) gets a higher percentage of assists per pass that he throws, or (2) increases the FG% of the players he throws passes to moreso that other passers do, then there should probably be a better measure than simply the number of assists thrown for by a player...
...because there are and have been many players who throw for alot of assists simply because they throw a ton of passes, as something like 2/3 of all FGM that were preceded by a pass had that pass recorded as an assist...
again for example looking at the past 15 seasons and all those players that have (a) played at least 3000 total minutes up to and through the age of 30, and (b) have thrown for assists at a rate of (again an arbitrary number here) say 9 ast/40min, you will find not only names like chris paul, deron williams, jason kidd, and steve nash, but also names like brevin knight, robert pack, sergio rodriguez, and jamaal tinsley...
are there people out there that consider players like pack, rodriguez, and tinsley to have been some of the best "passers" over the last decade and a half? or just as players who threw for a ton of assists because they handled the ball so much, shot infrequently, and threw a ton of passes?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Apr 02, 2010 7:59 am Post subject: Reply with quote
bchaikin wrote:
are there people out there that consider players like pack, rodriguez, and tinsley to have been some of the best "passers" over the last decade and a half? or just as players who threw for a ton of assists because they handled the ball so much, shot infrequently, and threw a ton of passes?...
Because I am looking at where the assist went to, I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. I don't care if they didn't shoot much--if their assists were going for dunks and layups, they should get some credit for it!
Interestingly, the players that get the biggest boost percentage-wise are the post players. Last year, Yao and KG both got a 30% boost because of where their assists went to. The ones hurt the most are players who just made passes to the midrange--Stuckey, Chalmers, Prince, Lee.
Oddly enough, despite the ones helped often being post players, Andrew Bynum is one hurt the most by the adjustment.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
habetw4
Joined: 12 Nov 2009
Posts: 22
Location: CT
PostPosted: Fri Apr 02, 2010 9:04 am Post subject: Reply with quote
Are we to assume that there aren't good passers who throw a ton of passes? Are the two mutually exclusive? Surely, there are players who specialize in ball distribution just like there are those who exclusively specialize in rebounding, guarding the opponent's superstar, or defending the rim.
_________________
I'm a twitterererer: @tomhaberstroh.
bchaikin
Joined: 27 Jan 2005
Posts: 681
Location: cleveland, ohio
PostPosted: Fri Apr 02, 2010 1:48 pm Post subject: Reply with quote
Because I am looking at where the assist went to, I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. I don't care if they didn't shoot much--if their assists were going for dunks and layups, they should get some credit for it!
who should get some credit for it? the passer or the player who made the cut to the basket or got into position close enough to the basket to get the dunk in the first place?...
sine 04-05 both amare stoudemire and zach randolph have both been 20/10 players (or close to it). stoudemire has shot 57% on 2s, randolph 47%. stoudemire has played with primarily one PG (nash), randolph with at least four (four different teams) if not more. are you of the opinion that if randolph had played all 5 seasons with phoenix, and stoudemire with the 4 teams zach had played with, that those two 2pt FG%s would be reversed? all because nash supposedly throws passes that lead to layups and dunks (with no "assist" credit to the player who got the layup or dunk)?...
because if you watch both players stoudemire cuts to the basket far more often than does randolph. shouldn't he get credit for getting himself open often for a player like nash to get him the pass in the first place?...
was there a better "passer" than john stockton? from 1987-88 to 1996-97, ten years, john stockton passed for about 1000 assists/yr. no one else in the league passed for even 700 ast/yr. the following players played for utah during this time, but also for other teams (karl malone, mark eaton, and thurl bailey played primarily for utah):
tyrone corbin
jeff malone
jeff hornacek
david benoit
mike brown
blue edwards
antoine carr
felton spencer
adam keefe
tom chambers
when these 10 players all played for utah (and stockton was there throwing for his 1000 ast/yr from 8788-9697) as a collective group they shot a 49.9% 2pt FG%. for all the other teams they played with (22 in all) during their careers as a collective group they shot a 48.4% 2pt FG%...
now i would assume that of that 48.4% 2pt FG% with other teams that with some of those PGs they shot a bit worse than 48.4% and with some a bit better, meaning that some of those PGs threw for assists to these players where they shot 49%-50%, similar to stockton...
does that mean those PGs were as good of "passers" as stockton was?...
Are we to assume that there aren't good passers who throw a ton of passes?
again - what if your definition of a "good passer"?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Apr 02, 2010 2:05 pm Post subject: Reply with quote
bchaikin wrote:
who should get some credit for it? the passer or the player who made the cut to the basket or got into position close enough to the basket to get the dunk in the first place?...
sine 04-05 both amare stoudemire and zach randolph have both been 20/10 players (or close to it). stoudemire has shot 57% on 2s, randolph 47%. stoudemire has played with primarily one PG (nash), randolph with at least four (four different teams) if not more. are you of the opinion that if randolph had played all 5 seasons with phoenix, and stoudemire with the 4 teams zach had played with, that those two 2pt FG%s would be reversed? all because nash supposedly throws passes that lead to layups and dunks (with no "assist" credit to the player who got the layup or dunk)?...
because if you watch both players stoudemire cuts to the basket far more often than does randolph. shouldn't he get credit for getting himself open often for a player like nash to get him the pass in the first place?...
was there a better "passer" than john stockton? from 1987-88 to 1996-97, ten years, john stockton passed for about 1000 assists/yr. no one else in the league passed for even 700 ast/yr. the following players played for utah during this time, but also for other teams (karl malone, mark eaton, and thurl bailey played primarily for utah):
tyrone corbin
jeff malone
jeff hornacek
david benoit
mike brown
blue edwards
antoine carr
felton spencer
adam keefe
tom chambers
when these 10 players all played for utah (and stockton was there throwing for his 1000 ast/yr from 8788-9697) as a collective group they shot a 49.9% 2pt FG%. for all the other teams they played with (22 in all) during their careers as a collective group they shot a 48.4% 2pt FG%...
now i would assume that of that 48.4% 2pt FG% with other teams that with some of those PGs they shot a bit worse than 48.4% and with some a bit better, meaning that some of those PGs threw for assists to these players where they shot 49%-50%, similar to stockton...
does that mean those PGs were as good of "passers" as stockton was?...
Are we to assume that there aren't good passers who throw a ton of passes?
again - what if your definition of a "good passer"?...
I have no idea about actual "point" credit. I'm simply trying to see which assist is worth more than the others.
Here is what I am doing.
1. Calculate the UNASSISTED team eFG%. This is critical. THIS is where there is a correction for teammates.
2. Calculate the eFG% for all potential assist passes by the player. (This differs depending on how good the team is from EACH LOCATION on ASSISTED eFG%)
3. Subtract unassisted team eFG% from player assisted eFG% and multiply the difference times total # of opportunities.
4. Total all of the NBA; normalize the value to the total number of assists handed out.
Sure, some of the credit for actually being able to make the pass goes to the recipient of the pass... but they already get credit for the score, anyway.
If a player is playing for a good midrange team (where the eFG% is high on assisted midrange jumpers) the player will get credit. If the assisted eFG% on midrange jumpers is lower than the team unassisted eFG%... there was basically no value added. (Why was the shot taken?)
All of this is predicated on all players improving their opponents shot in a given position on the court by the same amount, which is obviously shaky ground. Perhaps Kidd's pass to Dirk is better than JJ Barea's pass to Dirk in the same location. Perhaps CP3's lob to the rim is better than Darren Collison's. But until we can estimate this (which will be impossible until "missed assists" are kept track of in PbP), this is about the best we can do I think.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3535
Location: Hendersonville, NC
PostPosted: Fri Apr 02, 2010 5:54 pm Post subject: Reply with quote
DSMok1 wrote:
... I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. ..
Isn't the 'problem' resolved already? When you pass to high-% areas, you get more assists.
If you have 4 seconds to get a shot, a 33% shot might be as good as you are gonna get. So you get an assist on 1/3 of those passes.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Mon Apr 05, 2010 7:58 am Post subject: Reply with quote
Mike G wrote:
DSMok1 wrote:
... I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. ..
Isn't the 'problem' resolved already? When you pass to high-% areas, you get more assists.
If you have 4 seconds to get a shot, a 33% shot might be as good as you are gonna get. So you get an assist on 1/3 of those passes.
The "truth" is between my "adjusted assists" and the actual assists. One measures successes, the other tries to quantify value added.
True, that may be the best possibility given the situation, but it is also likely to not the best possibility if the person doing the passing were a bit better player, like a Steve Nash. Wouldn't you consider a player with 5 assists on dunks a better passer than a player with 5 assists on midrange jumpers?
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3535
Location: Hendersonville, NC
PostPosted: Mon Apr 05, 2010 8:19 am Post subject: Reply with quote
DSMok1 wrote:
Mike G wrote:
When you pass to high-% areas, you get more assists.
.
Wouldn't you consider a player with 5 assists on dunks a better passer than a player with 5 assists on midrange jumpers?
Suppose these players are equal in every respect, except that in a span of 20 possessions:
- Player A gets 6 passes into the hands of someone at the rim, resulting in 5 successful shots (dunks).
- Player B gets the ball 12 times to jumpshooters, who convert 5 times.
We don't know what they did with/without the ball on their other possessions. It could be A is the 'risk taker' and B is the 'play it safe' guy.
Maybe the teams play different styles (fastbreak vs halfcourt).
One guy did not feed a teammate for a shot on 8 possessions, and the other guy did not on 14 occassions.
Assists are equal to attempts * effectiveness.
If Player A could get 12 passes to players in dunking range, and they convert 10, then he has twice as many Assists as player B has. Now he's a better passer.[/i]
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Mon Apr 05, 2010 11:22 am Post subject: Reply with quote
Mike G wrote:
Suppose these players are equal in every respect, except that in a span of 20 possessions:
- Player A gets 6 passes into the hands of someone at the rim, resulting in 5 successful shots (dunks).
- Player B gets the ball 12 times to jumpshooters, who convert 5 times.
We don't know what they did with/without the ball on their other possessions. It could be A is the 'risk taker' and B is the 'play it safe' guy.
Maybe the teams play different styles (fastbreak vs halfcourt).
One guy did not feed a teammate for a shot on 8 possessions, and the other guy did not on 14 occassions.
Assists are equal to attempts * effectiveness.
If Player A could get 12 passes to players in dunking range, and they convert 10, then he has twice as many Assists as player B has. Now he's a better passer.[/i]
effectiveness. The baseline effectiveness is NOT 0. This is where assists go wrong. What is the value added by the sequence of pass-shot? One must compare to what the team would do without the passing, the baseline UNASSISTED fg% of the team.
The current valuation of assists is attempts*effectiveness. My valuation is attempts*(effectiveness-unassisted effectiveness), scaled so the total is the same as just plain assists. I think it's a valuable number to be able to look at.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3535
Location: Hendersonville, NC
PostPosted: Wed Apr 07, 2010 11:31 am Post subject: Reply with quote
DSMok1 wrote:
...
The current valuation of assists is attempts*effectiveness. My valuation is attempts*(effectiveness-unassisted effectiveness)..
Quote:
One must compare to what the team would do without the passing, the baseline UNASSISTED fg% of the team.
So, where is a table of teams' Unassisted FG% ?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Wed Apr 07, 2010 1:32 pm Post subject: Reply with quote
Mike G wrote:
So, where is a table of teams' Unassisted FG% ?
That's the rub. If the play-by-play recorded whether there was a potential assist on the play, then it would be easy to account for. As it is, I had to estimate it based on the FG%, AST%, and estimated (from The Value of a Good Pass) FG% added by assist for each portion of the floor.
The estimated unassisted eFG% for each team:
Team UAeFG%
PHO 49.1%
ORL 48.5%
CLE 48.1%
TOR 46.6%
BOS 46.5%
GSW 46.4%
SAS 46.2%
DEN 46.1%
NYK 45.7%
UTH 45.6%
ATL 45.3%
MEM 45.1%
NOR 45.0%
LAL 44.8%
DAL 44.7%
POR 44.6%
MIA 44.3%
SAC 44.2%
HOU 44.0%
OKC 43.7%
PHI 43.6%
IND 43.4%
LAC 43.4%
CHA 43.3%
MIL 43.2%
WAS 43.1%
MIN 42.5%
CHI 42.2%
DET 41.9%
NJN 40.2%
If a team gets all of their dunks off of assists, no dunks are assumed in calculating the UAeFG%. Phoenix is so high because they shoot a ridiculously high percentage from 3, and unusually few of their 3's are assisted (I'm guessing Nash is shooting a lot of those!). Orlando is so high because unusually few of their layups/dunks are assisted and they shoot very well from the midrange (and few of those are assisted).
Do you see what I was trying to do? Obviously, this is incomplete unless we know the true UAeFG% and how much each team's eFG% improves in each area with each assist. It's a start, though.
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Mar 26, 2010 1:30 pm Post subject: Adjusting Assists Reply with quote
Tom Haberstroh had an excellent article over at Hardwood Paroxysm a few days ago about "Dismantling the Assist". The basic concept was to look at each player's assists and to value them based on how much better an assisted shot from that range was than an unassisted shot.
This struck me as the wrong way to approach it. I believe the better way to attack it is to compare the value of the shot from that location to a generic unassisted shot. For dunks in particular, this makes a huge difference.
The groundwork for all this comes from 82games.com's "The Value of a Good Pass". From there it is possible to see how much value an assist at each location is worth. Unfortunately, these bins don't match the HoopData location tables for assists. I therefore had to do some math to approximate how many assists of each type were represented in the HoopData table. Approximately: 33% of "At Rim" assists were dunks (the rest were "Close Shots"), and 50% of "<10 feet" assists were "Close Shots" (the rest were 2pt Jumpers in 82Games.com parlance).
An unassisted shot has an EFG% of approximately 44% (from what I calculated). To value the assists, then: compare the potential assists' EFG% to the 44% baseline.
The new point value for potential assists to each area:
Dunk: 0.960 pts
Close Shots: 0.346 pts
2 Pt. Jumper: 0.036 pts
3 Pt. Jumper: 0.257 pts
I then summed the points added and normalized so the new, tweaked "assists" still sum to the same total across the NBA. Basically, 1 point added = 2 assists.
So here are the new, adjusted assist leaders (Ast+ is a HoopData metric; my new metric is the "Adj. Assists") (stats per 40 min, NOT pace adjusted):
Code:
# Player Assists Ast+ Total Pts Adj. Assists Change (from Assists)
1 Steve Nash 13.4 15.0 7.0 14.0 0.6
2 Deron Williams 11.2 12.2 5.6 11.1 -0.1
3 Chris Paul 11.4 13.0 5.5 11.0 -0.4
4 Rajon Rondo 10.6 11.8 5.4 10.8 0.2
5 LeBron James 8.8 10.1 5.0 10.0 1.2
6 Baron Davis 9.4 10.2 4.7 9.4 0.0
7 Jameer Nelson 7.4 8.9 4.5 8.9 1.5
8 Russell Westbrook 9.1 10.1 4.4 8.7 -0.4
9 Jason Kidd 10.2 10.9 4.3 8.5 -1.7
10 Jose Calderon 8.6 9.6 4.1 8.2 -0.4
11 Darren Collison 8.3 9.4 4.0 7.9 -0.4
12 Devin Harris 7.7 8.3 3.9 7.8 0.1
13 Chris Duhon 7.4 8.5 3.9 7.8 0.4
14 Raymond Felton 6.5 7.4 3.8 7.5 1.0
15 Brandon Jennings 7.2 8.3 3.7 7.4 0.2
16 Gilbert Arenas 7.8 8.5 3.7 7.3 -0.5
17 Dwyane Wade 7.3 8.4 3.6 7.2 -0.1
18 Earl Watson 6.6 7.8 3.5 7.0 0.4
19 Tony Parker 7.2 8.4 3.5 7.0 -0.2
20 Manu Ginobili 6.9 7.6 3.5 7.0 0.1
21 Mike Conley 6.7 7.3 3.5 6.9 0.2
22 Andre Miller 7.1 8.0 3.4 6.8 -0.3
23 Mo Williams 5.9 6.4 3.4 6.7 0.8
24 Will Bynum 7.2 8.2 3.4 6.7 -0.5
25 Chauncey Billups 6.8 7.6 3.3 6.6 -0.2
26 T.J. Ford 6.1 7.1 3.2 6.4 0.3
27 Jarrett Jack 7.1 7.8 3.2 6.4 -0.7
28 Stephen Curry 6.2 6.9 3.2 6.3 0.1
29 Aaron Brooks 5.7 6.5 3.1 6.1 0.4
30 Larry Hughes 5.3 5.9 3.0 6.0 0.7
EDIT: I just found that Tom had an older post that looked at things more similarly to this (except without using the 44% baseline or using actual ASSISTED FG% for each location). Here that is: http://www.hardwoodparoxysm.com/2010/03 ... w-assists/.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
habetw4
Joined: 12 Nov 2009
Posts: 22
Location: CT
PostPosted: Fri Mar 26, 2010 2:09 pm Post subject: Reply with quote
I don't mean to reinvent the assist statistic because I don't think we have enough granular data, at least publicly, to do the makeover. The two articles were more of an exploration of various ways to record an assist and how that reflects the player's value as a ball distributor. There are many ways to look at the value of the assist. I hope the 82games study and two pieces gets us closer to finding the answer.
The issue essentially boils down to the chicken or the egg problem. Does the passer create the open shots for the offense or does the offense create open shots for the passer? More to the point, should we discount, says, Jason Kidd's assists because his personnel/system environment lends itself to mid range shots, which, in general, are the least efficient shots in the game? Of course not. But then again, we'd like to reward players for opening up opportunities for teammates to get higher percentage shots that they wouldn't have normally received. This is where point expectancy comes in. That, I think, is the ultimate piece to the puzzle.
Like in all player evaluation, we want to assign credit where credit is due. The contextual factor of assists creates a host of misattribution errors that makes player evaluation more difficult. At this point, I think the answer is a blend of different metrics (on/off court, player pairs, type of assist breakdowns, etc) before we are able to delve into point expectancy.
Sometimes, I think assist as a statistic is so flawed it isn't worth the effort and time. But then I feel as though I'm missing an opportunity to find inefficiencies if I ignore them all together.
_________________
I'm a twitterererer: @tomhaberstroh.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Mar 26, 2010 2:27 pm Post subject: Reply with quote
habetw4 wrote:
I don't mean to reinvent the assist statistic because I don't think we have enough granular data, at least publicly, to do the makeover. The two articles were more of an exploration of various ways to record an assist and how that reflects the player's value as a ball distributor. There are many ways to look at the value of the assist. I hope the 82games study and two pieces gets us closer to finding the answer.
The issue essentially boils down to the chicken or the egg problem. Does the passer create the open shots for the offense or does the offense create open shots for the passer? More to the point, should we discount, says, Jason Kidd's assists because his personnel/system environment lends itself to mid range shots, which, in general, are the least efficient shots in the game? Of course not. But then again, we'd like to reward players for opening up opportunities for teammates to get higher percentage shots that they wouldn't have normally received. This is where point expectancy comes in. That, I think, is the ultimate piece to the puzzle.
Like in all player evaluation, we want to assign credit where credit is due. The contextual factor of assists creates a host of misattribution errors that makes player evaluation more difficult. At this point, I think the answer is a blend of different metrics (on/off court, player pairs, type of assist breakdowns, etc) before we are able to delve into point expectancy.
Sometimes, I think assist as a statistic is so flawed it isn't worth the effort and time. But then I feel as though I'm missing an opportunity to find inefficiencies if I ignore them all together.
I see what you were trying to do--more exploration than anything.
I would say that the assist is definitely a valid measure; there is a reason it is quite significant in the SPM regression. It does measure a player's impact in a certain dimension.
As for my version of "assists" and Jason Kidd: it appears that the best valuation would be somewhere in between pure assists (a success metric) vs. my adjusted assist (an efficiency metric). Similar to the difference between yards/carry in football and FirstDowns/carry. One measures "success"; the other "efficiency". True value is a synthesis of the two.
I suspect that using an adjusted version of assists such as this would allow a stronger correlation when doing an SPM-on-APM regression.
A few notable players that increased or decreased on this list:
Code:
Player Assists Adj. Ast. Change (from Assists)
Jameer Nelson 7.4 8.9 1.5
LeBron James 8.8 10.0 1.2
Raymond Felton 6.5 7.5 1.0
Mo Williams 5.9 6.7 0.8
Vince Carter 3.9 4.7 0.8
Larry Hughes 5.3 6.0 0.7
Stephen Jackson 3.7 4.4 0.7
Carmelo Anthony 3.6 4.3 0.7
Kevin Garnett 3.6 4.3 0.7
Steve Nash 13.4 14.0 0.6
Player Assists Adj. Ast. Change (from Assists)
Jason Kidd 10.2 8.5 -1.7
Kirk Hinrich 5.5 4.3 -1.2
Tayshaun Prince 3.4 2.2 -1.2
Derrick Rose 6.4 5.5 -0.9
Rafer Alston 5.4 4.6 -0.8
Rodney Stuckey 5.6 4.9 -0.7
Jarrett Jack 7.1 6.4 -0.7
Mike Miller 4.5 3.8 -0.7
Rafer Alston 4.4 3.8 -0.6
John Salmons 2.9 2.3 -0.6
Interesting!
Back to top
View user's profile Send private message Send e-mail Visit poster's website
erivera7
Joined: 19 Jan 2009
Posts: 178
Location: Chicago, IL
PostPosted: Sat Mar 27, 2010 4:28 pm Post subject: Reply with quote
I like the idea of the "adjusted assist" ... that's certainly one metric, in theory, that could be effectively used to find out who really is a good passer vis-a-vis the shots he creates for his teammates. The assist is a flawed statistic but I don't see why it can't be improved upon so it better captures what people are looking for.
I like what I see from the list provided above, so far.
_________________
@erivera7
I cover the Orlando Magic - Magic Basketball
Back to top
View user's profile Send private message
John Hollinger
Joined: 14 Feb 2005
Posts: 175
PostPosted: Tue Mar 30, 2010 11:09 am Post subject: Reply with quote
I think Tom just hit upon something very important which is why I've restrained myself from doing more in ratings to reflect 'assist quality' -- Jameer Nelson has about a hundred times more opportunity to make a pass leading to a dunk than Jason Kidd, simply because of who he plays with. (ok, slight exaggeration). If Jameer played with Caron Butler and Dirk while Kidd playe dwiht Dwight Howard I suspect the numbers would flip.
An interesting case study for the chicken-egg proposition would be to look at Nash in the two MVP years -- one with Amare, one without. Suspect he was setting up a lot more long Js the second year.
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Tue Mar 30, 2010 11:13 am Post subject: Reply with quote
I think Dean makes a good argument for how to value assist quality in BoP. This is one of the many things I lose sleep over (think impact on usage), but in general I think it's a good way of attacking the problem.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
habetw4
Joined: 12 Nov 2009
Posts: 22
Location: CT
PostPosted: Tue Mar 30, 2010 11:35 am Post subject: Reply with quote
Quote:
I think Dean makes a good argument for how to value assist quality in BoP. This is one of the many things I lose sleep over (think impact on usage), but in general I think it's a good way of attacking the problem.
His discussion in BoP was the inspiration for my assist analysis. IIRC, at its core it is a point expectancy argument. The greater the change in point expectancy, the more credit we assign to the passer. We don't have the luxury of within-play point expectancy so a lot of assumptions would have to be made.
_________________
I'm a twitterererer: @tomhaberstroh.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Mar 30, 2010 11:55 am Post subject: Reply with quote
Good discussion, guys.
I thought my initial evaluation wasn't granular enough, so I went back and calculated an approximate unassisted and assisted FG% for each team at each location range. Estimating the overall unassisted FG% for each team gave me a baseline for each assist to each bin for each team. Now, Jason Kidd gets more credit for midrange assists; Nelson no longer gets overwhelming credit for dishing the ball to Howard.
A bye-product of this approach is that good passers on bad teams get adjusted upwards because of all the potential assists that were missed by their poor-shooting teammates. Devin Harris and Rip Hamilton are the biggest beneficiaries of this adjustment--both of their teams are about 41% estimated unassisted EFG%.
Until we know true unassisted FG% and assisted FG% for each range for each team, though, this will still just be an estimate. Perhaps Kidd's dishes to the midrange make a huge difference in Dirk's FG% for that range... maybe...
The new top 20:
Code:
Player Assists Ast+ Total Pts Adj. Ast. Change (from Assists)
Steve Nash 13.4 15.0 5.3 13.5 0.1
Chris Paul 11.4 13.0 4.4 11.1 -0.3
Deron Williams 11.2 12.2 4.1 10.4 -0.8
LeBron James 8.8 10.1 3.9 10.0 1.2
Rajon Rondo 10.6 11.8 3.8 9.7 -0.9
Baron Davis 9.4 10.2 3.8 9.6 0.2
Jason Kidd 10.2 10.9 3.7 9.4 -0.8
Devin Harris 7.7 8.3 3.5 8.8 1.1
Russell Westbrook 9.1 10.1 3.4 8.6 -0.5
Gilbert Arenas 7.8 8.5 3.4 8.5 0.7
Brandon Jennings 7.2 8.3 3.3 8.3 1.1
Jose Calderon 8.6 9.6 3.3 8.3 -0.3
Darren Collison 8.3 9.4 3.2 8.0 -0.3
Jameer Nelson 7.4 8.9 3.0 7.7 0.3
Dwyane Wade 7.3 8.4 3.0 7.6 0.3
Raymond Felton 6.5 7.4 2.9 7.3 0.8
Andre Miller 7.1 8.0 2.8 7.0 -0.1
Earl Watson 6.6 7.8 2.7 7.0 0.4
Manu Ginobili 6.9 7.6 2.7 6.9 0.0
Mike Conley 6.7 7.3 2.7 6.8 0.1
And the new "biggest movers":
Code:
Player Assists Ast+ Total Pts Adj. Ast. Change (from Assists)
LeBron James 8.8 10.1 3.9 10.0 1.2
Richard Hamilton 5.1 5.5 2.5 6.2 1.1
Brandon Jennings 7.2 8.3 3.3 8.3 1.1
Devin Harris 7.7 8.3 3.5 8.8 1.1
Raymond Felton 6.5 7.4 2.9 7.3 0.8
Carmelo Anthony 3.6 4.0 1.7 4.4 0.8
Andrew Bogut 2.4 2.8 1.3 3.2 0.8
Stephen Jackson 3.7 4.0 1.8 4.5 0.8
Mo Williams 5.9 6.4 2.6 6.7 0.8
Lou Williams 5.3 5.8 2.4 6.1 0.8
Kevin Garnett 3.6 4.1 1.7 4.3 0.7
Gilbert Arenas 7.8 8.5 3.4 8.5 0.7
Andray Blatche 2.5 2.8 1.2 3.2 0.7
Josh Smith 4.9 5.6 2.2 5.5 0.6
Andre Iguodala 5.8 6.3 2.5 6.4 0.6
Player Assists Ast+ Total Pts Adj. Ast. Change (from Assists)
Chauncey Billups 6.8 7.6 2.4 6.2 -0.6
Mehmet Okur 2.4 2.5 0.7 1.8 -0.6
Boris Diaw 4.3 4.9 1.4 3.7 -0.6
Beno Udrih 5.7 6.3 2.0 5.0 -0.7
Derrick Rose 6.4 7.1 2.3 5.7 -0.7
Kirk Hinrich 5.5 5.8 1.9 4.8 -0.7
Tony Parker 7.2 8.4 2.5 6.4 -0.8
Tyreke Evans 6.0 6.9 2.1 5.2 -0.8
Jarrett Jack 7.1 7.8 2.5 6.3 -0.8
Jason Kidd 10.2 10.9 3.7 9.4 -0.8
George Hill 3.8 4.3 1.2 3.0 -0.8
Deron Williams 11.2 12.2 4.1 10.4 -0.8
Rajon Rondo 10.6 11.8 3.8 9.7 -0.9
Paul Pierce 3.7 4.0 1.1 2.8 -0.9
Tayshaun Prince 3.4 3.8 0.9 2.3 -1.1
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Tue Mar 30, 2010 12:37 pm Post subject: Reply with quote
habetw4 wrote:
Quote:
I think Dean makes a good argument for how to value assist quality in BoP. This is one of the many things I lose sleep over (think impact on usage), but in general I think it's a good way of attacking the problem.
His discussion in BoP was the inspiration for my assist analysis. IIRC, at its core it is a point expectancy argument. The greater the change in point expectancy, the more credit we assign to the passer. We don't have the luxury of within-play point expectancy so a lot of assumptions would have to be made.
Yeah I knew you were doing this, just wanted to point that out to JH Cool
Ultimately he's right in that assist totals aren't that useful, even if you do adjust them.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
PostPosted: Thu Apr 01, 2010 6:12 pm Post subject: Reply with quote
John Hollinger wrote:
An interesting case study for the chicken-egg proposition would be to look at Nash in the two MVP years -- one with Amare, one without. Suspect he was setting up a lot more long Js the second year.
um, yes:
05-06:
222 assisted 3's
307 assisted 2pt Jump shots
209 assisted close shots (not dunks)
88 assisted dunks
04-05:
284 assisted 3's
169 assisted 2pt Jump shots
206 assisted close shots (not dunks)
202 assisted dunks
...
Back to top
View user's profile Send private message Visit poster's website
bchaikin
Joined: 27 Jan 2005
Posts: 686
Location: cleveland, ohio
PostPosted: Thu Apr 01, 2010 8:48 pm Post subject: Reply with quote
A bye-product of this approach is that good passers on bad teams get adjusted upwards because of all the potential assists that were missed by their poor-shooting teammates.
fwiw you might want to rethink this line of approach - if you look at say players up and through the age of 30 (just to even out the field as many do not play past that age) over the last 15 years of the nba, a player for example like brevin knight threw for assists at a better ast/min rate than did players like jason kidd, steve nash, andre miller, and a host of others. his ast/min rate was bettered by only chris paul and deron williams...
yet the teams he played on averaged only about 30 wins a season. do you really think adjusting his assists upwards would make him in the minds of some the best "passer" in the game over the last decade and a half?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
BobboFitos
Joined: 21 Feb 2009
Posts: 192
Location: Cambridge, MA
PostPosted: Thu Apr 01, 2010 11:35 pm Post subject: Reply with quote
bchaikin wrote:
A bye-product of this approach is that good passers on bad teams get adjusted upwards because of all the potential assists that were missed by their poor-shooting teammates.
fwiw you might want to rethink this line of approach - if you look at say players up and through the age of 30 (just to even out the field as many do not play past that age) over the last 15 years of the nba, a player for example like brevin knight threw for assists at a better ast/min rate than did players like jason kidd, steve nash, andre miller, and a host of others. his ast/min rate was bettered by only chris paul and deron williams...
yet the teams he played on averaged only about 30 wins a season. do you really think adjusting his assists upwards would make him in the minds of some the best "passer" in the game over the last decade and a half?...
why is it hard to believe brevin knight *could* be one of, if not the best, pure passers in the entire game? brevin was 5'10 (optimistically speaking) without a lot of athletic ability and 0 shooting ability. He had unbelievably quick hands, but even with a great STL%, wasn't ever considered a positive on defense. (And possibly was a detriment) But he had an 11 year career playing over 18,000 minutes. Clearly he was doing *something* well, right?
_________________
-Rob
Back to top
View user's profile Send private message AIM Address
bchaikin
Joined: 27 Jan 2005
Posts: 686
Location: cleveland, ohio
PostPosted: Fri Apr 02, 2010 1:16 am Post subject: Reply with quote
why is it hard to believe brevin knight *could* be one of, if not the best, pure passers in the entire game?... Clearly he was doing *something* well, right?
well then it depends on what your definition of what a "pure passer" is. if its a player that throws a ton of passes and gets alot of assists simply because he throws a ton of passes, then i'd say he could be...
the fact is that in today's game 56%-57% of all FGM have an associated assist. over the past 15 years that percentage goes up to close to 60%. if you assume say just 1/10 of all FGM are clearly on one-on-one plays where a basket is scored and there was no assist (and i'm guessing on that 1/10 number), then close to 2/3 of all of the other FGM had an associated assist - i.e. something like 2/3 of all other FGM preceded by a pass had that pass recorded as an assist...
so the question becomes what proportion of those assists on those other 2/3 of FGM were due more to the passing ability of the passer rather than the scoring ability of the player receiving the pass? or...
The issue essentially boils down to the chicken or the egg problem. Does the passer create the open shots for the offense or does the offense create open shots for the passer?
however if your definition of what a "pure passer" is is not simply just someone who gets alot of assists because he throws alot of passes, but someone who either (1) gets a higher percentage of assists per pass that he throws, or (2) increases the FG% of the players he throws passes to moreso that other passers do, then there should probably be a better measure than simply the number of assists thrown for by a player...
...because there are and have been many players who throw for alot of assists simply because they throw a ton of passes, as something like 2/3 of all FGM that were preceded by a pass had that pass recorded as an assist...
again for example looking at the past 15 seasons and all those players that have (a) played at least 3000 total minutes up to and through the age of 30, and (b) have thrown for assists at a rate of (again an arbitrary number here) say 9 ast/40min, you will find not only names like chris paul, deron williams, jason kidd, and steve nash, but also names like brevin knight, robert pack, sergio rodriguez, and jamaal tinsley...
are there people out there that consider players like pack, rodriguez, and tinsley to have been some of the best "passers" over the last decade and a half? or just as players who threw for a ton of assists because they handled the ball so much, shot infrequently, and threw a ton of passes?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Apr 02, 2010 7:59 am Post subject: Reply with quote
bchaikin wrote:
are there people out there that consider players like pack, rodriguez, and tinsley to have been some of the best "passers" over the last decade and a half? or just as players who threw for a ton of assists because they handled the ball so much, shot infrequently, and threw a ton of passes?...
Because I am looking at where the assist went to, I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. I don't care if they didn't shoot much--if their assists were going for dunks and layups, they should get some credit for it!
Interestingly, the players that get the biggest boost percentage-wise are the post players. Last year, Yao and KG both got a 30% boost because of where their assists went to. The ones hurt the most are players who just made passes to the midrange--Stuckey, Chalmers, Prince, Lee.
Oddly enough, despite the ones helped often being post players, Andrew Bynum is one hurt the most by the adjustment.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
habetw4
Joined: 12 Nov 2009
Posts: 22
Location: CT
PostPosted: Fri Apr 02, 2010 9:04 am Post subject: Reply with quote
Are we to assume that there aren't good passers who throw a ton of passes? Are the two mutually exclusive? Surely, there are players who specialize in ball distribution just like there are those who exclusively specialize in rebounding, guarding the opponent's superstar, or defending the rim.
_________________
I'm a twitterererer: @tomhaberstroh.
bchaikin
Joined: 27 Jan 2005
Posts: 681
Location: cleveland, ohio
PostPosted: Fri Apr 02, 2010 1:48 pm Post subject: Reply with quote
Because I am looking at where the assist went to, I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. I don't care if they didn't shoot much--if their assists were going for dunks and layups, they should get some credit for it!
who should get some credit for it? the passer or the player who made the cut to the basket or got into position close enough to the basket to get the dunk in the first place?...
sine 04-05 both amare stoudemire and zach randolph have both been 20/10 players (or close to it). stoudemire has shot 57% on 2s, randolph 47%. stoudemire has played with primarily one PG (nash), randolph with at least four (four different teams) if not more. are you of the opinion that if randolph had played all 5 seasons with phoenix, and stoudemire with the 4 teams zach had played with, that those two 2pt FG%s would be reversed? all because nash supposedly throws passes that lead to layups and dunks (with no "assist" credit to the player who got the layup or dunk)?...
because if you watch both players stoudemire cuts to the basket far more often than does randolph. shouldn't he get credit for getting himself open often for a player like nash to get him the pass in the first place?...
was there a better "passer" than john stockton? from 1987-88 to 1996-97, ten years, john stockton passed for about 1000 assists/yr. no one else in the league passed for even 700 ast/yr. the following players played for utah during this time, but also for other teams (karl malone, mark eaton, and thurl bailey played primarily for utah):
tyrone corbin
jeff malone
jeff hornacek
david benoit
mike brown
blue edwards
antoine carr
felton spencer
adam keefe
tom chambers
when these 10 players all played for utah (and stockton was there throwing for his 1000 ast/yr from 8788-9697) as a collective group they shot a 49.9% 2pt FG%. for all the other teams they played with (22 in all) during their careers as a collective group they shot a 48.4% 2pt FG%...
now i would assume that of that 48.4% 2pt FG% with other teams that with some of those PGs they shot a bit worse than 48.4% and with some a bit better, meaning that some of those PGs threw for assists to these players where they shot 49%-50%, similar to stockton...
does that mean those PGs were as good of "passers" as stockton was?...
Are we to assume that there aren't good passers who throw a ton of passes?
again - what if your definition of a "good passer"?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Apr 02, 2010 2:05 pm Post subject: Reply with quote
bchaikin wrote:
who should get some credit for it? the passer or the player who made the cut to the basket or got into position close enough to the basket to get the dunk in the first place?...
sine 04-05 both amare stoudemire and zach randolph have both been 20/10 players (or close to it). stoudemire has shot 57% on 2s, randolph 47%. stoudemire has played with primarily one PG (nash), randolph with at least four (four different teams) if not more. are you of the opinion that if randolph had played all 5 seasons with phoenix, and stoudemire with the 4 teams zach had played with, that those two 2pt FG%s would be reversed? all because nash supposedly throws passes that lead to layups and dunks (with no "assist" credit to the player who got the layup or dunk)?...
because if you watch both players stoudemire cuts to the basket far more often than does randolph. shouldn't he get credit for getting himself open often for a player like nash to get him the pass in the first place?...
was there a better "passer" than john stockton? from 1987-88 to 1996-97, ten years, john stockton passed for about 1000 assists/yr. no one else in the league passed for even 700 ast/yr. the following players played for utah during this time, but also for other teams (karl malone, mark eaton, and thurl bailey played primarily for utah):
tyrone corbin
jeff malone
jeff hornacek
david benoit
mike brown
blue edwards
antoine carr
felton spencer
adam keefe
tom chambers
when these 10 players all played for utah (and stockton was there throwing for his 1000 ast/yr from 8788-9697) as a collective group they shot a 49.9% 2pt FG%. for all the other teams they played with (22 in all) during their careers as a collective group they shot a 48.4% 2pt FG%...
now i would assume that of that 48.4% 2pt FG% with other teams that with some of those PGs they shot a bit worse than 48.4% and with some a bit better, meaning that some of those PGs threw for assists to these players where they shot 49%-50%, similar to stockton...
does that mean those PGs were as good of "passers" as stockton was?...
Are we to assume that there aren't good passers who throw a ton of passes?
again - what if your definition of a "good passer"?...
I have no idea about actual "point" credit. I'm simply trying to see which assist is worth more than the others.
Here is what I am doing.
1. Calculate the UNASSISTED team eFG%. This is critical. THIS is where there is a correction for teammates.
2. Calculate the eFG% for all potential assist passes by the player. (This differs depending on how good the team is from EACH LOCATION on ASSISTED eFG%)
3. Subtract unassisted team eFG% from player assisted eFG% and multiply the difference times total # of opportunities.
4. Total all of the NBA; normalize the value to the total number of assists handed out.
Sure, some of the credit for actually being able to make the pass goes to the recipient of the pass... but they already get credit for the score, anyway.
If a player is playing for a good midrange team (where the eFG% is high on assisted midrange jumpers) the player will get credit. If the assisted eFG% on midrange jumpers is lower than the team unassisted eFG%... there was basically no value added. (Why was the shot taken?)
All of this is predicated on all players improving their opponents shot in a given position on the court by the same amount, which is obviously shaky ground. Perhaps Kidd's pass to Dirk is better than JJ Barea's pass to Dirk in the same location. Perhaps CP3's lob to the rim is better than Darren Collison's. But until we can estimate this (which will be impossible until "missed assists" are kept track of in PbP), this is about the best we can do I think.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3535
Location: Hendersonville, NC
PostPosted: Fri Apr 02, 2010 5:54 pm Post subject: Reply with quote
DSMok1 wrote:
... I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. ..
Isn't the 'problem' resolved already? When you pass to high-% areas, you get more assists.
If you have 4 seconds to get a shot, a 33% shot might be as good as you are gonna get. So you get an assist on 1/3 of those passes.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Mon Apr 05, 2010 7:58 am Post subject: Reply with quote
Mike G wrote:
DSMok1 wrote:
... I think the problem is resolved--there is almost no credit for an assist to a midrange jumper but there is a lot of credit for an assist to a layup. ..
Isn't the 'problem' resolved already? When you pass to high-% areas, you get more assists.
If you have 4 seconds to get a shot, a 33% shot might be as good as you are gonna get. So you get an assist on 1/3 of those passes.
The "truth" is between my "adjusted assists" and the actual assists. One measures successes, the other tries to quantify value added.
True, that may be the best possibility given the situation, but it is also likely to not the best possibility if the person doing the passing were a bit better player, like a Steve Nash. Wouldn't you consider a player with 5 assists on dunks a better passer than a player with 5 assists on midrange jumpers?
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3535
Location: Hendersonville, NC
PostPosted: Mon Apr 05, 2010 8:19 am Post subject: Reply with quote
DSMok1 wrote:
Mike G wrote:
When you pass to high-% areas, you get more assists.
.
Wouldn't you consider a player with 5 assists on dunks a better passer than a player with 5 assists on midrange jumpers?
Suppose these players are equal in every respect, except that in a span of 20 possessions:
- Player A gets 6 passes into the hands of someone at the rim, resulting in 5 successful shots (dunks).
- Player B gets the ball 12 times to jumpshooters, who convert 5 times.
We don't know what they did with/without the ball on their other possessions. It could be A is the 'risk taker' and B is the 'play it safe' guy.
Maybe the teams play different styles (fastbreak vs halfcourt).
One guy did not feed a teammate for a shot on 8 possessions, and the other guy did not on 14 occassions.
Assists are equal to attempts * effectiveness.
If Player A could get 12 passes to players in dunking range, and they convert 10, then he has twice as many Assists as player B has. Now he's a better passer.[/i]
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Mon Apr 05, 2010 11:22 am Post subject: Reply with quote
Mike G wrote:
Suppose these players are equal in every respect, except that in a span of 20 possessions:
- Player A gets 6 passes into the hands of someone at the rim, resulting in 5 successful shots (dunks).
- Player B gets the ball 12 times to jumpshooters, who convert 5 times.
We don't know what they did with/without the ball on their other possessions. It could be A is the 'risk taker' and B is the 'play it safe' guy.
Maybe the teams play different styles (fastbreak vs halfcourt).
One guy did not feed a teammate for a shot on 8 possessions, and the other guy did not on 14 occassions.
Assists are equal to attempts * effectiveness.
If Player A could get 12 passes to players in dunking range, and they convert 10, then he has twice as many Assists as player B has. Now he's a better passer.[/i]
effectiveness. The baseline effectiveness is NOT 0. This is where assists go wrong. What is the value added by the sequence of pass-shot? One must compare to what the team would do without the passing, the baseline UNASSISTED fg% of the team.
The current valuation of assists is attempts*effectiveness. My valuation is attempts*(effectiveness-unassisted effectiveness), scaled so the total is the same as just plain assists. I think it's a valuable number to be able to look at.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3535
Location: Hendersonville, NC
PostPosted: Wed Apr 07, 2010 11:31 am Post subject: Reply with quote
DSMok1 wrote:
...
The current valuation of assists is attempts*effectiveness. My valuation is attempts*(effectiveness-unassisted effectiveness)..
Quote:
One must compare to what the team would do without the passing, the baseline UNASSISTED fg% of the team.
So, where is a table of teams' Unassisted FG% ?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 585
Location: Where the wind comes sweeping down the plains
PostPosted: Wed Apr 07, 2010 1:32 pm Post subject: Reply with quote
Mike G wrote:
So, where is a table of teams' Unassisted FG% ?
That's the rub. If the play-by-play recorded whether there was a potential assist on the play, then it would be easy to account for. As it is, I had to estimate it based on the FG%, AST%, and estimated (from The Value of a Good Pass) FG% added by assist for each portion of the floor.
The estimated unassisted eFG% for each team:
Team UAeFG%
PHO 49.1%
ORL 48.5%
CLE 48.1%
TOR 46.6%
BOS 46.5%
GSW 46.4%
SAS 46.2%
DEN 46.1%
NYK 45.7%
UTH 45.6%
ATL 45.3%
MEM 45.1%
NOR 45.0%
LAL 44.8%
DAL 44.7%
POR 44.6%
MIA 44.3%
SAC 44.2%
HOU 44.0%
OKC 43.7%
PHI 43.6%
IND 43.4%
LAC 43.4%
CHA 43.3%
MIL 43.2%
WAS 43.1%
MIN 42.5%
CHI 42.2%
DET 41.9%
NJN 40.2%
If a team gets all of their dunks off of assists, no dunks are assumed in calculating the UAeFG%. Phoenix is so high because they shoot a ridiculously high percentage from 3, and unusually few of their 3's are assisted (I'm guessing Nash is shooting a lot of those!). Orlando is so high because unusually few of their layups/dunks are assisted and they shoot very well from the midrange (and few of those are assisted).
Do you see what I was trying to do? Obviously, this is incomplete unless we know the true UAeFG% and how much each team's eFG% improves in each area with each assist. It's a start, though.
Re: More recovered assist threads
see also
A different way of thinking of assists
http://www.sonicscentral.com/apbrmetric ... 95d0bf1a52
Weight of Assists
http://sonicscentral.com/apbrmetrics/vi ... p?f=2&t=73
Are top-rated point guards under-rated by PER?
http://sonicscentral.com/apbrmetrics/vi ... ?f=2&p=457
and the latter part of this one
http://sonicscentral.com/apbrmetrics/vi ... ?f=2&t=209
and mentioned / endorsed in Ed K's rules of thumb
http://sonicscentral.com/apbrmetrics/vi ... .php?t=166
A different way of thinking of assists
http://www.sonicscentral.com/apbrmetric ... 95d0bf1a52
Weight of Assists
http://sonicscentral.com/apbrmetrics/vi ... p?f=2&t=73
Are top-rated point guards under-rated by PER?
http://sonicscentral.com/apbrmetrics/vi ... ?f=2&p=457
and the latter part of this one
http://sonicscentral.com/apbrmetrics/vi ... ?f=2&t=209
and mentioned / endorsed in Ed K's rules of thumb
http://sonicscentral.com/apbrmetrics/vi ... .php?t=166
Last edited by Crow on Wed May 04, 2011 7:21 pm, edited 5 times in total.
Re: More recovered assist threads
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Wed Mar 22, 2006 2:57 am Post subject: 82games: Potential Assists Reply with quote
http://www.82games.com/assisted.htm
The value of an assist has long been a somewhat contentious subject for this group, and 82games' charting project sheds some light on the subject in its most valuable contribution to the public domain thus far.
Thoughts?
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mark
Joined: 20 Aug 2005
Posts: 807
PostPosted: Wed Mar 22, 2006 11:43 am Post subject: Gamecharting method Reply with quote
I am still digesting it but I tend to accept the concept of potential assists as having some discussion value and the modest point value derived for them sounds fair.
I dont think the turnover rate can be cut out of the comparison with non-assisted. The high turnover rate of completely not assisted is significant and real. Getting anothe rplayer to the shot or getting themn into a move and a foul or getting closer to doing something even it is a turnover trying to make a move/shot does have value reducing exposure to "futsing" around turnovers and some of the low percentage self-made / forced shots that occur more often in the not assisted distibution.
Last edited by Mark on Wed Mar 22, 2006 3:28 pm; edited 5 times in total
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 787
Location: Toronto
PostPosted: Wed Mar 22, 2006 12:05 pm Post subject: Reply with quote
Some of you may remember that I charted a bunch of games a few years ago to study this very thing, and my results were consistent with the results at 82games. My problem is that no sample sizes or confidence intervals or standard errors are shown, so it's hard to know the significance of those results. When I did my own study (20 or 30 games charted), the two-point shots were significant, but the three-pointers and FTs were nowhere near significant — in the first case because the number of shots was relatively small and in the latter because the difference was small. I wish the number of attempts, or at least games charted, was reported.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
94by50
Joined: 01 Jan 2006
Posts: 499
Location: Phoenix
PostPosted: Wed Mar 22, 2006 12:11 pm Post subject: Reply with quote
Ed Küpfer wrote:
Some of you may remember that I charted a bunch of games a few years ago to study this very thing, and my results were consistent with the results at 82games. My problem is that no sample sizes or confidence intervals or standard errors are shown, so it's hard to know the significance of those results. When I did my own study (20 or 30 games charted), the two-point shots were significant, but the three-pointers and FTs were nowhere near significant — in the first case because the number of shots was relatively small and in the latter because the difference was small. I wish the number of attempts, or at least games charted, was reported.
I had the same concern when I first read the column - that since there were no totals given, we don't know if the differences are significant or not. I assumed that they were, since there are thousands of shots and potential assists during a season. However, it's probably better to be sure than to assume that's the case.
Back to top
View user's profile Send private message
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Wed Mar 22, 2006 12:17 pm Post subject: Reply with quote
One thing not clear to me from the article was the turnover rate on passes deemed to be "potential assists." That's important information for determining the worth of an assist.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Ed Küpfer
Joined: 30 Dec 2004
Posts: 787
Location: Toronto
PostPosted: Wed Mar 22, 2006 1:57 pm Post subject: Reply with quote
94by50 wrote:
I had the same concern when I first read the column - that since there were no totals given, we don't know if the differences are significant or not. I assumed that they were, since there are thousands of shots and potential assists during a season. However, it's probably better to be sure than to assume that's the case.
It's an important point, given the small differences in some categories. For example, the difference in FG% between assisted 3pt% and unassisted is .379 - .342 = .037, with the assisted category accounting for 81% of all attempts. Fiddling around with my chi-sq calulator, the 2-tail p-value of that FG% difference given 1000 3-point attempts is 0.34. You need 3000 attempts to get the p-value below 0.1, and 4300 attempts before it gets below 0.05. That is a lot of attempts — you'd need to chart about 173 games, or about 1/8 of of the games played this season, to see that many attempts.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
THWilson
Joined: 19 Jul 2005
Posts: 164
Location: phoenix
PostPosted: Wed Mar 22, 2006 2:21 pm Post subject: Reply with quote
Ed Küpfer wrote:
Some of you may remember that I charted a bunch of games a few years ago to study this very thing, and my results were consistent with the results at 82games. My problem is that no sample sizes or confidence intervals or standard errors are shown, so it's hard to know the significance of those results. When I did my own study (20 or 30 games charted), the two-point shots were significant, but the three-pointers and FTs were nowhere near significant — in the first case because the number of shots was relatively small and in the latter because the difference was small. I wish the number of attempts, or at least games charted, was reported.
http://www.82games.com/charters/chart1.htm
29 games are listed, one assumes the number was not larger than that...
Back to top
View user's profile Send private message
Carlos
Joined: 21 Jan 2005
Posts: 64
Location: Montevideo, Uruguay
PostPosted: Wed Mar 22, 2006 3:29 pm Post subject: Reply with quote
Besides the sample size and the small difference between assisted and unassisted on some type of shots, I'm dubious of the idea of comparing the fg% of an assisted shot (of a certain type) with an unassisted one of the same type. Some type of shots are much more easily available than others; the alternative to an assisted dunk is probably not an unassisted dunk but an unassisted jump shot.
Back to top
View user's profile Send private message
Mark
Joined: 20 Aug 2005
Posts: 807
PostPosted: Wed Mar 22, 2006 3:38 pm Post subject: Reply with quote
Interesting. The game video charting does sound project specific and short-term and called "noble" so I'd guess probably volunteer. Thanks guys for the new info you are generating and thanks as always for sharing a lot of it Roland.
So I now assume the main data set is indeed based off written play by play exclusively and lay the general data collection method question to rest unless informed directly otherwise. Wouldnt need a separate special proect video tape exercise if you were already doing video analysis of every regular season game, I assume you'd just tell those folks to do a special add-on? I guess the pre-season test must have hit some obstacle- time, funding, etc. Wonder if something bigger will be tried again next year, or even just the playoffs (this year or next)- that might be a smaller piece to bite off and quite valuable.
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 787
Location: Toronto
PostPosted: Wed Mar 22, 2006 4:04 pm Post subject: Reply with quote
Man, here I am sounding all negative. I don't want to give that off. It's a worthwhile project, and the results can be useful even if they are not statistically significant, especially when they are congruent with our common sense, as these seem to be. The solution to small samples is more samples, not me dissing the whole thing.
Carlos: divvying up the categories is problematic, as you mention, but still very useful. We can still combine the categories after the fact.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Wed Mar 22, 2006 6:48 pm Post subject: Reply with quote
The one thing I'd be curious to see is a breakdown by passer or toucher. I had the honor of charting half of the NJN/NOK game, and I noted (subjectively and anecdotally) that JKidd and CPaul seemed to generate potential assists on about 75-80% of their touches. In addition, the overall prevalence of potential assists for their respective teams seemed to be higher while they were on the floor.
Perhaps it was just confirmation bias on my part to pick out those two. However, in general I would speculate that if a player generates a higher PAst/touch, it would seem that they are doing something right.
Back to top
View user's profile Send private message Send e-mail AIM Address
ziller
Joined: 30 Jun 2005
Posts: 126
Location: Sac Metro
PostPosted: Thu Mar 23, 2006 1:07 pm Post subject: Reply with quote
gabefarkas wrote:
The one thing I'd be curious to see is a breakdown by passer or toucher. I had the honor of charting half of the NJN/NOK game, and I noted (subjectively and anecdotally) that JKidd and CPaul seemed to generate potential assists on about 75-80% of their touches. In addition, the overall prevalence of potential assists for their respective teams seemed to be higher while they were on the floor.
Perhaps it was just confirmation bias on my part to pick out those two. However, in general I would speculate that if a player generates a higher PAst/touch, it would seem that they are doing something right.
And this is what I think is key about the record-keeping of potential assists: finding better metrics than assists and assists per 100 possessions for passing skills. Potential assists to turnovers seems like ratio that could tell us a lot, and potential assists to field goal attempts would give us a terrific metric on "point-guardedness." I think with potential assists, you can start mixing and matching tandems - guards who can set up a dunk or close shot efficiently with great finishers (Nash, Stoudemire) and bigs who thrive on getting wing players free for open jumpers and wing players who can nail open jumpers (B. Miller, Bibby). This, mixed with 82games' existing close/jump shot stats, could be powerful.
_________________
SactownRoyalty.com
tziller@gmail.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Fri Mar 31, 2006 1:51 pm Post subject: Reply with quote
This thread was about to slide off of page 1 without an attempt to analyze. So here I go. The data at 82games can be summarized thus:
pAst = potential assist
uAst = unassisted
shot type: 3FGA - 2 FGA
pAst FG% - .379 - .502
uAst FG% - .342 - .421
pAst/FGA - .810 - .560
Thus 3's are 81% potentially-assisted.
Here's what I've derived:
shot type : 3-FGA - 2-FGA -- derivations
pAstPts/Att - 1.14 - 1.00 = .379*3 -- .502*2
uAstPts/Att - 1.03 - 0.84 = .342*3 -- .421*2
diff Pts/Att - .111 - .162 = 1.14-1.03 -- 1.00-.84
diff Pts/Ast - .293 - .323 = .111/.379 -- .162/.502
While an assist-worthy pass adds some .16 to each 2-pt FGA, only about half of those passes are registered (by converting the shot) as assists. So each (2Pt) assist represents 2 potential ones.
Tentative conclusions:
A good pass to a player in 2-pt range is worth about 10% more than one to a player at the arc. On average.
A value of roughly 0.3 Pts for an Ast doesn't seem like much. The very good point was made, that the alternative to an assisted dunk is likely an unassisted jump shot.
And if the passer is given credit of .3 pts for his Ast, then the scorer should get only 1.7 (or 2.7) for his FG
?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
cherokee_ACB
Joined: 22 Mar 2006
Posts: 157
PostPosted: Fri Mar 31, 2006 2:52 pm Post subject: Reply with quote
Mike G wrote:
A value of roughly 0.3 Pts for an Ast doesn't seem like much. The very good point was made, that the alternative to an assisted dunk is likely an unassisted jump shot.
Or a turnover, in roughly 20% of cases. That's why I don't think you can ignore turnovers and FTs in the analysis. If you factor them in, you get to a value between 0.6 and 0.7, consistent with the one used by PER.
Back to top
View user's profile Send private message
capnhistory
Joined: 27 Jul 2005
Posts: 62
Location: Durham, North Carolina
PostPosted: Mon Apr 10, 2006 6:20 am Post subject: Reply with quote
cherokee_ACB wrote:
Mike G wrote:
A value of roughly 0.3 Pts for an Ast doesn't seem like much. The very good point was made, that the alternative to an assisted dunk is likely an unassisted jump shot.
Or a turnover, in roughly 20% of cases. That's why I don't think you can ignore turnovers and FTs in the analysis. If you factor them in, you get to a value between 0.6 and 0.7, consistent with the one used by PER.
I think these two facts will be crucial to determining th utility of this research. First, I am glad to see the evidence appears to confirm something we already suspected (that an AST tends to be worth between 0.6-0.7 points). It also helps that the value of an AST for each kind of shot lines up with at least my current understanding. Potential AST for three-pointers should be valued least see those shots are the hardest to make, and therefore the passes are least likely to pay off. Dunks are more valuable since they are a higher percentage shot. Howeverthe fact that it's a dunk probably means the player making the shot positioned himself well, while the passer made a pretty obvious choice to throw to the big guy right under the basket. So it seems logical that the more valuable passes should be those that turn into close contested shots where a bad choice by the passer could hurt the team more. Second, I think this could open the door to a better understanding of the mental aspects of the game. By assessing the value of passing decisions we may get more insight into which players actually have the mental edge (call it court vision call it basketball IQ, call it whatever you like). This is huge to me because it always seemed to be the aspect of the game stats couldn't penetrate.
_________________
Throw it down big man!
More expansive basketball babble at a slower pace The Captain of History
Site Admin
Joined: 30 Dec 2004
Posts: 979
Location: Seattle
PostPosted: Wed Mar 22, 2006 2:57 am Post subject: 82games: Potential Assists Reply with quote
http://www.82games.com/assisted.htm
The value of an assist has long been a somewhat contentious subject for this group, and 82games' charting project sheds some light on the subject in its most valuable contribution to the public domain thus far.
Thoughts?
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mark
Joined: 20 Aug 2005
Posts: 807
PostPosted: Wed Mar 22, 2006 11:43 am Post subject: Gamecharting method Reply with quote
I am still digesting it but I tend to accept the concept of potential assists as having some discussion value and the modest point value derived for them sounds fair.
I dont think the turnover rate can be cut out of the comparison with non-assisted. The high turnover rate of completely not assisted is significant and real. Getting anothe rplayer to the shot or getting themn into a move and a foul or getting closer to doing something even it is a turnover trying to make a move/shot does have value reducing exposure to "futsing" around turnovers and some of the low percentage self-made / forced shots that occur more often in the not assisted distibution.
Last edited by Mark on Wed Mar 22, 2006 3:28 pm; edited 5 times in total
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 787
Location: Toronto
PostPosted: Wed Mar 22, 2006 12:05 pm Post subject: Reply with quote
Some of you may remember that I charted a bunch of games a few years ago to study this very thing, and my results were consistent with the results at 82games. My problem is that no sample sizes or confidence intervals or standard errors are shown, so it's hard to know the significance of those results. When I did my own study (20 or 30 games charted), the two-point shots were significant, but the three-pointers and FTs were nowhere near significant — in the first case because the number of shots was relatively small and in the latter because the difference was small. I wish the number of attempts, or at least games charted, was reported.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
94by50
Joined: 01 Jan 2006
Posts: 499
Location: Phoenix
PostPosted: Wed Mar 22, 2006 12:11 pm Post subject: Reply with quote
Ed Küpfer wrote:
Some of you may remember that I charted a bunch of games a few years ago to study this very thing, and my results were consistent with the results at 82games. My problem is that no sample sizes or confidence intervals or standard errors are shown, so it's hard to know the significance of those results. When I did my own study (20 or 30 games charted), the two-point shots were significant, but the three-pointers and FTs were nowhere near significant — in the first case because the number of shots was relatively small and in the latter because the difference was small. I wish the number of attempts, or at least games charted, was reported.
I had the same concern when I first read the column - that since there were no totals given, we don't know if the differences are significant or not. I assumed that they were, since there are thousands of shots and potential assists during a season. However, it's probably better to be sure than to assume that's the case.
Back to top
View user's profile Send private message
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Wed Mar 22, 2006 12:17 pm Post subject: Reply with quote
One thing not clear to me from the article was the turnover rate on passes deemed to be "potential assists." That's important information for determining the worth of an assist.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Ed Küpfer
Joined: 30 Dec 2004
Posts: 787
Location: Toronto
PostPosted: Wed Mar 22, 2006 1:57 pm Post subject: Reply with quote
94by50 wrote:
I had the same concern when I first read the column - that since there were no totals given, we don't know if the differences are significant or not. I assumed that they were, since there are thousands of shots and potential assists during a season. However, it's probably better to be sure than to assume that's the case.
It's an important point, given the small differences in some categories. For example, the difference in FG% between assisted 3pt% and unassisted is .379 - .342 = .037, with the assisted category accounting for 81% of all attempts. Fiddling around with my chi-sq calulator, the 2-tail p-value of that FG% difference given 1000 3-point attempts is 0.34. You need 3000 attempts to get the p-value below 0.1, and 4300 attempts before it gets below 0.05. That is a lot of attempts — you'd need to chart about 173 games, or about 1/8 of of the games played this season, to see that many attempts.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
THWilson
Joined: 19 Jul 2005
Posts: 164
Location: phoenix
PostPosted: Wed Mar 22, 2006 2:21 pm Post subject: Reply with quote
Ed Küpfer wrote:
Some of you may remember that I charted a bunch of games a few years ago to study this very thing, and my results were consistent with the results at 82games. My problem is that no sample sizes or confidence intervals or standard errors are shown, so it's hard to know the significance of those results. When I did my own study (20 or 30 games charted), the two-point shots were significant, but the three-pointers and FTs were nowhere near significant — in the first case because the number of shots was relatively small and in the latter because the difference was small. I wish the number of attempts, or at least games charted, was reported.
http://www.82games.com/charters/chart1.htm
29 games are listed, one assumes the number was not larger than that...
Back to top
View user's profile Send private message
Carlos
Joined: 21 Jan 2005
Posts: 64
Location: Montevideo, Uruguay
PostPosted: Wed Mar 22, 2006 3:29 pm Post subject: Reply with quote
Besides the sample size and the small difference between assisted and unassisted on some type of shots, I'm dubious of the idea of comparing the fg% of an assisted shot (of a certain type) with an unassisted one of the same type. Some type of shots are much more easily available than others; the alternative to an assisted dunk is probably not an unassisted dunk but an unassisted jump shot.
Back to top
View user's profile Send private message
Mark
Joined: 20 Aug 2005
Posts: 807
PostPosted: Wed Mar 22, 2006 3:38 pm Post subject: Reply with quote
Interesting. The game video charting does sound project specific and short-term and called "noble" so I'd guess probably volunteer. Thanks guys for the new info you are generating and thanks as always for sharing a lot of it Roland.
So I now assume the main data set is indeed based off written play by play exclusively and lay the general data collection method question to rest unless informed directly otherwise. Wouldnt need a separate special proect video tape exercise if you were already doing video analysis of every regular season game, I assume you'd just tell those folks to do a special add-on? I guess the pre-season test must have hit some obstacle- time, funding, etc. Wonder if something bigger will be tried again next year, or even just the playoffs (this year or next)- that might be a smaller piece to bite off and quite valuable.
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 787
Location: Toronto
PostPosted: Wed Mar 22, 2006 4:04 pm Post subject: Reply with quote
Man, here I am sounding all negative. I don't want to give that off. It's a worthwhile project, and the results can be useful even if they are not statistically significant, especially when they are congruent with our common sense, as these seem to be. The solution to small samples is more samples, not me dissing the whole thing.
Carlos: divvying up the categories is problematic, as you mention, but still very useful. We can still combine the categories after the fact.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Wed Mar 22, 2006 6:48 pm Post subject: Reply with quote
The one thing I'd be curious to see is a breakdown by passer or toucher. I had the honor of charting half of the NJN/NOK game, and I noted (subjectively and anecdotally) that JKidd and CPaul seemed to generate potential assists on about 75-80% of their touches. In addition, the overall prevalence of potential assists for their respective teams seemed to be higher while they were on the floor.
Perhaps it was just confirmation bias on my part to pick out those two. However, in general I would speculate that if a player generates a higher PAst/touch, it would seem that they are doing something right.
Back to top
View user's profile Send private message Send e-mail AIM Address
ziller
Joined: 30 Jun 2005
Posts: 126
Location: Sac Metro
PostPosted: Thu Mar 23, 2006 1:07 pm Post subject: Reply with quote
gabefarkas wrote:
The one thing I'd be curious to see is a breakdown by passer or toucher. I had the honor of charting half of the NJN/NOK game, and I noted (subjectively and anecdotally) that JKidd and CPaul seemed to generate potential assists on about 75-80% of their touches. In addition, the overall prevalence of potential assists for their respective teams seemed to be higher while they were on the floor.
Perhaps it was just confirmation bias on my part to pick out those two. However, in general I would speculate that if a player generates a higher PAst/touch, it would seem that they are doing something right.
And this is what I think is key about the record-keeping of potential assists: finding better metrics than assists and assists per 100 possessions for passing skills. Potential assists to turnovers seems like ratio that could tell us a lot, and potential assists to field goal attempts would give us a terrific metric on "point-guardedness." I think with potential assists, you can start mixing and matching tandems - guards who can set up a dunk or close shot efficiently with great finishers (Nash, Stoudemire) and bigs who thrive on getting wing players free for open jumpers and wing players who can nail open jumpers (B. Miller, Bibby). This, mixed with 82games' existing close/jump shot stats, could be powerful.
_________________
SactownRoyalty.com
tziller@gmail.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Fri Mar 31, 2006 1:51 pm Post subject: Reply with quote
This thread was about to slide off of page 1 without an attempt to analyze. So here I go. The data at 82games can be summarized thus:
pAst = potential assist
uAst = unassisted
shot type: 3FGA - 2 FGA
pAst FG% - .379 - .502
uAst FG% - .342 - .421
pAst/FGA - .810 - .560
Thus 3's are 81% potentially-assisted.
Here's what I've derived:
shot type : 3-FGA - 2-FGA -- derivations
pAstPts/Att - 1.14 - 1.00 = .379*3 -- .502*2
uAstPts/Att - 1.03 - 0.84 = .342*3 -- .421*2
diff Pts/Att - .111 - .162 = 1.14-1.03 -- 1.00-.84
diff Pts/Ast - .293 - .323 = .111/.379 -- .162/.502
While an assist-worthy pass adds some .16 to each 2-pt FGA, only about half of those passes are registered (by converting the shot) as assists. So each (2Pt) assist represents 2 potential ones.
Tentative conclusions:
A good pass to a player in 2-pt range is worth about 10% more than one to a player at the arc. On average.
A value of roughly 0.3 Pts for an Ast doesn't seem like much. The very good point was made, that the alternative to an assisted dunk is likely an unassisted jump shot.
And if the passer is given credit of .3 pts for his Ast, then the scorer should get only 1.7 (or 2.7) for his FG
?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
cherokee_ACB
Joined: 22 Mar 2006
Posts: 157
PostPosted: Fri Mar 31, 2006 2:52 pm Post subject: Reply with quote
Mike G wrote:
A value of roughly 0.3 Pts for an Ast doesn't seem like much. The very good point was made, that the alternative to an assisted dunk is likely an unassisted jump shot.
Or a turnover, in roughly 20% of cases. That's why I don't think you can ignore turnovers and FTs in the analysis. If you factor them in, you get to a value between 0.6 and 0.7, consistent with the one used by PER.
Back to top
View user's profile Send private message
capnhistory
Joined: 27 Jul 2005
Posts: 62
Location: Durham, North Carolina
PostPosted: Mon Apr 10, 2006 6:20 am Post subject: Reply with quote
cherokee_ACB wrote:
Mike G wrote:
A value of roughly 0.3 Pts for an Ast doesn't seem like much. The very good point was made, that the alternative to an assisted dunk is likely an unassisted jump shot.
Or a turnover, in roughly 20% of cases. That's why I don't think you can ignore turnovers and FTs in the analysis. If you factor them in, you get to a value between 0.6 and 0.7, consistent with the one used by PER.
I think these two facts will be crucial to determining th utility of this research. First, I am glad to see the evidence appears to confirm something we already suspected (that an AST tends to be worth between 0.6-0.7 points). It also helps that the value of an AST for each kind of shot lines up with at least my current understanding. Potential AST for three-pointers should be valued least see those shots are the hardest to make, and therefore the passes are least likely to pay off. Dunks are more valuable since they are a higher percentage shot. Howeverthe fact that it's a dunk probably means the player making the shot positioned himself well, while the passer made a pretty obvious choice to throw to the big guy right under the basket. So it seems logical that the more valuable passes should be those that turn into close contested shots where a bad choice by the passer could hurt the team more. Second, I think this could open the door to a better understanding of the mental aspects of the game. By assessing the value of passing decisions we may get more insight into which players actually have the mental edge (call it court vision call it basketball IQ, call it whatever you like). This is huge to me because it always seemed to be the aspect of the game stats couldn't penetrate.
_________________
Throw it down big man!
More expansive basketball babble at a slower pace The Captain of History
Re: More recovered assist threads
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Wed Dec 29, 2010 8:30 pm Post subject: Value of Assists Reply with quote
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 547
Location: Where the wind comes sweeping down the plains
PostPosted: Wed Dec 29, 2010 11:00 pm Post subject: Re: Value of Assists Reply with quote
EvanZ wrote:
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
There's lots of threads on this topic around here... you might search for them.
And yes, I think they are worth quite a bit.
_________________
GodismyJudgeOK.com/DStats
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3484
Location: Hendersonville, NC
PostPosted: Thu Dec 30, 2010 7:41 am Post subject: Reply with quote
If league FG% is .45, and .60 of all FG are assisted, then --
given that an assisted FG is by definition a 100%-made shot --
if unassisted FGA are .08 lower FG% than potentially-assisted FGA,
this indicates pretty close to .42 unassisted FG% vs .50 on p-a FGA.
However, the assist is only counted on the made (100%) FG.
Shouldn't the Assisted FGA actually be considered 2.4 times (100/42) as effective as the unassisted FGA?
In other words, potentially assisted FGA may be .080 better than unassisted FGA; but if only half of these are converted, the average assist yields 2 FGA which have a 20%-better (50/42) chance of going in.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Thu Dec 30, 2010 8:24 am Post subject: Reply with quote
Mike, I did take DSMok1's advice (admittedly, I should have done this before posting) and read several previous threads on the topic. In one exchange between you and DSMok1, you had pointed out - and I'm trying to paraphrase here - that the value of an assist is already sort of baked into the numbers, since a player will get more assists by finding open teammates at shot locations that yield higher percentages.
The reason I bring it up is that I've been having an internal debate about whether to assign different weights to an assist depending on shot location (inside vs. mid-range vs. 3-pt) and also quality of shooter (i.e. more valuable assists could be considered those that go to better shooter). But each time I've thought about this, I sort of come to the same conclusion that you seemed to previously - i.e. that the value is already in the numbers. The good point guard gets the ball to open players who are in a position to take good shots. Ideally, the good passer causes an improvement in the eFG% of the shooter.
Is that still your feeling on it?
(I don't suppose at this point that Synergy tracks unassisted %? Seems like such an obvious use of that platform.)
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3484
Location: Hendersonville, NC
PostPosted: Thu Dec 30, 2010 11:16 am Post subject: Reply with quote
EvanZ wrote:
... the value of an assist is already sort of baked into the numbers, since a player will get more assists by finding open teammates at shot locations that yield higher percentages. ..
Two separate issues:
- the accruing of assists, a function of quantity and quality of passes
- the value of an individual assist, over the value of an unassisted shot
Quote:
Is that still your feeling on it?
Yes, plus home/away Ast/FG disparity, plus varying Pts/FG ratio.
Quote:
I don't suppose at this point that Synergy tracks unassisted %?
Doesn't 82games.com have that for each player?
I have a formula that estimates it. It's not great, also not bad. Within a few % for most players.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1311
Location: Durham, NC
PostPosted: Thu Dec 30, 2010 11:19 am Post subject: Re: Value of Assists Reply with quote
EvanZ wrote:
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
The confounding issue is the consistency in assists being assigned. It's subjective, so it's hard to say how much of an effect they have. In general, it probably is beneficial to emphasize "good passing" or "smart passing". But whether or not this leads to an assist getting recorded, I don't think we can say for sure.
Back to top
View user's profile Send private message Send e-mail AIM Address
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Thu Dec 30, 2010 11:44 am Post subject: Re: Value of Assists Reply with quote
gabefarkas wrote:
EvanZ wrote:
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
The confounding issue is the consistency in assists being assigned. It's subjective, so it's hard to say how much of an effect they have. In general, it probably is beneficial to emphasize "good passing" or "smart passing". But whether or not this leads to an assist getting recorded, I don't think we can say for sure.
Is the inconsistency systematic or random (between players, teams, home/away)? Can you point me to a study that addressed this? thanks.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
greyberger
Joined: 27 Sep 2010
Posts: 34
PostPosted: Thu Dec 30, 2010 3:57 pm Post subject: Reply with quote
Quote:
The reason I bring it up is that I've been having an internal debate about whether to assign different weights to an assist depending on shot location (inside vs. mid-range vs. 3-pt) and also quality of shooter (i.e. more valuable assists could be considered those that go to better shooter). But each time I've thought about this, I sort of come to the same conclusion that you seemed to previously - i.e. that the value is already in the numbers. The good point guard gets the ball to open players who are in a position to take good shots. Ideally, the good passer causes an improvement in the eFG% of the shooter.
Scott Sereday took on what (I think) is the same conundrum, distinguishing between types of assists in a linear SPM model. His articles on the reasoning behind it are on http://www.basketball-analysis.com/page/2/
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Thu Dec 30, 2010 4:07 pm Post subject: Reply with quote
greyberger wrote:
Scott Sereday took on what (I think) is the same conundrum, distinguishing between types of assists in a linear SPM model. His articles on the reasoning behind it are on http://www.basketball-analysis.com/page/2/
Thanks, I need to sit down and read that carefully.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Joined: 22 Nov 2010
Posts: 199
PostPosted: Wed Dec 29, 2010 8:30 pm Post subject: Value of Assists Reply with quote
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 547
Location: Where the wind comes sweeping down the plains
PostPosted: Wed Dec 29, 2010 11:00 pm Post subject: Re: Value of Assists Reply with quote
EvanZ wrote:
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
There's lots of threads on this topic around here... you might search for them.
And yes, I think they are worth quite a bit.
_________________
GodismyJudgeOK.com/DStats
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3484
Location: Hendersonville, NC
PostPosted: Thu Dec 30, 2010 7:41 am Post subject: Reply with quote
If league FG% is .45, and .60 of all FG are assisted, then --
given that an assisted FG is by definition a 100%-made shot --
if unassisted FGA are .08 lower FG% than potentially-assisted FGA,
this indicates pretty close to .42 unassisted FG% vs .50 on p-a FGA.
However, the assist is only counted on the made (100%) FG.
Shouldn't the Assisted FGA actually be considered 2.4 times (100/42) as effective as the unassisted FGA?
In other words, potentially assisted FGA may be .080 better than unassisted FGA; but if only half of these are converted, the average assist yields 2 FGA which have a 20%-better (50/42) chance of going in.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Thu Dec 30, 2010 8:24 am Post subject: Reply with quote
Mike, I did take DSMok1's advice (admittedly, I should have done this before posting) and read several previous threads on the topic. In one exchange between you and DSMok1, you had pointed out - and I'm trying to paraphrase here - that the value of an assist is already sort of baked into the numbers, since a player will get more assists by finding open teammates at shot locations that yield higher percentages.
The reason I bring it up is that I've been having an internal debate about whether to assign different weights to an assist depending on shot location (inside vs. mid-range vs. 3-pt) and also quality of shooter (i.e. more valuable assists could be considered those that go to better shooter). But each time I've thought about this, I sort of come to the same conclusion that you seemed to previously - i.e. that the value is already in the numbers. The good point guard gets the ball to open players who are in a position to take good shots. Ideally, the good passer causes an improvement in the eFG% of the shooter.
Is that still your feeling on it?
(I don't suppose at this point that Synergy tracks unassisted %? Seems like such an obvious use of that platform.)
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3484
Location: Hendersonville, NC
PostPosted: Thu Dec 30, 2010 11:16 am Post subject: Reply with quote
EvanZ wrote:
... the value of an assist is already sort of baked into the numbers, since a player will get more assists by finding open teammates at shot locations that yield higher percentages. ..
Two separate issues:
- the accruing of assists, a function of quantity and quality of passes
- the value of an individual assist, over the value of an unassisted shot
Quote:
Is that still your feeling on it?
Yes, plus home/away Ast/FG disparity, plus varying Pts/FG ratio.
Quote:
I don't suppose at this point that Synergy tracks unassisted %?
Doesn't 82games.com have that for each player?
I have a formula that estimates it. It's not great, also not bad. Within a few % for most players.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1311
Location: Durham, NC
PostPosted: Thu Dec 30, 2010 11:19 am Post subject: Re: Value of Assists Reply with quote
EvanZ wrote:
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
The confounding issue is the consistency in assists being assigned. It's subjective, so it's hard to say how much of an effect they have. In general, it probably is beneficial to emphasize "good passing" or "smart passing". But whether or not this leads to an assist getting recorded, I don't think we can say for sure.
Back to top
View user's profile Send private message Send e-mail AIM Address
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Thu Dec 30, 2010 11:44 am Post subject: Re: Value of Assists Reply with quote
gabefarkas wrote:
EvanZ wrote:
The 82games study found that fg% from potential assist were on average 8 %-points higher. Doesn't seem huge at first glance, but 8 points in eFG% is about 3 standard deviations. A team that shoots 8 points higher than another team would gain about ~24 wins. Huge difference, right? Doesn't this mean an assist should be worth quite a bit? (Or am I missing something?)
The confounding issue is the consistency in assists being assigned. It's subjective, so it's hard to say how much of an effect they have. In general, it probably is beneficial to emphasize "good passing" or "smart passing". But whether or not this leads to an assist getting recorded, I don't think we can say for sure.
Is the inconsistency systematic or random (between players, teams, home/away)? Can you point me to a study that addressed this? thanks.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
greyberger
Joined: 27 Sep 2010
Posts: 34
PostPosted: Thu Dec 30, 2010 3:57 pm Post subject: Reply with quote
Quote:
The reason I bring it up is that I've been having an internal debate about whether to assign different weights to an assist depending on shot location (inside vs. mid-range vs. 3-pt) and also quality of shooter (i.e. more valuable assists could be considered those that go to better shooter). But each time I've thought about this, I sort of come to the same conclusion that you seemed to previously - i.e. that the value is already in the numbers. The good point guard gets the ball to open players who are in a position to take good shots. Ideally, the good passer causes an improvement in the eFG% of the shooter.
Scott Sereday took on what (I think) is the same conundrum, distinguishing between types of assists in a linear SPM model. His articles on the reasoning behind it are on http://www.basketball-analysis.com/page/2/
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 199
PostPosted: Thu Dec 30, 2010 4:07 pm Post subject: Reply with quote
greyberger wrote:
Scott Sereday took on what (I think) is the same conundrum, distinguishing between types of assists in a linear SPM model. His articles on the reasoning behind it are on http://www.basketball-analysis.com/page/2/
Thanks, I need to sit down and read that carefully.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Re: More recovered assist threads
Author Message
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Sun Jan 22, 2006 1:56 pm Post subject: Immediate shots created, aka "The Steve Nash measure&qu Reply with quote
Longtime reader, first-time poster.
One perplexing element of basketball statistics is "the Steve Nash factor." Any time I watch the Suns (which is infrequently as I live in the Upper Midwest) I am convinced that Steve Nash is making a beneficial contribution. However I'm also a believer in the numerical evaluation of players, and by most measures - plus/minus, Roland rating, PER - Steve Nash is just not that good a player. Definitely above average but far from the best player on his own team. (This could also be called "the Jason Terry factor," another player who appears to be contributing mightily but has good-but-not-great numbers)
I just looked on 82games and it seems that in the past month of the season Nash has joined the elite players statistics-wise (80% win%, 162 plus/minus). But still, the statistics seem to be missing something about his game. What I'm looking for, which is probably virtually impossible to calculate with any existing statistics, is something like this:
"Percentage of a player's passes that lead directly to a shot"
Nash is a leader in assists, but I wonder if he creates more wide-open looks than other PGs. That is, does he pass in such a way that players have more open shots and are more likely to shoot directly off a Nash feed than a pass from another player?
We would have to assume that a high percentage is a positive element - that facilitating your teammates shooting immediately is good thing.
I wonder if people think this would be a good measure, or whether other statistics already account for this.
The problem with this metric is that it's dependent on who's on the floor with you, and Phoenix has a lot of gunners. Pass the ball to Eddie House or Leandro Barbosa and there's a pretty good chance he shoots it right away. Teams with fewer effective scorers, like Charlotte, would be low in this category, because when you pass it to Jake Voskuhl he probably won't do much with it.
This is probably further problematized by the fact that the Suns made their off-season moves to enhance Nash's strengths, which means getting players like Eddie House who pull the trigger quickly (although to be fair Kurt Thomas and Brian Grant were also picked up).
Side note:
Another interesting element of Nash is that he seems to be improving most quickly in the later stages of his career. On basketball-reference, his similarity scores graduate from Lee Mayberry and Derek Fisher (age 24), to Mark Price and Terry Porter (age 26), and now to esteemed company like Isiah Thomas, John Stockton and Kevin Johnson (age 30).
http://www.basketball-reference.com/pla ... l#advanced
Stockton and Thomas were good much earlier in their careers - at age 25 Stockton was averaging 15 points, 14 assists, 3 steals a game; Thomas had 21 points, 10 assists, 2 steals a game; Nash was 26 before he even scored double-digits and it wasn't until last year that he averaged double-digit assists.
I've heard that players peak around 30, but is Nash's pattern of a slow start with gradual ascension to all-star status extremely unusual?
Apologies if any of this has already been discussed on the board.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3532
Location: Hendersonville, NC
PostPosted: Fri Jan 27, 2006 7:43 am Post subject: Reply with quote
Some pretty good questions here, which deserve reply.
We've discussed 'quality of potential assists' before. I tend to think your assist rate is pretty much the sum of all passes X their potential; that is, passes to high-% shooters get you more assists. Both as % of receptions that become a shot, and as % of shots that are made.
But I am leery of segregating out % of passes which become FGA. Some 'bad' passers just won't give up the ball except (a) to shoot, or (b) to get an assist. In other words, they won't make the safe, easy pass to a teammate who is in a better overall position (to shoot or to pass). In fact, some notorious 'ball hogs' have this tendency.
Nash has both quantity and quality in his 'pass selection'; thus, he leads the league in assists. That may be the stat you're looking for.
As for his late development: I believe he was off to a good start in his 2nd year (Phx); then regressed for 2 full years in Dallas. Suddenly blooming into stardom in year 5; then allstar-level in year 7. Where he remains.
There's no real typical time track for development of players. Other PG who peaked late include Lenny Wilkens, Magic Johnson, Payton, Cassell, Kidd; how about Billups?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Fri Jan 27, 2006 11:28 am Post subject: Reply with quote
Mike G wrote:
But I am leery of segregating out % of passes which become FGA. Some 'bad' passers just won't give up the ball except (a) to shoot, or (b) to get an assist. In other words, they won't make the safe, easy pass to a teammate who is in a better overall position (to shoot or to pass). In fact, some notorious 'ball hogs' have this tendency.
Shouldn't that be captured in the statistic though?
For scoring, it's important to look at per-minute stats and scoring efficiency, but also to look at points/game, right?
Back to top
View user's profile Send private message Send e-mail AIM Address
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Sun Jan 29, 2006 1:04 pm Post subject: Maybe I want "potential assists"? Reply with quote
Thanks for your replies. What I was looking for with this idea is something about the "value-added" of Steve Nash besides the great assists numbers. More than other point guards, he's creating great shots (I think) even when people are missing those shots.
Maybe what I want is something like:
"Offense-independent assist rate" or
"Potential assists"
That is, a metric that somehow adjusts for the shooting percentage of your teammates by calculating potential assists.
I hypothesize that Nash creates at least 20-25 potential assists per game with his passing. How many he actually records is dependent on the shooting of his teammates, what kind of defense they're playing against, and who's in the lineup. But even when PHX is losing, I think he's still creating those extra shots.
This is all speculation. To get the data, you'd need to look at the 82games charting. It may be that Nash is not creating any more "potential assists" than Kidd, TJ Ford, Jason Terry.
As far as the career development, I suppose their is no typical arc. Billups is obviously another example of graduation from decent starter to all-star late in his career.
Back to top
View user's profile Send private message Visit poster's website
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Sun Jan 22, 2006 1:56 pm Post subject: Immediate shots created, aka "The Steve Nash measure&qu Reply with quote
Longtime reader, first-time poster.
One perplexing element of basketball statistics is "the Steve Nash factor." Any time I watch the Suns (which is infrequently as I live in the Upper Midwest) I am convinced that Steve Nash is making a beneficial contribution. However I'm also a believer in the numerical evaluation of players, and by most measures - plus/minus, Roland rating, PER - Steve Nash is just not that good a player. Definitely above average but far from the best player on his own team. (This could also be called "the Jason Terry factor," another player who appears to be contributing mightily but has good-but-not-great numbers)
I just looked on 82games and it seems that in the past month of the season Nash has joined the elite players statistics-wise (80% win%, 162 plus/minus). But still, the statistics seem to be missing something about his game. What I'm looking for, which is probably virtually impossible to calculate with any existing statistics, is something like this:
"Percentage of a player's passes that lead directly to a shot"
Nash is a leader in assists, but I wonder if he creates more wide-open looks than other PGs. That is, does he pass in such a way that players have more open shots and are more likely to shoot directly off a Nash feed than a pass from another player?
We would have to assume that a high percentage is a positive element - that facilitating your teammates shooting immediately is good thing.
I wonder if people think this would be a good measure, or whether other statistics already account for this.
The problem with this metric is that it's dependent on who's on the floor with you, and Phoenix has a lot of gunners. Pass the ball to Eddie House or Leandro Barbosa and there's a pretty good chance he shoots it right away. Teams with fewer effective scorers, like Charlotte, would be low in this category, because when you pass it to Jake Voskuhl he probably won't do much with it.
This is probably further problematized by the fact that the Suns made their off-season moves to enhance Nash's strengths, which means getting players like Eddie House who pull the trigger quickly (although to be fair Kurt Thomas and Brian Grant were also picked up).
Side note:
Another interesting element of Nash is that he seems to be improving most quickly in the later stages of his career. On basketball-reference, his similarity scores graduate from Lee Mayberry and Derek Fisher (age 24), to Mark Price and Terry Porter (age 26), and now to esteemed company like Isiah Thomas, John Stockton and Kevin Johnson (age 30).
http://www.basketball-reference.com/pla ... l#advanced
Stockton and Thomas were good much earlier in their careers - at age 25 Stockton was averaging 15 points, 14 assists, 3 steals a game; Thomas had 21 points, 10 assists, 2 steals a game; Nash was 26 before he even scored double-digits and it wasn't until last year that he averaged double-digit assists.
I've heard that players peak around 30, but is Nash's pattern of a slow start with gradual ascension to all-star status extremely unusual?
Apologies if any of this has already been discussed on the board.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3532
Location: Hendersonville, NC
PostPosted: Fri Jan 27, 2006 7:43 am Post subject: Reply with quote
Some pretty good questions here, which deserve reply.
We've discussed 'quality of potential assists' before. I tend to think your assist rate is pretty much the sum of all passes X their potential; that is, passes to high-% shooters get you more assists. Both as % of receptions that become a shot, and as % of shots that are made.
But I am leery of segregating out % of passes which become FGA. Some 'bad' passers just won't give up the ball except (a) to shoot, or (b) to get an assist. In other words, they won't make the safe, easy pass to a teammate who is in a better overall position (to shoot or to pass). In fact, some notorious 'ball hogs' have this tendency.
Nash has both quantity and quality in his 'pass selection'; thus, he leads the league in assists. That may be the stat you're looking for.
As for his late development: I believe he was off to a good start in his 2nd year (Phx); then regressed for 2 full years in Dallas. Suddenly blooming into stardom in year 5; then allstar-level in year 7. Where he remains.
There's no real typical time track for development of players. Other PG who peaked late include Lenny Wilkens, Magic Johnson, Payton, Cassell, Kidd; how about Billups?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Fri Jan 27, 2006 11:28 am Post subject: Reply with quote
Mike G wrote:
But I am leery of segregating out % of passes which become FGA. Some 'bad' passers just won't give up the ball except (a) to shoot, or (b) to get an assist. In other words, they won't make the safe, easy pass to a teammate who is in a better overall position (to shoot or to pass). In fact, some notorious 'ball hogs' have this tendency.
Shouldn't that be captured in the statistic though?
For scoring, it's important to look at per-minute stats and scoring efficiency, but also to look at points/game, right?
Back to top
View user's profile Send private message Send e-mail AIM Address
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Sun Jan 29, 2006 1:04 pm Post subject: Maybe I want "potential assists"? Reply with quote
Thanks for your replies. What I was looking for with this idea is something about the "value-added" of Steve Nash besides the great assists numbers. More than other point guards, he's creating great shots (I think) even when people are missing those shots.
Maybe what I want is something like:
"Offense-independent assist rate" or
"Potential assists"
That is, a metric that somehow adjusts for the shooting percentage of your teammates by calculating potential assists.
I hypothesize that Nash creates at least 20-25 potential assists per game with his passing. How many he actually records is dependent on the shooting of his teammates, what kind of defense they're playing against, and who's in the lineup. But even when PHX is losing, I think he's still creating those extra shots.
This is all speculation. To get the data, you'd need to look at the 82games charting. It may be that Nash is not creating any more "potential assists" than Kidd, TJ Ford, Jason Terry.
As far as the career development, I suppose their is no typical arc. Billups is obviously another example of graduation from decent starter to all-star late in his career.
Back to top
View user's profile Send private message Visit poster's website
Re: More recovered assist threads
Author Message
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Thu Dec 16, 2010 9:33 am Post subject: Are assists on 3-pt FG worth more than assists for 2pt FG ? Reply with quote
While this could open a lot of tangential debate about the relative value of various assists, is the distinction worthwhile?
If a 3-pointer is 1.5 times as valuable as a 2, isn't the assist also worth 1.5 times as much?
Since and-1 FG are much more likely on a 2FGA, that diminishes the difference somewhat.
And this begs the question: should 'assisted' FT opportunities be counted?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 303
PostPosted: Thu Dec 16, 2010 9:37 am Post subject: Reply with quote
In my own model, I value it 2x, but I could be convinced otherwise. Maybe the assist should be worth the same, and the rest of the team could get some of the leftover credit.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Dec 16, 2010 9:49 am Post subject: Reply with quote
The definitive work on assists and questions of this sort is found at 82games.com: http://www.82games.com/assisted.htm
A repost of their table:
Code:
Measuring FG% from a "potential assist" pass versus being unassisted
Potential Not Difference Pot.Ast%
Shot Type Assist Assisted in FG% of Att
3-Pointers 37.9% 34.2% 3.7% 81.0%
2-Point Jumpers 45.8% 36.3% 9.5% 52.0%
Close Shots 61.3% 48.7% 12.6% 43.0%
Dunks 91.0% 84.0% 7.0% 76.0%
All Shots (excluding Tips) 50.2% 42.1% 8.1% 56.0%
I would say that assists to 3-pointers are probably worth less than most assists to 2-pointers, judging by this table. It just happens that most 3-pointers are spot-ups; spot-ups almost always have an assist recorded if the shot is made, even if there is nothing special to the assist.
In another thread, I attempted to account for whether a specific shot would have even been available without the pass. http://sonicscentral.com/apbrmetrics/vi ... 1065#31065
_________________
GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Back to top
View user's profile Send private message Send e-mail Visit poster's website
EvanZ
Joined: 22 Nov 2010
Posts: 303
PostPosted: Thu Dec 16, 2010 10:10 am Post subject: Reply with quote
But there are players that create their own shot and can still make 3's at a good rate, versus catch-and-shoot types that can't create their own shots. I think we would need to separate out the %'s for those two types of players and make the comparison. If that makes sense to you guys...
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Thu Dec 16, 2010 10:15 am Post subject: Reply with quote
We should distinguish between actual, counted Assists, and those passes which will be counted Assists IF the guy makes the shot.
If you make 10 passes to shooters with a 40% shot, you'll get about 4 assists.
If you complete 10 passes to 80% range, you're apt to get 8 assists.
So the improvement in FG% created by your passes creates more FG, therefore more assists. It's built in.
The 10 passes to close range create twice as many assists, but perhaps not twice as many points. If you get 4 successful 3-pointers among those 40% shots, those 4 assists led to 12 points.
With the better likelihood of a FT tacked on to the close (assisted) shot, there's also a greater chance for a turnover. EvanZ's 2.0 factor may be high, and 1.5 may be about right.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Dec 16, 2010 11:05 am Post subject: Reply with quote
We also need to consider--for some of the assists, a similar quality of shot could be found without the assist. For others (a D-Howard dunk) the shot was created by the good positioning of the player and the good pass.
_________________
GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Back to top
View user's profile Send private message Send e-mail Visit poster's website
EvanZ
Joined: 22 Nov 2010
Posts: 303
PostPosted: Thu Dec 16, 2010 11:05 am Post subject: Re: Are assists on 3-pt FG worth more than assists for 2pt F Reply with quote
Mike G wrote:
While this could open a lot of tangential debate about the relative value of various assists, is the distinction worthwhile?
If a 3-pointer is 1.5 times as valuable as a 2, isn't the assist also worth 1.5 times as much?
I read this again. If the average output of a possession is ~1.0 PPP, then isn't a 3-PT field goal worth +2 marginal points, which would be 2 times a 2-pt FG (+1 marginal points).
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Thu Dec 23, 2010 1:00 pm Post subject: Reply with quote
If you got 1 point for a missed FGA, then yes, a 3 is twice as valuable as a 2.
Here's what I've done: Each player's Ast rate is additionally factored by:
((TeamPts - PlayerPts)/(TeamFG - PlayerFG)*(LeaguePts/LeagueFG))^N
... and N is to be determined.
(Currently N = 1.33 creates best fit between eWins and pythagorean.)
This factor gives an Ast boost for creating more points via FT as well as by 3FG.
Ast/FG is easily available for each team, and of course Ast are only granted for FG made. But other points accrue, and are just as much assisted. Namely, from FT and 3's.
Getting the fewest Pts/FG are the Celtics and Grizzlies, at 2.58
The NBA average is 2.70
At the top in this ratio are Den 2.89 and SAS 2.87.
I'll propose that the average Nuggets assist leads to (289/258) 1.12 times as many points as the avg Celtic assist.
Of course, this varies by the player. Rajon Rondo does not make 3's (nor FT, really), so even though his team is last in Pts/FG -- just 95.5% of league avg -- his assists are still bringing 97% of league-avg points.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Fri Dec 24, 2010 10:18 am Post subject: Reply with quote
Using the adjustment described above, here are players whose assist rates are most affected by the factor.
Ast1 is N = 0, and Ast2 is N = 1.34.
Code:
delta per36 rates tm Ast1 Ast2 delta per36 rates tm Ast1 Ast2
.68 Lawson,Ty Den 6.1 6.8 -.48 Williams,Deron Uta 9.7 9.2
.62 Westbrook,Russel Okl 8.3 8.9 -.46 Conley,Mike Mem 7.1 6.6
.52 Parker,Tony SAS 8.5 9.0 -.46 Rondo,Rajon Bos 14.2 13.7
.50 Maynor,Eric Okl 5.2 5.7 -.36 Williams,Lou Phi 5.5 5.1
.38 Felton,Raymond NYK 8.6 9.0 -.36 Vasquez,Greivis Mem 5.9 5.5
.36 Smith,Ishmael Hou 7.0 7.3 -.34 Rose,Derrick Chi 8.4 8.1
.33 Anthony,Carmelo Den 3.4 3.7 -.33 Pierce,Paul Bos 3.8 3.5
.28 Lowry,Kyle Hou 7.9 8.2 -.32 Curry,Stephen GSW 5.7 5.4
.27 Hilario,Nene Den 2.7 3.0 -.32 Robinson,Nate Bos 5.2 4.9
.26 Afflalo,Arron Den 2.5 2.8 -.32 Stuckey,Rodney Det 5.7 5.4
Gainers (on the left) play for teams with high Pts/FG and/or don't make many of their teams' FT and 3FG.
On the right, players from teams with lower Pts/FG and/or who have relatively high Pts/FG.
_________________
`
36% of all statistics are wrong
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Thu Dec 16, 2010 9:33 am Post subject: Are assists on 3-pt FG worth more than assists for 2pt FG ? Reply with quote
While this could open a lot of tangential debate about the relative value of various assists, is the distinction worthwhile?
If a 3-pointer is 1.5 times as valuable as a 2, isn't the assist also worth 1.5 times as much?
Since and-1 FG are much more likely on a 2FGA, that diminishes the difference somewhat.
And this begs the question: should 'assisted' FT opportunities be counted?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 303
PostPosted: Thu Dec 16, 2010 9:37 am Post subject: Reply with quote
In my own model, I value it 2x, but I could be convinced otherwise. Maybe the assist should be worth the same, and the rest of the team could get some of the leftover credit.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Dec 16, 2010 9:49 am Post subject: Reply with quote
The definitive work on assists and questions of this sort is found at 82games.com: http://www.82games.com/assisted.htm
A repost of their table:
Code:
Measuring FG% from a "potential assist" pass versus being unassisted
Potential Not Difference Pot.Ast%
Shot Type Assist Assisted in FG% of Att
3-Pointers 37.9% 34.2% 3.7% 81.0%
2-Point Jumpers 45.8% 36.3% 9.5% 52.0%
Close Shots 61.3% 48.7% 12.6% 43.0%
Dunks 91.0% 84.0% 7.0% 76.0%
All Shots (excluding Tips) 50.2% 42.1% 8.1% 56.0%
I would say that assists to 3-pointers are probably worth less than most assists to 2-pointers, judging by this table. It just happens that most 3-pointers are spot-ups; spot-ups almost always have an assist recorded if the shot is made, even if there is nothing special to the assist.
In another thread, I attempted to account for whether a specific shot would have even been available without the pass. http://sonicscentral.com/apbrmetrics/vi ... 1065#31065
_________________
GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Back to top
View user's profile Send private message Send e-mail Visit poster's website
EvanZ
Joined: 22 Nov 2010
Posts: 303
PostPosted: Thu Dec 16, 2010 10:10 am Post subject: Reply with quote
But there are players that create their own shot and can still make 3's at a good rate, versus catch-and-shoot types that can't create their own shots. I think we would need to separate out the %'s for those two types of players and make the comparison. If that makes sense to you guys...
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Thu Dec 16, 2010 10:15 am Post subject: Reply with quote
We should distinguish between actual, counted Assists, and those passes which will be counted Assists IF the guy makes the shot.
If you make 10 passes to shooters with a 40% shot, you'll get about 4 assists.
If you complete 10 passes to 80% range, you're apt to get 8 assists.
So the improvement in FG% created by your passes creates more FG, therefore more assists. It's built in.
The 10 passes to close range create twice as many assists, but perhaps not twice as many points. If you get 4 successful 3-pointers among those 40% shots, those 4 assists led to 12 points.
With the better likelihood of a FT tacked on to the close (assisted) shot, there's also a greater chance for a turnover. EvanZ's 2.0 factor may be high, and 1.5 may be about right.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 611
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Dec 16, 2010 11:05 am Post subject: Reply with quote
We also need to consider--for some of the assists, a similar quality of shot could be found without the assist. For others (a D-Howard dunk) the shot was created by the good positioning of the player and the good pass.
_________________
GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Back to top
View user's profile Send private message Send e-mail Visit poster's website
EvanZ
Joined: 22 Nov 2010
Posts: 303
PostPosted: Thu Dec 16, 2010 11:05 am Post subject: Re: Are assists on 3-pt FG worth more than assists for 2pt F Reply with quote
Mike G wrote:
While this could open a lot of tangential debate about the relative value of various assists, is the distinction worthwhile?
If a 3-pointer is 1.5 times as valuable as a 2, isn't the assist also worth 1.5 times as much?
I read this again. If the average output of a possession is ~1.0 PPP, then isn't a 3-PT field goal worth +2 marginal points, which would be 2 times a 2-pt FG (+1 marginal points).
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Thu Dec 23, 2010 1:00 pm Post subject: Reply with quote
If you got 1 point for a missed FGA, then yes, a 3 is twice as valuable as a 2.
Here's what I've done: Each player's Ast rate is additionally factored by:
((TeamPts - PlayerPts)/(TeamFG - PlayerFG)*(LeaguePts/LeagueFG))^N
... and N is to be determined.
(Currently N = 1.33 creates best fit between eWins and pythagorean.)
This factor gives an Ast boost for creating more points via FT as well as by 3FG.
Ast/FG is easily available for each team, and of course Ast are only granted for FG made. But other points accrue, and are just as much assisted. Namely, from FT and 3's.
Getting the fewest Pts/FG are the Celtics and Grizzlies, at 2.58
The NBA average is 2.70
At the top in this ratio are Den 2.89 and SAS 2.87.
I'll propose that the average Nuggets assist leads to (289/258) 1.12 times as many points as the avg Celtic assist.
Of course, this varies by the player. Rajon Rondo does not make 3's (nor FT, really), so even though his team is last in Pts/FG -- just 95.5% of league avg -- his assists are still bringing 97% of league-avg points.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3618
Location: Hendersonville, NC
PostPosted: Fri Dec 24, 2010 10:18 am Post subject: Reply with quote
Using the adjustment described above, here are players whose assist rates are most affected by the factor.
Ast1 is N = 0, and Ast2 is N = 1.34.
Code:
delta per36 rates tm Ast1 Ast2 delta per36 rates tm Ast1 Ast2
.68 Lawson,Ty Den 6.1 6.8 -.48 Williams,Deron Uta 9.7 9.2
.62 Westbrook,Russel Okl 8.3 8.9 -.46 Conley,Mike Mem 7.1 6.6
.52 Parker,Tony SAS 8.5 9.0 -.46 Rondo,Rajon Bos 14.2 13.7
.50 Maynor,Eric Okl 5.2 5.7 -.36 Williams,Lou Phi 5.5 5.1
.38 Felton,Raymond NYK 8.6 9.0 -.36 Vasquez,Greivis Mem 5.9 5.5
.36 Smith,Ishmael Hou 7.0 7.3 -.34 Rose,Derrick Chi 8.4 8.1
.33 Anthony,Carmelo Den 3.4 3.7 -.33 Pierce,Paul Bos 3.8 3.5
.28 Lowry,Kyle Hou 7.9 8.2 -.32 Curry,Stephen GSW 5.7 5.4
.27 Hilario,Nene Den 2.7 3.0 -.32 Robinson,Nate Bos 5.2 4.9
.26 Afflalo,Arron Den 2.5 2.8 -.32 Stuckey,Rodney Det 5.7 5.4
Gainers (on the left) play for teams with high Pts/FG and/or don't make many of their teams' FT and 3FG.
On the right, players from teams with lower Pts/FG and/or who have relatively high Pts/FG.
_________________
`
36% of all statistics are wrong
Re: More recovered assist threads
rlee
Joined: 20 Aug 2005
Posts: 62
PostPosted: Mon Oct 22, 2007 10:20 pm Post subject: Is it time to add second NBA assist? Reply with quote
Dave Feschuk
Toronto Star
Quote:
"while the club doesn't make those numbers public, web-savvy enthusiasts such as the video excavators at 82games.com, provide the rest of us no end of interesting fodder not found in the regular stats package – not only, for instance, the team that committed the most charges last season (the ball-hogging New York Knicks), but the team that took the most (the Phoenix Suns).
Still, there is always room for more information. Sam Mitchell, the Raptors coach, said he was ranting to his players just the other day about the need for another stat. Call it the hockey assist, as Mike Brown, the Cleveland coach, recently dubbed it."
http://www.thestar.com/printArticle/268885
Back to top
View user's profile Send private message Send e-mail
BorisD
Joined: 29 Jul 2007
Posts: 14
PostPosted: Tue Oct 23, 2007 1:05 am Post subject: Reply with quote
Personally, no, I don't think the notion of a second assist adds any understanding of value in terms of analytics. For one, there is a low degree of difficulty in many second assists. In most instances I daresay those passes are uncontested. I know that probably every team has a guy (usually a big) who has a limited offensive game and is used simply to reverse the ball when the offense is stymied on the strongside. To give them an assist stat for just swinging the ball when they really don't add much to the offense seems rather pointless.
I think this is a noble attempt to quantify something important and meaningful on the floor, but is otherwise unquantifiable. Same thing with keeping track of number of screens set: how do you reward a guy statistically for doing what an offense calls him to do by rote? Do you punish him for setting an ill-advised ball screen that clogs up spacing on the floor so a guy with a mismatch can no longer drive? And who keeps track of this, because it all seems so subjective and situational that it would likely defy consistent monitoring.
And, for example, do we say Rasho Nesterovic is a better team offensive player than (for instance) Brad Miller or Andrew Bynum because he would likely set more screens per possession/minute/whatever? Sounds good, until you realize that Rasho runs a lot in a flex offense which calls for a lot of screens, and Miller runs in a Princeton offense which generally seeks to maintain spacing and goes away from pressure with fewer screens, or Bynum who plays in a Triple-Post Offense which calls for very little screening from its centers?
Now, assisted FTA...I'm all for that.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3613
Location: Hendersonville, NC
PostPosted: Tue Oct 23, 2007 4:10 am Post subject: Reply with quote
When Bill Russell was calling games (in the '80s) he liked to assert that not only do easy passes tend to produce easy baskets, but that low-% shots should be rare events. Of course, he liked the Bird-DJ Celtics -- a team without a true PG, and lots of assists from lots of guys.
A team could track 'hockey assists' and not share them with the league (or the general public). This would give them a competitive advantage; for example, 'knowing' that TJ Ford's assists were accompanied by too-few h-Ast, trade him for a less-selfish player who doesn't have the market value commensurate with his 'conventional' stats.
I don't think 'degree of difficulty' has much to do with value. Like Bill said, keep it simple. The center who reverses the ball can do it well or not so well. There are many times when the next-to-last pass is the great one, and the assist is very simple.
When playing, the ballhog who only passes when there's an assist to be had is a worse teammate than the one who always shoots. Just giving up the ball to someone in better position -- a basic offensive move if there ever was one -- occasionally gets you an h-Ast; if anyone's counting.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
ziller
Joined: 30 Jun 2005
Posts: 126
Location: Sac Metro
PostPosted: Tue Oct 23, 2007 11:36 am Post subject: Reply with quote
It's nice to think about, but we need to track potential assists and FTA-assists before we consider getting deep into 'hockey assists.' As Boris indicates, 'hockey assists' are about as valuable as screens, perhaps less so in some offenses.
_________________
SactownRoyalty.com
tziller@gmail.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Tue Oct 23, 2007 11:59 am Post subject: Reply with quote
As a first step with assists, I think the league should address awarding them properly. No preferential treatment for stars and PGs. Uniformly applied criteria.
I like the idea of tracking "potential assists" and FT assists. But I want to see the awarding of assists fixed.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
BorisD
Joined: 29 Jul 2007
Posts: 14
PostPosted: Tue Oct 23, 2007 12:55 pm Post subject: Reply with quote
Mike G wrote:
When Bill Russell was calling games (in the '80s) he liked to assert that not only do easy passes tend to produce easy baskets, but that low-% shots should be rare events. Of course, he liked the Bird-DJ Celtics -- a team without a true PG, and lots of assists from lots of guys.
I definitely agree with Russell's philosophy in that regard, and some systems are designed specifically with avoiding pressure in mind (the Princeton offense is a great example) so it's encouraged to make the uncontested pass vs. the contested pass. However, often some plays/passes are only uncontested because that's what the defense wants you to do. If I'm playing the 1998 Utah Jazz, do I want the ball in the hands of Karl Malone on the low block, John Stockton up top, Jeff Hornacek on the wing, or Greg Ostertag at the top of the 3 point arc? Well, I'm going to let that pass go to Ostertag outside of his 4' range 100 times out of 100, with a margin of error of zero.
This is where the idea of "hockey assists" gets dicey IMO from a basketball standpoint (keep in mind the highest level I've coached at is high school, so I'm far from an expert): is a "hockey assist" coming from a guy doing something very important (i.e. quickly initiating a weak-side attack) or he is basically just offensive spare parts for 90% of the possession and just kind of tossing the ball laterally because that's all he can do? In other words, the offense plays 4-on-5 while he's on the floor, so all he does is reverse the ball? Or is that function perhaps more valuable than I'm giving it credit for? I honestly don't know.
As a Raptors fan, I have pictures of Jorge Garbajosa flashing through my mind, who has no post game, and his offensive game largely consists of hovering either in the corner or at the top of the arc (depending upon the set) and throwing either reversal passes or low percentage post entry passes from 22' that more often than not end up in the opponent's hands, or the fourth row. Does he deserve credit for that?
And, being a Canadian, I've had to watch a lot of hockey in my day. So when I hear "hockey assists", you might get the same situation in basketball as you do in hockey. The second assist is highly controversial, because sometimes it doesn't communicate much. Sometimes the second assist is given on a dump-and-chase, where a defenceman will dump the puck from the redline deep into the offensive zone (sometimes to force a line change), and skate to the bench. Some enterprising forward will crash hard into the corner, fight off a defender, bring out the puck under pressure, put a centering pace right on the tape of a forward on the far side for an easy open-net goal. Yet the defenceman who is now celebrating on the bench gets equal credit (1 point) with the forward who did the bulk of the work and the forward who banged home the easy goal. A lot of hockey fans always gripe about the second assist making very poor playmakers look better than they are and decent playmakers looking very pedestrian.
I have no problem with a second assist when a guy makes a pass that makes something happen, but hockey attributes credit just for touching the puck - including, oddly, on situations where a rebound(s) are scored on. So I could dump the puck into the offensive zone again from the redline, you get the puck out from the corner, and let's say mountain could be open on the far side, but you elect to shoot towards the far side because you think the goalie is anticipating the pass and not the shot. The goalie stops the shot, but the rebound falls to mountain, who scores. In that scenario, both you and I get assists, even though you didn't pass and I'm on the bench after mindlessly throwing the puck into the corner. Was I really a playmaker on that play?
I have no problem with tracking the hAst - as long as the pass actually accomplishes something and doesn't reward doing little or nothing.
Quote:
A team could track 'hockey assists' and not share them with the league (or the general public). This would give them a competitive advantage; for example, 'knowing' that TJ Ford's assists were accompanied by too-few h-Ast, trade him for a less-selfish player who doesn't have the market value commensurate with his 'conventional' stats.
Okay, but by the same token, how many hAst do you credit on a play? Using TJ Ford as an example, and a fairly common one in Toronto, they'll run a 1-2-2 Horns offense where he'll drive off the high ball screen and down into the paint. The defense will collapse, he'll pass the ball to the player in the strongside corner, who (if it wasn't Morris Peterson) will pass to the big who is now on the right wing, who will pass to the top, who will pass to the weakside big, who will pass to the weakside wing player, who shoots a 3. Which of those players get the hAst?
Quote:
I don't think 'degree of difficulty' has much to do with value. Like Bill said, keep it simple. The center who reverses the ball can do it well or not so well. There are many times when the next-to-last pass is the great one, and the assist is very simple.
When playing, the ballhog who only passes when there's an assist to be had is a worse teammate than the one who always shoots. Just giving up the ball to someone in better position -- a basic offensive move if there ever was one -- occasionally gets you an h-Ast; if anyone's counting.
I agree in principle, I just think that keeping this as a metric would be very complicated and communicate little.
In my opinion, there's a lot of aspects of teamwork that will forever defy quantification. Do you give assists to the big man who clears out of the post to allow a wing player with a severe mismatch to drive to the basket? In my opinion, that's no less important and no less selfless than a direct pass, but I don't see that being quantified anytime soon. Every coach in the world will hug the guy when he gets to the bench, but that's not necessarily something that can be directly measured.
Back to top
View user's profile Send private message
tsherkin
Joined: 31 Jan 2005
Posts: 247
PostPosted: Tue Oct 23, 2007 1:35 pm Post subject: Reply with quote
Potential assists would be fantastic if you could use it to produce something like % of realized assists or whatever, to see how many opportunities were available to which point guards and how effectively they took advantage of those opportunities.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3613
Location: Hendersonville, NC
PostPosted: Tue Oct 23, 2007 3:04 pm Post subject: Reply with quote
BorisD wrote:
...passes from 22' that more often than not end up in the opponent's hands, or the fourth row. Does he deserve credit for that?
.
Definitely. That is a turnover.
Quote:
...he'll pass the ball to the player in the strongside corner, who ... will pass to the big who is now on the right wing, who will pass to the top, who will pass to the weakside big, who will pass to the weakside wing player, who shoots a 3. Which of those players get the hAst?
The next-to-last pass before the made FG. This can't be the sequence for every play, or the opponent would intervene. Other times, the shot comes earlier in the sequence, and someone else gets the assist(s).
The whole idea (I think) is that almost any pass may turn into an h-Ast. Therefore, make that practical pass.
Quote:
...there's a lot of aspects of teamwork that will forever defy quantification...
But that shouldn't make anyone reject an obvious avenue towards quantifying something that everyone agrees is positive. Maybe part of the problem with assists being awarded so subjectively is that scorekeepers are forever trying to make up for nice passes that weren't assists. With the h-Ast, many of them will be recorded and credited.
Assists don't fully measure passing acuity. H-Ast + Ast (however weighted) would just be a better measure.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Tue Oct 23, 2007 4:11 pm Post subject: Reply with quote
What about boxing out, or HRebound?, Deflections or HSteals?
I think is a very good thing for the game, maybe not so important to metrics, the track and weight of every kind of indirect assists. It's already hard to rightly boxscore an assist, and adjusting methods does exist.
Back to top
View user's profile Send private message
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Tue Oct 23, 2007 9:08 pm Post subject: Reply with quote
When it comes to weight sharing stats, it's used to credit 2/3 (66%) to the possession end accomplisher, and 1/3 (33%) to the assistant (although PER just gives 1/3 to DReb). If one or several new potential assistant-s were introduced in calcs, and some people talks about a weighting of about an 8% for them (at least at passing); Who must be cut on credits, the finisher or the direct assistant, or both?
Back to top
View user's profile Send private message
Charles
Joined: 16 May 2005
Posts: 134
PostPosted: Tue Oct 23, 2007 9:49 pm Post subject: Reply with quote
I see no reason not to collect new stats as long as the score-keeper is not taxed to the point that accuracy is compromised. If you can find a legitimate way to use second assists to understand the game better, great. If not, no real harm done. And as Sam says it might encourage some players to move the ball more even when no direct assist is likely.
As for PER, I believe it is currently, one third to the scorer for getting open, one-third to the scorer for making the shot, one-third to the [assumed] assister for making the pass. So, I'll say 55% to the shooter, 20% to the assister, 10% for the hAssister, 5% to the screener, 5% to the guy spotted up behind the arc for spreading the floor and 5% to the coach for drawing up the play (assumed.)
I'm kidding, but the entire system is arbitrary to begin with, so there is really no reason you can't just make up whatever number you like.
Joined: 20 Aug 2005
Posts: 62
PostPosted: Mon Oct 22, 2007 10:20 pm Post subject: Is it time to add second NBA assist? Reply with quote
Dave Feschuk
Toronto Star
Quote:
"while the club doesn't make those numbers public, web-savvy enthusiasts such as the video excavators at 82games.com, provide the rest of us no end of interesting fodder not found in the regular stats package – not only, for instance, the team that committed the most charges last season (the ball-hogging New York Knicks), but the team that took the most (the Phoenix Suns).
Still, there is always room for more information. Sam Mitchell, the Raptors coach, said he was ranting to his players just the other day about the need for another stat. Call it the hockey assist, as Mike Brown, the Cleveland coach, recently dubbed it."
http://www.thestar.com/printArticle/268885
Back to top
View user's profile Send private message Send e-mail
BorisD
Joined: 29 Jul 2007
Posts: 14
PostPosted: Tue Oct 23, 2007 1:05 am Post subject: Reply with quote
Personally, no, I don't think the notion of a second assist adds any understanding of value in terms of analytics. For one, there is a low degree of difficulty in many second assists. In most instances I daresay those passes are uncontested. I know that probably every team has a guy (usually a big) who has a limited offensive game and is used simply to reverse the ball when the offense is stymied on the strongside. To give them an assist stat for just swinging the ball when they really don't add much to the offense seems rather pointless.
I think this is a noble attempt to quantify something important and meaningful on the floor, but is otherwise unquantifiable. Same thing with keeping track of number of screens set: how do you reward a guy statistically for doing what an offense calls him to do by rote? Do you punish him for setting an ill-advised ball screen that clogs up spacing on the floor so a guy with a mismatch can no longer drive? And who keeps track of this, because it all seems so subjective and situational that it would likely defy consistent monitoring.
And, for example, do we say Rasho Nesterovic is a better team offensive player than (for instance) Brad Miller or Andrew Bynum because he would likely set more screens per possession/minute/whatever? Sounds good, until you realize that Rasho runs a lot in a flex offense which calls for a lot of screens, and Miller runs in a Princeton offense which generally seeks to maintain spacing and goes away from pressure with fewer screens, or Bynum who plays in a Triple-Post Offense which calls for very little screening from its centers?
Now, assisted FTA...I'm all for that.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3613
Location: Hendersonville, NC
PostPosted: Tue Oct 23, 2007 4:10 am Post subject: Reply with quote
When Bill Russell was calling games (in the '80s) he liked to assert that not only do easy passes tend to produce easy baskets, but that low-% shots should be rare events. Of course, he liked the Bird-DJ Celtics -- a team without a true PG, and lots of assists from lots of guys.
A team could track 'hockey assists' and not share them with the league (or the general public). This would give them a competitive advantage; for example, 'knowing' that TJ Ford's assists were accompanied by too-few h-Ast, trade him for a less-selfish player who doesn't have the market value commensurate with his 'conventional' stats.
I don't think 'degree of difficulty' has much to do with value. Like Bill said, keep it simple. The center who reverses the ball can do it well or not so well. There are many times when the next-to-last pass is the great one, and the assist is very simple.
When playing, the ballhog who only passes when there's an assist to be had is a worse teammate than the one who always shoots. Just giving up the ball to someone in better position -- a basic offensive move if there ever was one -- occasionally gets you an h-Ast; if anyone's counting.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
ziller
Joined: 30 Jun 2005
Posts: 126
Location: Sac Metro
PostPosted: Tue Oct 23, 2007 11:36 am Post subject: Reply with quote
It's nice to think about, but we need to track potential assists and FTA-assists before we consider getting deep into 'hockey assists.' As Boris indicates, 'hockey assists' are about as valuable as screens, perhaps less so in some offenses.
_________________
SactownRoyalty.com
tziller@gmail.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Tue Oct 23, 2007 11:59 am Post subject: Reply with quote
As a first step with assists, I think the league should address awarding them properly. No preferential treatment for stars and PGs. Uniformly applied criteria.
I like the idea of tracking "potential assists" and FT assists. But I want to see the awarding of assists fixed.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
BorisD
Joined: 29 Jul 2007
Posts: 14
PostPosted: Tue Oct 23, 2007 12:55 pm Post subject: Reply with quote
Mike G wrote:
When Bill Russell was calling games (in the '80s) he liked to assert that not only do easy passes tend to produce easy baskets, but that low-% shots should be rare events. Of course, he liked the Bird-DJ Celtics -- a team without a true PG, and lots of assists from lots of guys.
I definitely agree with Russell's philosophy in that regard, and some systems are designed specifically with avoiding pressure in mind (the Princeton offense is a great example) so it's encouraged to make the uncontested pass vs. the contested pass. However, often some plays/passes are only uncontested because that's what the defense wants you to do. If I'm playing the 1998 Utah Jazz, do I want the ball in the hands of Karl Malone on the low block, John Stockton up top, Jeff Hornacek on the wing, or Greg Ostertag at the top of the 3 point arc? Well, I'm going to let that pass go to Ostertag outside of his 4' range 100 times out of 100, with a margin of error of zero.
This is where the idea of "hockey assists" gets dicey IMO from a basketball standpoint (keep in mind the highest level I've coached at is high school, so I'm far from an expert): is a "hockey assist" coming from a guy doing something very important (i.e. quickly initiating a weak-side attack) or he is basically just offensive spare parts for 90% of the possession and just kind of tossing the ball laterally because that's all he can do? In other words, the offense plays 4-on-5 while he's on the floor, so all he does is reverse the ball? Or is that function perhaps more valuable than I'm giving it credit for? I honestly don't know.
As a Raptors fan, I have pictures of Jorge Garbajosa flashing through my mind, who has no post game, and his offensive game largely consists of hovering either in the corner or at the top of the arc (depending upon the set) and throwing either reversal passes or low percentage post entry passes from 22' that more often than not end up in the opponent's hands, or the fourth row. Does he deserve credit for that?
And, being a Canadian, I've had to watch a lot of hockey in my day. So when I hear "hockey assists", you might get the same situation in basketball as you do in hockey. The second assist is highly controversial, because sometimes it doesn't communicate much. Sometimes the second assist is given on a dump-and-chase, where a defenceman will dump the puck from the redline deep into the offensive zone (sometimes to force a line change), and skate to the bench. Some enterprising forward will crash hard into the corner, fight off a defender, bring out the puck under pressure, put a centering pace right on the tape of a forward on the far side for an easy open-net goal. Yet the defenceman who is now celebrating on the bench gets equal credit (1 point) with the forward who did the bulk of the work and the forward who banged home the easy goal. A lot of hockey fans always gripe about the second assist making very poor playmakers look better than they are and decent playmakers looking very pedestrian.
I have no problem with a second assist when a guy makes a pass that makes something happen, but hockey attributes credit just for touching the puck - including, oddly, on situations where a rebound(s) are scored on. So I could dump the puck into the offensive zone again from the redline, you get the puck out from the corner, and let's say mountain could be open on the far side, but you elect to shoot towards the far side because you think the goalie is anticipating the pass and not the shot. The goalie stops the shot, but the rebound falls to mountain, who scores. In that scenario, both you and I get assists, even though you didn't pass and I'm on the bench after mindlessly throwing the puck into the corner. Was I really a playmaker on that play?
I have no problem with tracking the hAst - as long as the pass actually accomplishes something and doesn't reward doing little or nothing.
Quote:
A team could track 'hockey assists' and not share them with the league (or the general public). This would give them a competitive advantage; for example, 'knowing' that TJ Ford's assists were accompanied by too-few h-Ast, trade him for a less-selfish player who doesn't have the market value commensurate with his 'conventional' stats.
Okay, but by the same token, how many hAst do you credit on a play? Using TJ Ford as an example, and a fairly common one in Toronto, they'll run a 1-2-2 Horns offense where he'll drive off the high ball screen and down into the paint. The defense will collapse, he'll pass the ball to the player in the strongside corner, who (if it wasn't Morris Peterson) will pass to the big who is now on the right wing, who will pass to the top, who will pass to the weakside big, who will pass to the weakside wing player, who shoots a 3. Which of those players get the hAst?
Quote:
I don't think 'degree of difficulty' has much to do with value. Like Bill said, keep it simple. The center who reverses the ball can do it well or not so well. There are many times when the next-to-last pass is the great one, and the assist is very simple.
When playing, the ballhog who only passes when there's an assist to be had is a worse teammate than the one who always shoots. Just giving up the ball to someone in better position -- a basic offensive move if there ever was one -- occasionally gets you an h-Ast; if anyone's counting.
I agree in principle, I just think that keeping this as a metric would be very complicated and communicate little.
In my opinion, there's a lot of aspects of teamwork that will forever defy quantification. Do you give assists to the big man who clears out of the post to allow a wing player with a severe mismatch to drive to the basket? In my opinion, that's no less important and no less selfless than a direct pass, but I don't see that being quantified anytime soon. Every coach in the world will hug the guy when he gets to the bench, but that's not necessarily something that can be directly measured.
Back to top
View user's profile Send private message
tsherkin
Joined: 31 Jan 2005
Posts: 247
PostPosted: Tue Oct 23, 2007 1:35 pm Post subject: Reply with quote
Potential assists would be fantastic if you could use it to produce something like % of realized assists or whatever, to see how many opportunities were available to which point guards and how effectively they took advantage of those opportunities.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3613
Location: Hendersonville, NC
PostPosted: Tue Oct 23, 2007 3:04 pm Post subject: Reply with quote
BorisD wrote:
...passes from 22' that more often than not end up in the opponent's hands, or the fourth row. Does he deserve credit for that?
.
Definitely. That is a turnover.
Quote:
...he'll pass the ball to the player in the strongside corner, who ... will pass to the big who is now on the right wing, who will pass to the top, who will pass to the weakside big, who will pass to the weakside wing player, who shoots a 3. Which of those players get the hAst?
The next-to-last pass before the made FG. This can't be the sequence for every play, or the opponent would intervene. Other times, the shot comes earlier in the sequence, and someone else gets the assist(s).
The whole idea (I think) is that almost any pass may turn into an h-Ast. Therefore, make that practical pass.
Quote:
...there's a lot of aspects of teamwork that will forever defy quantification...
But that shouldn't make anyone reject an obvious avenue towards quantifying something that everyone agrees is positive. Maybe part of the problem with assists being awarded so subjectively is that scorekeepers are forever trying to make up for nice passes that weren't assists. With the h-Ast, many of them will be recorded and credited.
Assists don't fully measure passing acuity. H-Ast + Ast (however weighted) would just be a better measure.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Tue Oct 23, 2007 4:11 pm Post subject: Reply with quote
What about boxing out, or HRebound?, Deflections or HSteals?
I think is a very good thing for the game, maybe not so important to metrics, the track and weight of every kind of indirect assists. It's already hard to rightly boxscore an assist, and adjusting methods does exist.
Back to top
View user's profile Send private message
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Tue Oct 23, 2007 9:08 pm Post subject: Reply with quote
When it comes to weight sharing stats, it's used to credit 2/3 (66%) to the possession end accomplisher, and 1/3 (33%) to the assistant (although PER just gives 1/3 to DReb). If one or several new potential assistant-s were introduced in calcs, and some people talks about a weighting of about an 8% for them (at least at passing); Who must be cut on credits, the finisher or the direct assistant, or both?
Back to top
View user's profile Send private message
Charles
Joined: 16 May 2005
Posts: 134
PostPosted: Tue Oct 23, 2007 9:49 pm Post subject: Reply with quote
I see no reason not to collect new stats as long as the score-keeper is not taxed to the point that accuracy is compromised. If you can find a legitimate way to use second assists to understand the game better, great. If not, no real harm done. And as Sam says it might encourage some players to move the ball more even when no direct assist is likely.
As for PER, I believe it is currently, one third to the scorer for getting open, one-third to the scorer for making the shot, one-third to the [assumed] assister for making the pass. So, I'll say 55% to the shooter, 20% to the assister, 10% for the hAssister, 5% to the screener, 5% to the guy spotted up behind the arc for spreading the floor and 5% to the coach for drawing up the play (assumed.)
I'm kidding, but the entire system is arbitrary to begin with, so there is really no reason you can't just make up whatever number you like.
Re: More recovered assist threads
jaxx
Joined: 17 May 2005
Posts: 23
PostPosted: Sat Jan 07, 2006 2:59 pm Post subject: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
Since "assists" seem to be the theme of this month... What is the expert's opinion here about Hollinger's crediting of Assists in PER?
As you probably know, he sees the assisted basket as
1/3 getting open
1/3 making the pass
1/3 making the shot
Following this logic, he then attributes 2/3 of the credit to the scorer, and only 1/3 to the passer.
However, this 2/3-1/3 credit split seems very unfair towards Stephon Marbury-type players who drive, draw the double team and then dish to the open man. Or even towards guys like Shaq, who routinely draw double coverage and are adept at finding the open man on the perimeter. In both instances, the shotmaker hardly has to move his feet to free himself; he sees his defender leave, catches the pass and shoots. So following Hollinger's logic, here the PASSER 'takes care of two of the three acts of making an assisted basket', and should receive 2/3.
Watching actual games, these types of assists seem fairly common, and I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it. Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof. However, I know Hollinger is an intelligent guy who probably knows more than I do, so maybe his 1/3-2/3 decision is rooted in analysis too complex to share with the general public.
What's the APBR party line on this?
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
PostPosted: Sat Jan 07, 2006 4:27 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
jaxx wrote:
I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it.
I don't know if I agree with that. He also says in his explanation, "All assists are not equal, and one of my long-term goals is to develop better ways of measuring their differences."
Quote:
Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof.
I agree and have mentioned it before, but now that I think through the issue I think the bigger issue is not that point guards are not getting enough credit for their assists, they're not getting enough credit for not being assisted. Point guards almost always have very low percentages of their field goals that are assisted (Steve Nash, for example, has been assisted on only 22% of his field goals this year), but PER assumes his rate as the same as his entire team (66.3%). That takes away a lot of points that Nash should be getting credit for.
John has spoken about factoring this into PER v2.0.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Sat Jan 07, 2006 7:56 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
jaxx wrote:
Since "assists" seem to be the theme of this month... What is the expert's opinion here about Hollinger's crediting of Assists in PER?
As you probably know, he sees the assisted basket as
1/3 getting open
1/3 making the pass
1/3 making the shot
Following this logic, he then attributes 2/3 of the credit to the scorer, and only 1/3 to the passer.
Also by this logic, passers who pass to guys missing shots should get 1/3 credit (=blame) for the missed FGA. But there is nothing in the record showing passes on missed shots.
Assists are a pain in the butt. Argue as much as you like, but assists are so fundamentally biased that it's hard to get them right. I'm content (not happy) with my approach of giving variable credit based on estimates of the ease of shot created for teammates. But...
A better thing to do is to propose other measures of teammate cooperation. The simplest thing is potential assists - passes that resulted in a FGA. That in combination with recording assists on passes that lead to FTM/FTA would provide some balance to the assist stat.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
jaxx
Joined: 17 May 2005
Posts: 23
PostPosted: Mon Jan 09, 2006 2:56 pm Post subject: Reply with quote
Thanks for your thoughts. Probably everybody in this community has spent a lot of time thinking through assists, and I agree - they are a pain.
Still... Since you can deduce most offensive weights logically, shouldn't you be able to get the correct weight for assists by regression? (I half expect that Hollinger actually did this, and ended up with a value of 0.33.)
Quote:
I think the bigger issue is not that point guards are not getting enough credit for their assists, they're not getting enough credit for not being assisted.
Wow. I never thought about that, great point. This is also easy to solve in a linear-weights method such as PER; you can estimate the excess percentage of non-assisted PG shots and correct PG credits for that.
Back to top
View user's profile Send private message
kbche
Joined: 19 Jul 2005
Posts: 51
Location: washington d.c.
PostPosted: Mon Jan 09, 2006 11:11 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
jaxx wrote:
Since "assists" seem to be the theme of this month... What is the expert's opinion here about Hollinger's crediting of Assists in PER?
As you probably know, he sees the assisted basket as
1/3 getting open
1/3 making the pass
1/3 making the shot
Following this logic, he then attributes 2/3 of the credit to the scorer, and only 1/3 to the passer.
However, this 2/3-1/3 credit split seems very unfair towards Stephon Marbury-type players who drive, draw the double team and then dish to the open man. Or even towards guys like Shaq, who routinely draw double coverage and are adept at finding the open man on the perimeter. In both instances, the shotmaker hardly has to move his feet to free himself; he sees his defender leave, catches the pass and shoots. So following Hollinger's logic, here the PASSER 'takes care of two of the three acts of making an assisted basket', and should receive 2/3.
Watching actual games, these types of assists seem fairly common, and I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it. Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof. However, I know Hollinger is an intelligent guy who probably knows more than I do, so maybe his 1/3-2/3 decision is rooted in analysis too complex to share with the general public.
What's the APBR party line on this?
Power forwards had the highest PER values last year. However, power forwards are the highest paid, had the highest NBA efficiency, and had the most impressive stats last season. Thus, PER is in agreement with last seasons stats in that regard. Check out my blog dated 5 november (http://kabnba.blogspot.com/2005/11/what ... ition.html).
Kimberly
Back to top
View user's profile Send private message Send e-mail
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
PostPosted: Mon Jan 09, 2006 11:19 pm Post subject: Reply with quote
jaxx wrote:
Still... Since you can deduce most offensive weights logically, shouldn't you be able to get the correct weight for assists by regression? (I half expect that Hollinger actually did this, and ended up with a value of 0.33.)
It's tough, because assists don't end up of much use at the team level. You can regress FG% against team performance, for example, but with assists the value is in increasing FG%, not so much the assist itself. There is a correlation between assists and wins, but as is discussed in the other recent thread, that's because assists are so strongly correlated with FGM and FGM with wins. Assist/FGM is occasionally a negative indicator at the team level in the NBA, because guys who can create their own shot are valuable.
So the two methods you're left with are:
1. A comprehensive study of how much a pass that would be an assist if the shot is made increases the efficiency of a shooter.
2. Using some sort of adjusted plus-minus rating and regressing against individual statistics, including assists.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Tue Jan 10, 2006 7:09 am Post subject: Reply with quote
At the end of last season, Roland at 82games.com linked to a list of all players and the % of their FG which were unassisted. Here are a few examples, from the Timberwolves:
%unA fac Minnesota
.335 .88 Garnett,Kevin
.252 .85 Szczerbiak,Wally
.342 .88 Sprewell,Latrell
.324 .88 Hassell,Trenton
.613 .94 Hudson,Troy
.639 .95 Cassell,Sam
.292 .87 Griffin,Eddie
.058 .72 Hoiberg,Fred
.404 .90 Olowokandi,Michael
Garnett 'created' 33.5% of his shots, Wally only 25%. Playmakers Cassell and Hudson had the highest 'self-made' rates; and TS% champ Hoiberg only created his own FG once every 10 games.
Column 2 is column 1 raised to the .116 power. This was the value which yielded the highest correlation between team stats and team pt.-differential. It's factored (along with other elements, like TS%) into a player's scoring rate. I gave Hoiberg only 72% of the credit for his own scoring; Cassell 95% for his.
Before this adjustment, an NBA assist appeared (by regression) to be worth only .722 as much as an NBA point. After this transferrence of power, the assist seemed to be worth 1.20.
Before I incorporated this factor, it seemed Hoiberg was an under-utilized great weapon, who should get more minutes than Spree or Hudson; afterward, he looks like the weakest player here, except for the defensive specialist Hassell.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Ben
Joined: 13 Jan 2005
Posts: 264
Location: Iowa City
PostPosted: Tue Jan 10, 2006 2:10 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
admin wrote:
jaxx wrote:
I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it.
I don't know if I agree with that. He also says in his explanation, "All assists are not equal, and one of my long-term goals is to develop better ways of measuring their differences."
Quote:
Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof.
I agree and have mentioned it before, but now that I think through the issue I think the bigger issue is not that point guards are not getting enough credit for their assists, they're not getting enough credit for not being assisted. Point guards almost always have very low percentages of their field goals that are assisted (Steve Nash, for example, has been assisted on only 22% of his field goals this year), but PER assumes his rate as the same as his entire team (66.3%). That takes away a lot of points that Nash should be getting credit for.
John has spoken about factoring this into PER v2.0.
Has anybody posted anything on how much this changes the numbers?
Back to top
View user's profile Send private message
jaxx
Joined: 17 May 2005
Posts: 23
PostPosted: Tue Jan 10, 2006 4:43 pm Post subject: Reply with quote
Quote:
Before this adjustment, an NBA assist appeared (by regression) to be worth only .722 as much as an NBA point. After this transferrence of power, the assist seemed to be worth 1.20.
Great stuff, Mike. Illuminating.
Just to clarify: does this mean the average value should be 1.20, but that it rises to ~1.40 for Cassell-types and dips to ~1.00 for Hoiberg-types?
Back to top
View user's profile Send private message
Ben
Joined: 13 Jan 2005
Posts: 264
Location: Iowa City
PostPosted: Tue Jan 10, 2006 4:54 pm Post subject: Reply with quote
Jaxx,
From looking at your posts, I thought you might be interested in Dan Rosenbaum's work if you haven't seen it already:
http://www.82games.com/comm30.htm
He found the geometric mean of Points, Rebounds, and Assists to be quite valuable, but assists of no value over and above that. I believe he stated somewhere that big men with high assist rates seemed to be particularly valuable.
Back to top
View user's profile Send private message
Eli W
Joined: 01 Feb 2005
Posts: 402
PostPosted: Tue Jan 10, 2006 4:57 pm Post subject: Reply with quote
Dan's regressions are here:
http://sonicscentral.com/apbrmetrics/vi ... .php?t=327
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Tue Jan 10, 2006 5:17 pm Post subject: Reply with quote
[quote="jaxx"]
Quote:
..does this mean the average value should be 1.20, but that it rises to ~1.40 for Cassell-types and dips to ~1.00 for Hoiberg-types?
No, I haven't seen any reason an assist should be worth more when coming from one player or another. And there's no solid relation between assist rate and % unassisted FG. Team makeup, playing style, etc., go into that stuff. If you pass to your high-% shooters (in good position), you get more assists; and your team does better.
_________________
`
36% of all statistics are wrong
Joined: 17 May 2005
Posts: 23
PostPosted: Sat Jan 07, 2006 2:59 pm Post subject: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
Since "assists" seem to be the theme of this month... What is the expert's opinion here about Hollinger's crediting of Assists in PER?
As you probably know, he sees the assisted basket as
1/3 getting open
1/3 making the pass
1/3 making the shot
Following this logic, he then attributes 2/3 of the credit to the scorer, and only 1/3 to the passer.
However, this 2/3-1/3 credit split seems very unfair towards Stephon Marbury-type players who drive, draw the double team and then dish to the open man. Or even towards guys like Shaq, who routinely draw double coverage and are adept at finding the open man on the perimeter. In both instances, the shotmaker hardly has to move his feet to free himself; he sees his defender leave, catches the pass and shoots. So following Hollinger's logic, here the PASSER 'takes care of two of the three acts of making an assisted basket', and should receive 2/3.
Watching actual games, these types of assists seem fairly common, and I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it. Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof. However, I know Hollinger is an intelligent guy who probably knows more than I do, so maybe his 1/3-2/3 decision is rooted in analysis too complex to share with the general public.
What's the APBR party line on this?
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
PostPosted: Sat Jan 07, 2006 4:27 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
jaxx wrote:
I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it.
I don't know if I agree with that. He also says in his explanation, "All assists are not equal, and one of my long-term goals is to develop better ways of measuring their differences."
Quote:
Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof.
I agree and have mentioned it before, but now that I think through the issue I think the bigger issue is not that point guards are not getting enough credit for their assists, they're not getting enough credit for not being assisted. Point guards almost always have very low percentages of their field goals that are assisted (Steve Nash, for example, has been assisted on only 22% of his field goals this year), but PER assumes his rate as the same as his entire team (66.3%). That takes away a lot of points that Nash should be getting credit for.
John has spoken about factoring this into PER v2.0.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Sat Jan 07, 2006 7:56 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
jaxx wrote:
Since "assists" seem to be the theme of this month... What is the expert's opinion here about Hollinger's crediting of Assists in PER?
As you probably know, he sees the assisted basket as
1/3 getting open
1/3 making the pass
1/3 making the shot
Following this logic, he then attributes 2/3 of the credit to the scorer, and only 1/3 to the passer.
Also by this logic, passers who pass to guys missing shots should get 1/3 credit (=blame) for the missed FGA. But there is nothing in the record showing passes on missed shots.
Assists are a pain in the butt. Argue as much as you like, but assists are so fundamentally biased that it's hard to get them right. I'm content (not happy) with my approach of giving variable credit based on estimates of the ease of shot created for teammates. But...
A better thing to do is to propose other measures of teammate cooperation. The simplest thing is potential assists - passes that resulted in a FGA. That in combination with recording assists on passes that lead to FTM/FTA would provide some balance to the assist stat.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
jaxx
Joined: 17 May 2005
Posts: 23
PostPosted: Mon Jan 09, 2006 2:56 pm Post subject: Reply with quote
Thanks for your thoughts. Probably everybody in this community has spent a lot of time thinking through assists, and I agree - they are a pain.
Still... Since you can deduce most offensive weights logically, shouldn't you be able to get the correct weight for assists by regression? (I half expect that Hollinger actually did this, and ended up with a value of 0.33.)
Quote:
I think the bigger issue is not that point guards are not getting enough credit for their assists, they're not getting enough credit for not being assisted.
Wow. I never thought about that, great point. This is also easy to solve in a linear-weights method such as PER; you can estimate the excess percentage of non-assisted PG shots and correct PG credits for that.
Back to top
View user's profile Send private message
kbche
Joined: 19 Jul 2005
Posts: 51
Location: washington d.c.
PostPosted: Mon Jan 09, 2006 11:11 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
jaxx wrote:
Since "assists" seem to be the theme of this month... What is the expert's opinion here about Hollinger's crediting of Assists in PER?
As you probably know, he sees the assisted basket as
1/3 getting open
1/3 making the pass
1/3 making the shot
Following this logic, he then attributes 2/3 of the credit to the scorer, and only 1/3 to the passer.
However, this 2/3-1/3 credit split seems very unfair towards Stephon Marbury-type players who drive, draw the double team and then dish to the open man. Or even towards guys like Shaq, who routinely draw double coverage and are adept at finding the open man on the perimeter. In both instances, the shotmaker hardly has to move his feet to free himself; he sees his defender leave, catches the pass and shoots. So following Hollinger's logic, here the PASSER 'takes care of two of the three acts of making an assisted basket', and should receive 2/3.
Watching actual games, these types of assists seem fairly common, and I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it. Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof. However, I know Hollinger is an intelligent guy who probably knows more than I do, so maybe his 1/3-2/3 decision is rooted in analysis too complex to share with the general public.
What's the APBR party line on this?
Power forwards had the highest PER values last year. However, power forwards are the highest paid, had the highest NBA efficiency, and had the most impressive stats last season. Thus, PER is in agreement with last seasons stats in that regard. Check out my blog dated 5 november (http://kabnba.blogspot.com/2005/11/what ... ition.html).
Kimberly
Back to top
View user's profile Send private message Send e-mail
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
PostPosted: Mon Jan 09, 2006 11:19 pm Post subject: Reply with quote
jaxx wrote:
Still... Since you can deduce most offensive weights logically, shouldn't you be able to get the correct weight for assists by regression? (I half expect that Hollinger actually did this, and ended up with a value of 0.33.)
It's tough, because assists don't end up of much use at the team level. You can regress FG% against team performance, for example, but with assists the value is in increasing FG%, not so much the assist itself. There is a correlation between assists and wins, but as is discussed in the other recent thread, that's because assists are so strongly correlated with FGM and FGM with wins. Assist/FGM is occasionally a negative indicator at the team level in the NBA, because guys who can create their own shot are valuable.
So the two methods you're left with are:
1. A comprehensive study of how much a pass that would be an assist if the shot is made increases the efficiency of a shooter.
2. Using some sort of adjusted plus-minus rating and regressing against individual statistics, including assists.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Tue Jan 10, 2006 7:09 am Post subject: Reply with quote
At the end of last season, Roland at 82games.com linked to a list of all players and the % of their FG which were unassisted. Here are a few examples, from the Timberwolves:
%unA fac Minnesota
.335 .88 Garnett,Kevin
.252 .85 Szczerbiak,Wally
.342 .88 Sprewell,Latrell
.324 .88 Hassell,Trenton
.613 .94 Hudson,Troy
.639 .95 Cassell,Sam
.292 .87 Griffin,Eddie
.058 .72 Hoiberg,Fred
.404 .90 Olowokandi,Michael
Garnett 'created' 33.5% of his shots, Wally only 25%. Playmakers Cassell and Hudson had the highest 'self-made' rates; and TS% champ Hoiberg only created his own FG once every 10 games.
Column 2 is column 1 raised to the .116 power. This was the value which yielded the highest correlation between team stats and team pt.-differential. It's factored (along with other elements, like TS%) into a player's scoring rate. I gave Hoiberg only 72% of the credit for his own scoring; Cassell 95% for his.
Before this adjustment, an NBA assist appeared (by regression) to be worth only .722 as much as an NBA point. After this transferrence of power, the assist seemed to be worth 1.20.
Before I incorporated this factor, it seemed Hoiberg was an under-utilized great weapon, who should get more minutes than Spree or Hudson; afterward, he looks like the weakest player here, except for the defensive specialist Hassell.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Ben
Joined: 13 Jan 2005
Posts: 264
Location: Iowa City
PostPosted: Tue Jan 10, 2006 2:10 pm Post subject: Re: Assists in PER - strange decisionmaking by Hollinger? Reply with quote
admin wrote:
jaxx wrote:
I'm a bit befuddled that Hollinger doesn't correct his Assist credit split for it, especially since he is so outspoken about the "factual basis" behind it.
I don't know if I agree with that. He also says in his explanation, "All assists are not equal, and one of my long-term goals is to develop better ways of measuring their differences."
Quote:
Also, I've always had the feeling that PER undervalues point guards - look at the career PER totals for proof.
I agree and have mentioned it before, but now that I think through the issue I think the bigger issue is not that point guards are not getting enough credit for their assists, they're not getting enough credit for not being assisted. Point guards almost always have very low percentages of their field goals that are assisted (Steve Nash, for example, has been assisted on only 22% of his field goals this year), but PER assumes his rate as the same as his entire team (66.3%). That takes away a lot of points that Nash should be getting credit for.
John has spoken about factoring this into PER v2.0.
Has anybody posted anything on how much this changes the numbers?
Back to top
View user's profile Send private message
jaxx
Joined: 17 May 2005
Posts: 23
PostPosted: Tue Jan 10, 2006 4:43 pm Post subject: Reply with quote
Quote:
Before this adjustment, an NBA assist appeared (by regression) to be worth only .722 as much as an NBA point. After this transferrence of power, the assist seemed to be worth 1.20.
Great stuff, Mike. Illuminating.
Just to clarify: does this mean the average value should be 1.20, but that it rises to ~1.40 for Cassell-types and dips to ~1.00 for Hoiberg-types?
Back to top
View user's profile Send private message
Ben
Joined: 13 Jan 2005
Posts: 264
Location: Iowa City
PostPosted: Tue Jan 10, 2006 4:54 pm Post subject: Reply with quote
Jaxx,
From looking at your posts, I thought you might be interested in Dan Rosenbaum's work if you haven't seen it already:
http://www.82games.com/comm30.htm
He found the geometric mean of Points, Rebounds, and Assists to be quite valuable, but assists of no value over and above that. I believe he stated somewhere that big men with high assist rates seemed to be particularly valuable.
Back to top
View user's profile Send private message
Eli W
Joined: 01 Feb 2005
Posts: 402
PostPosted: Tue Jan 10, 2006 4:57 pm Post subject: Reply with quote
Dan's regressions are here:
http://sonicscentral.com/apbrmetrics/vi ... .php?t=327
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Tue Jan 10, 2006 5:17 pm Post subject: Reply with quote
[quote="jaxx"]
Quote:
..does this mean the average value should be 1.20, but that it rises to ~1.40 for Cassell-types and dips to ~1.00 for Hoiberg-types?
No, I haven't seen any reason an assist should be worth more when coming from one player or another. And there's no solid relation between assist rate and % unassisted FG. Team makeup, playing style, etc., go into that stuff. If you pass to your high-% shooters (in good position), you get more assists; and your team does better.
_________________
`
36% of all statistics are wrong
Re: More recovered assist threads
omverve
Joined: 24 Feb 2005
Posts: 17
PostPosted: Mon Mar 21, 2005 11:28 am Post subject: Assist attempts Reply with quote
I was reading over FFS's thread about assisted% and FG%, and it got me thinking. Wouldn't it be useful to have stats that track assists in much the same way we track FGs? That is, for each FGA that is eligible to generate an assist if it turns out to be a FGM, keep an analogous stat for assists-- "assist attempts," or AA, for lack of a better term. If the FGA is successful, we notch down a successful assist (call it assist make, or AM) as usual, but if the FGA misses, we still track the information that there would have been an assist, had the shot gone down-- that is, the missed FGA does not generate an AM, but it does generate an AA.
It seems there would be a wide range of applicability for such a stat. For instance, it would be very insightful for FFS's study. We could get a better handle on how a shooter is set up by teammates by charting his FG% (or eFG%, or whatever) on FGAs that qualify for AAs, vs. those that don't. As it stands, we only track AMs, so in the context of FFS's study, we count all missed 'assist attempted' shots against the player's ability to successfully create his own shot. Obviously, this artifact skews our ability to grasp what we really want to get at by making the player seem poorer at successfully making his own created FGAs than he really is.
The AA stat could also be useful in a range of other areas. For instance, we could get a better handle on to what extent a player's assist/possession numbers are skewed by how well or poorly his teammates shoot the ball. Intuitively, we know a player on a poor shooting team is going to have his assist numbers artificially decreased, but this stat would let us quantify that effect.
AAs would also help us get a grasp on something like 'passing efficiency,' or the quality of shots a player typically creates for his teammates. For instance, maybe it turns out that someone like a Francis does not create very good opportunities for his teammates, and thus their FG% on assist attempted shots from Francis is not significantly better than it is from, say, Doug Christie, or the 'average' NBA player, etc. I imagine we'd see a guy like Shaq have a significant effect on elevating his teammates' FG% on assist attempted shots, since he creates so many wide open looks on the perimeter and is also a great passer in the paint. Likewise, we'd probably see Nash excelling at this stat, as he creates so many shots at the rim and so on.
Of course, to do all this we'd need even more refined information-- something like separate stats for a player's FG% on assist attempted shots from each of his teammates. And even the more general idea of AAs, without keeping track of whom the pass comes from, seems impossible to recover from play-by-play data. Are there any thoughts on the potential utility of AAs, or ways to keep track of it in the future?
Back to top
View user's profile Send private message
tomverve
Joined: 24 Feb 2005
Posts: 17
PostPosted: Mon Mar 21, 2005 11:45 am Post subject: Reply with quote
A quick addendum: I also would love to see assist-type passes that lead to fouls on shot attempts accounted for somehow. I imagine one could simply count such passes as successful assists, or better, keep a separate stat for passes that lead to fouls on shot attempts. It seems that with such stats, in addition to some form of AA considerations as mentioned above, we could get a much more sophisticated and insightful analysis of the passing abilities/values of various players.
Back to top
View user's profile Send private message
Joined: 24 Feb 2005
Posts: 17
PostPosted: Mon Mar 21, 2005 11:28 am Post subject: Assist attempts Reply with quote
I was reading over FFS's thread about assisted% and FG%, and it got me thinking. Wouldn't it be useful to have stats that track assists in much the same way we track FGs? That is, for each FGA that is eligible to generate an assist if it turns out to be a FGM, keep an analogous stat for assists-- "assist attempts," or AA, for lack of a better term. If the FGA is successful, we notch down a successful assist (call it assist make, or AM) as usual, but if the FGA misses, we still track the information that there would have been an assist, had the shot gone down-- that is, the missed FGA does not generate an AM, but it does generate an AA.
It seems there would be a wide range of applicability for such a stat. For instance, it would be very insightful for FFS's study. We could get a better handle on how a shooter is set up by teammates by charting his FG% (or eFG%, or whatever) on FGAs that qualify for AAs, vs. those that don't. As it stands, we only track AMs, so in the context of FFS's study, we count all missed 'assist attempted' shots against the player's ability to successfully create his own shot. Obviously, this artifact skews our ability to grasp what we really want to get at by making the player seem poorer at successfully making his own created FGAs than he really is.
The AA stat could also be useful in a range of other areas. For instance, we could get a better handle on to what extent a player's assist/possession numbers are skewed by how well or poorly his teammates shoot the ball. Intuitively, we know a player on a poor shooting team is going to have his assist numbers artificially decreased, but this stat would let us quantify that effect.
AAs would also help us get a grasp on something like 'passing efficiency,' or the quality of shots a player typically creates for his teammates. For instance, maybe it turns out that someone like a Francis does not create very good opportunities for his teammates, and thus their FG% on assist attempted shots from Francis is not significantly better than it is from, say, Doug Christie, or the 'average' NBA player, etc. I imagine we'd see a guy like Shaq have a significant effect on elevating his teammates' FG% on assist attempted shots, since he creates so many wide open looks on the perimeter and is also a great passer in the paint. Likewise, we'd probably see Nash excelling at this stat, as he creates so many shots at the rim and so on.
Of course, to do all this we'd need even more refined information-- something like separate stats for a player's FG% on assist attempted shots from each of his teammates. And even the more general idea of AAs, without keeping track of whom the pass comes from, seems impossible to recover from play-by-play data. Are there any thoughts on the potential utility of AAs, or ways to keep track of it in the future?
Back to top
View user's profile Send private message
tomverve
Joined: 24 Feb 2005
Posts: 17
PostPosted: Mon Mar 21, 2005 11:45 am Post subject: Reply with quote
A quick addendum: I also would love to see assist-type passes that lead to fouls on shot attempts accounted for somehow. I imagine one could simply count such passes as successful assists, or better, keep a separate stat for passes that lead to fouls on shot attempts. It seems that with such stats, in addition to some form of AA considerations as mentioned above, we could get a much more sophisticated and insightful analysis of the passing abilities/values of various players.
Back to top
View user's profile Send private message
Re: More recovered assist threads
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Wed Nov 22, 2006 8:11 pm Post subject: Bird vs Dirk Reply with quote
I was having a discussion about Dirk06/Bird88 in another forum and most of it pertained to their peak PER seasons.
Dirk seems to be a slightly better scorer in his era (pace and team adjusted). Most of the other stats are even, but Bird has a decisive passing edge.
But considering PER factors Player APG/Team APG it sort of makes Dirk look even more superior on offense, because Birds 6.1 - 2.8apg (+3.1apg) passing edge isn't as big because the Celtics of 1988 were a superior passing team. Meaning since Dirks team was a poor passing team meaning Bird's passing edge relative to his team context is less credited with the PER system.
Is this fair? Is it a weakness in the PER system? Or is Dirk actually just a little better than Bird offensively because his offensive PER is a little better?
Dirk had 15.6% of his teams assists, Bird had 20%. This means that Bird didn't have that much more proportion of team assists as the +3.1apg differential indicates. Its sort of like the Win Share Argument in a way. Dirk is more responsible for his teams offense if you factor his slight edge in scoring over Bird (only because of ERA and league rating -- this edge is not noticable in basic per 40, TS% stats).
Anyone agree with this potential assist flaw? Or does it make sense to say these guys were about dead even offensively, despite their basic per 40 stats look the same (even era adjusted) -- where Bird has more assists?
http://basketball-reference.com/teams/BOS/1988.html
http://basketball-reference.com/teams/DAL/2006.html
This would be a perfect time to see these players offensive PER's alone. But I don't think such a spreadsheet exists.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Thu Nov 23, 2006 6:40 am Post subject: Reply with quote
Is it better to have a higher % of your team's assists? Doesn't a good pass often lead to another good pass, which becomes the assist?
Some ballhogger types won't give it up until they might get an assist. Suppose Bird would not make the easy pass, unless it was to a likely shooter. Then his Ast/TmAst ratio would go up, as he became a poorer passer.
As it was, several Celtics made good/easy passes, and eventually someone got an assist. They enhanced one anothers' assist rates, and no one had a really high % of them. Had anyone become selfish, the team would score less, with fewer assists; and the one guy would have a higher % of them.
A lot more successful teams had balanced distribution of their assists, than had one-man focus of them. Other than late-80s Magic/Lakers, I can't even think of anyone else who won a title with a dominant passer. Cousy/Celtics, anyone else?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Thu Nov 23, 2006 9:50 am Post subject: Reply with quote
Mike G wrote:
Is it better to have a higher % of your team's assists? Doesn't a good pass often lead to another good pass, which becomes the assist?
Thats what I am pondering. Dirk is stylistically and from general observation a weaker passer. But when comparing these guys in their prime seasons objectively they look very similiar except in passing. Bird has a sizable edge in passing, Dirk has a small edge in scoring relative to era/team context.
Dirk looks better offensively than Bird despite his mediocre passing because he is contributing more offensively by having only a slightly smaller % ratio of PlayerAPG/Team APG relative to Larry Bird. So that one passing edge Bird has in the context of Offensive PER when comparing with Dirk.
Is this fair when comparing these guys objectively. Might Bird's APG drop if he were to replace Dirk?
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Thu Nov 23, 2006 10:49 am Post subject: Reply with quote
To state the obvious: the ultimate purpose of an offense is to score points, and to score as many points as possible in the finite possessions allowed.
Assists are one way to generate points, but they aren't the only way, and depending on the team and talents involved, they aren't even necessarily the best way. Dallas illustrated that last year. If I remember correctly, they were perhaps the top offensive team but had one of the lowest number of assisted baskets. I'm sure there are other examples.
This is part of what makes assists so difficult to value -- their optional nature in offense. But personally, I wouldn't hold it against a player how he creates his offense, I'd just look at how efficient it was and how much of it there was. I'd credit Bird for his assists, but mostly in how it influenced his Usage Rate and the amount of offense he generated
Of course, maybe with different teams these players would change in value? Was Nowitzki in a special situation that allowed him to shoot and not pass much? Would Bird's passing be more valuable to the average team? I don't know.
Back to top
View user's profile Send private message
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Thu Nov 23, 2006 11:36 am Post subject: Reply with quote
Bird's Usage Rate in this comparison will be higher, because his team generates more assists in general. So those extra assists he has on Dirk, while not being a large difference if you factor PlayerAPG/TeamAPG (aka OFFENSIVE PER) is large in terms of the Usage Rate Equation.
I guess this why a lot of players and guards especially seemed to have higher Usage Rates back in the 80s. (This was another question I posted before, but didn't generate any discussion).
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Thu Nov 23, 2006 2:59 pm Post subject: Reply with quote
Ack. I just realized this thread was not about what I initially thought. It was early in the morning here, and I wasn't paying close enough attention. Sorry.
Quote:
Meaning since Dirks team was a poor passing team meaning Bird's passing edge relative to his team context is less credited with the PER system.
I'm not quite understanding this part. I thought PER credited each assist as 0.67 points? I thought it was the FGM that varied in value depending on a team's assist rate? What am I missing?
Back to top
View user's profile Send private message
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Fri Nov 24, 2006 8:54 am Post subject: Reply with quote
Quote:
I'm not quite understanding this part. I thought PER credited each assist as 0.67 points? I thought it was the FGM that varied in value depending on a team's assist rate? What am I missing?
I think you are right actually but I am not positive. Even if a players PER is credit more for a higher proportion of Player APG/ Team FGM, doesn't that still sort of undersell Bird for having good passing teamattes? Or does it sound logical for Dirk because he is more responsible for his teams offense relative to Larry Bird -- despite being an objectively worse passer, and a seemingly equal scorer?
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Fri Nov 24, 2006 12:01 pm Post subject: Reply with quote
Quote:
doesn't that still sort of undersell Bird for having good passing teamattes?
It seems fair to me.
If Bird is playing with better passers then it's safe to assume that they are setting him up for a few easy shots per game. Among other things, this is going to boost his TS%, and it might inflate his scoring average. Nowitzki, on the other hand, doesn't get that benefit and has to create most of his shots himself. (And Nowtizki, if I remember right, had a lower TS% than Bird.)
It would seem to me that Bird would have an unfair advantage -- an advantage coming from his teammates and not his own abilities -- if we didn't even it out some. So he only gets partial credit on his estimated assisted baskets.
Quote:
despite being an objectively worse passer, and a seemingly equal scorer?
You could argue that Nowitzki is a superior scorer to Bird because he didn't get the freebie passes every game that Bird got. Nowitzki had to create each of his off the dribble against a defender. I think that's what PER is trying to say at least.
Back to top
View user's profile Send private message
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Fri Nov 24, 2006 3:31 pm Post subject: Reply with quote
Yeah I thought it made sense at first, but on second look and after arguing on the boards, I felt that Bird deserved an edge, because he is a better passer -- and that is easier to see in the general GLORY STATS (basic per 40 era adjusted stats) -- not as easy to tell that Dirk was the superior scorer if you UNLESS you adjust for everything -- in general Dirk is superior on offense SLIGHTLY.
I see them basically as equals I guess in terms of regular season prime production -- considering neither was a standout defender. Bird the better passer, Dirk the slightly better offensive player. Both equal in rebounding I guess.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Fri Nov 24, 2006 5:11 pm Post subject: Reply with quote
Nikos wrote:
I see them basically as equals I guess in terms of regular season prime production ... Bird the better passer, Dirk the slightly better offensive player. Both equal in rebounding I guess.
Equal in rebounding, Bird was equal or slightly better scorer and much better assister. How does Dirk seem to be better offensively?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Fri Nov 24, 2006 6:26 pm Post subject: Reply with quote
Mike G wrote:
Nikos wrote:
I see them basically as equals I guess in terms of regular season prime production ... Bird the better passer, Dirk the slightly better offensive player. Both equal in rebounding I guess.
Equal in rebounding, Bird was equal or slightly better scorer and much better assister. How does Dirk seem to be better offensively?
Why are there PER's even?
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Fri Nov 24, 2006 8:44 pm Post subject: Reply with quote
Nikos wrote:
Yeah I thought it made sense at first, but on second look and after arguing on the boards, I felt that Bird deserved an edge, because he is a better passer
Very reasonable. It all depends on how much you value assists, which are one of the toughest things in basketball to value. I'd say their value isn't even very constant -- some teams need them more than others. But yeah, if you think that assists (or at least Bird's assists) should be more valuable than 0.67 points, then Bird would rate as the better player.
If you accept Hollinger's arbitrary value of 0.67 points then they are about equal players -- though Nowitzki gets a slight edge.
Quote:
Why are there PER's even?
Because PER values scoring and avoiding turnovers more than creating assists.
Back to top
View user's profile Send private message
moneyp
Joined: 24 Feb 2005
Posts: 69
PostPosted: Mon Nov 27, 2006 2:29 pm Post subject: Reply with quote
What percentage of his team's shots did Bird take? What percentage did Nowitzki take? I think teams spread the ball around more in Bird's day.
One result of that, I think, is that the standard deviation of PER was lower in Bird's day than it is in Nowitzki's. I don't know if anyone's ever bothered to factor that in.
Joined: 16 Jan 2005
Posts: 346
PostPosted: Wed Nov 22, 2006 8:11 pm Post subject: Bird vs Dirk Reply with quote
I was having a discussion about Dirk06/Bird88 in another forum and most of it pertained to their peak PER seasons.
Dirk seems to be a slightly better scorer in his era (pace and team adjusted). Most of the other stats are even, but Bird has a decisive passing edge.
But considering PER factors Player APG/Team APG it sort of makes Dirk look even more superior on offense, because Birds 6.1 - 2.8apg (+3.1apg) passing edge isn't as big because the Celtics of 1988 were a superior passing team. Meaning since Dirks team was a poor passing team meaning Bird's passing edge relative to his team context is less credited with the PER system.
Is this fair? Is it a weakness in the PER system? Or is Dirk actually just a little better than Bird offensively because his offensive PER is a little better?
Dirk had 15.6% of his teams assists, Bird had 20%. This means that Bird didn't have that much more proportion of team assists as the +3.1apg differential indicates. Its sort of like the Win Share Argument in a way. Dirk is more responsible for his teams offense if you factor his slight edge in scoring over Bird (only because of ERA and league rating -- this edge is not noticable in basic per 40, TS% stats).
Anyone agree with this potential assist flaw? Or does it make sense to say these guys were about dead even offensively, despite their basic per 40 stats look the same (even era adjusted) -- where Bird has more assists?
http://basketball-reference.com/teams/BOS/1988.html
http://basketball-reference.com/teams/DAL/2006.html
This would be a perfect time to see these players offensive PER's alone. But I don't think such a spreadsheet exists.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Thu Nov 23, 2006 6:40 am Post subject: Reply with quote
Is it better to have a higher % of your team's assists? Doesn't a good pass often lead to another good pass, which becomes the assist?
Some ballhogger types won't give it up until they might get an assist. Suppose Bird would not make the easy pass, unless it was to a likely shooter. Then his Ast/TmAst ratio would go up, as he became a poorer passer.
As it was, several Celtics made good/easy passes, and eventually someone got an assist. They enhanced one anothers' assist rates, and no one had a really high % of them. Had anyone become selfish, the team would score less, with fewer assists; and the one guy would have a higher % of them.
A lot more successful teams had balanced distribution of their assists, than had one-man focus of them. Other than late-80s Magic/Lakers, I can't even think of anyone else who won a title with a dominant passer. Cousy/Celtics, anyone else?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Thu Nov 23, 2006 9:50 am Post subject: Reply with quote
Mike G wrote:
Is it better to have a higher % of your team's assists? Doesn't a good pass often lead to another good pass, which becomes the assist?
Thats what I am pondering. Dirk is stylistically and from general observation a weaker passer. But when comparing these guys in their prime seasons objectively they look very similiar except in passing. Bird has a sizable edge in passing, Dirk has a small edge in scoring relative to era/team context.
Dirk looks better offensively than Bird despite his mediocre passing because he is contributing more offensively by having only a slightly smaller % ratio of PlayerAPG/Team APG relative to Larry Bird. So that one passing edge Bird has in the context of Offensive PER when comparing with Dirk.
Is this fair when comparing these guys objectively. Might Bird's APG drop if he were to replace Dirk?
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Thu Nov 23, 2006 10:49 am Post subject: Reply with quote
To state the obvious: the ultimate purpose of an offense is to score points, and to score as many points as possible in the finite possessions allowed.
Assists are one way to generate points, but they aren't the only way, and depending on the team and talents involved, they aren't even necessarily the best way. Dallas illustrated that last year. If I remember correctly, they were perhaps the top offensive team but had one of the lowest number of assisted baskets. I'm sure there are other examples.
This is part of what makes assists so difficult to value -- their optional nature in offense. But personally, I wouldn't hold it against a player how he creates his offense, I'd just look at how efficient it was and how much of it there was. I'd credit Bird for his assists, but mostly in how it influenced his Usage Rate and the amount of offense he generated
Of course, maybe with different teams these players would change in value? Was Nowitzki in a special situation that allowed him to shoot and not pass much? Would Bird's passing be more valuable to the average team? I don't know.
Back to top
View user's profile Send private message
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Thu Nov 23, 2006 11:36 am Post subject: Reply with quote
Bird's Usage Rate in this comparison will be higher, because his team generates more assists in general. So those extra assists he has on Dirk, while not being a large difference if you factor PlayerAPG/TeamAPG (aka OFFENSIVE PER) is large in terms of the Usage Rate Equation.
I guess this why a lot of players and guards especially seemed to have higher Usage Rates back in the 80s. (This was another question I posted before, but didn't generate any discussion).
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Thu Nov 23, 2006 2:59 pm Post subject: Reply with quote
Ack. I just realized this thread was not about what I initially thought. It was early in the morning here, and I wasn't paying close enough attention. Sorry.
Quote:
Meaning since Dirks team was a poor passing team meaning Bird's passing edge relative to his team context is less credited with the PER system.
I'm not quite understanding this part. I thought PER credited each assist as 0.67 points? I thought it was the FGM that varied in value depending on a team's assist rate? What am I missing?
Back to top
View user's profile Send private message
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Fri Nov 24, 2006 8:54 am Post subject: Reply with quote
Quote:
I'm not quite understanding this part. I thought PER credited each assist as 0.67 points? I thought it was the FGM that varied in value depending on a team's assist rate? What am I missing?
I think you are right actually but I am not positive. Even if a players PER is credit more for a higher proportion of Player APG/ Team FGM, doesn't that still sort of undersell Bird for having good passing teamattes? Or does it sound logical for Dirk because he is more responsible for his teams offense relative to Larry Bird -- despite being an objectively worse passer, and a seemingly equal scorer?
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Fri Nov 24, 2006 12:01 pm Post subject: Reply with quote
Quote:
doesn't that still sort of undersell Bird for having good passing teamattes?
It seems fair to me.
If Bird is playing with better passers then it's safe to assume that they are setting him up for a few easy shots per game. Among other things, this is going to boost his TS%, and it might inflate his scoring average. Nowitzki, on the other hand, doesn't get that benefit and has to create most of his shots himself. (And Nowtizki, if I remember right, had a lower TS% than Bird.)
It would seem to me that Bird would have an unfair advantage -- an advantage coming from his teammates and not his own abilities -- if we didn't even it out some. So he only gets partial credit on his estimated assisted baskets.
Quote:
despite being an objectively worse passer, and a seemingly equal scorer?
You could argue that Nowitzki is a superior scorer to Bird because he didn't get the freebie passes every game that Bird got. Nowitzki had to create each of his off the dribble against a defender. I think that's what PER is trying to say at least.
Back to top
View user's profile Send private message
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Fri Nov 24, 2006 3:31 pm Post subject: Reply with quote
Yeah I thought it made sense at first, but on second look and after arguing on the boards, I felt that Bird deserved an edge, because he is a better passer -- and that is easier to see in the general GLORY STATS (basic per 40 era adjusted stats) -- not as easy to tell that Dirk was the superior scorer if you UNLESS you adjust for everything -- in general Dirk is superior on offense SLIGHTLY.
I see them basically as equals I guess in terms of regular season prime production -- considering neither was a standout defender. Bird the better passer, Dirk the slightly better offensive player. Both equal in rebounding I guess.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Fri Nov 24, 2006 5:11 pm Post subject: Reply with quote
Nikos wrote:
I see them basically as equals I guess in terms of regular season prime production ... Bird the better passer, Dirk the slightly better offensive player. Both equal in rebounding I guess.
Equal in rebounding, Bird was equal or slightly better scorer and much better assister. How does Dirk seem to be better offensively?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Nikos
Joined: 16 Jan 2005
Posts: 346
PostPosted: Fri Nov 24, 2006 6:26 pm Post subject: Reply with quote
Mike G wrote:
Nikos wrote:
I see them basically as equals I guess in terms of regular season prime production ... Bird the better passer, Dirk the slightly better offensive player. Both equal in rebounding I guess.
Equal in rebounding, Bird was equal or slightly better scorer and much better assister. How does Dirk seem to be better offensively?
Why are there PER's even?
Back to top
View user's profile Send private message
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Fri Nov 24, 2006 8:44 pm Post subject: Reply with quote
Nikos wrote:
Yeah I thought it made sense at first, but on second look and after arguing on the boards, I felt that Bird deserved an edge, because he is a better passer
Very reasonable. It all depends on how much you value assists, which are one of the toughest things in basketball to value. I'd say their value isn't even very constant -- some teams need them more than others. But yeah, if you think that assists (or at least Bird's assists) should be more valuable than 0.67 points, then Bird would rate as the better player.
If you accept Hollinger's arbitrary value of 0.67 points then they are about equal players -- though Nowitzki gets a slight edge.
Quote:
Why are there PER's even?
Because PER values scoring and avoiding turnovers more than creating assists.
Back to top
View user's profile Send private message
moneyp
Joined: 24 Feb 2005
Posts: 69
PostPosted: Mon Nov 27, 2006 2:29 pm Post subject: Reply with quote
What percentage of his team's shots did Bird take? What percentage did Nowitzki take? I think teams spread the ball around more in Bird's day.
One result of that, I think, is that the standard deviation of PER was lower in Bird's day than it is in Nowitzki's. I don't know if anyone's ever bothered to factor that in.
Re: More recovered assist threads
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 1:35 am Post subject: Individual Efficiency Ratings (BoP) from PBP Reply with quote
I would like to see what we can do with individual efficiency ratings (as Dean defines in BoP) that are extracted from the play-by-play.
My first result comes from individual offensive ratings. The spreadsheet of ratings can be found here. The data in this spreadsheet is:
Code:
* Ortg: the player’s offensive efficiency rating
* Usg%: the percentage of possessions used by this player while on the court
* Total Used: the total number of possessions this player used
* %Shots: percentage of possessions used that were shots
* %Free Throws: percentage of possessions used that were free throws
* %Assists: percentage of possessions used that were assists
* Assist eFG%: mean expected eFG% of assists
* %Oreb: percentage of possessions used that were offensive rebounds
* %Turnover: percentage of possessions used that were turnovers
I also used Eli's usg vs efficiency method and estimate there to be an effect, where as my estimate is 11 versus Eli's 25. Full details in the post.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Joe
Joined: 27 Sep 2009
Posts: 94
Location: Long Island, NY
PostPosted: Sun Oct 25, 2009 10:55 am Post subject: Reply with quote
Great stuff. Curious about one thing, though.
"A player uses a possession and produces points in one of the following ways:
* Making a field goal or free throw
* Assisting on a field goal
* Obtaining an offensive rebound that leads to a made field goal or free throw
* Missed field goals or free throws that are rebounded by the defense
* Turnovers"
I don't think you clarify anywhere, and I haven't read BoP in awhile, but how are possessions used affected on missed field goal attempts that are offensive rebounded? Does the shooter get credited for the used possession at all? Because it doesn't seem like that's the case here.
Looking forward to see the defensive ratings. This is really great stuff.
Back to top
View user's profile Send private message Visit poster's website
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 10:56 am Post subject: Reply with quote
If there is an offensive rebound then the guy that missed the shot gets a free pass. Very Happy
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
deepak
Joined: 26 Apr 2006
Posts: 665
PostPosted: Sun Oct 25, 2009 1:25 pm Post subject: Reply with quote
Nice work.
"Assist eFG%: mean expected eFG% of assists"
Can you elaborate on how you calculate this?
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 1:41 pm Post subject: Reply with quote
For each shot a player assists on I calculate an expected eFG% for that shot. So assist eFG% is simply the mean of all of these expected eFG%'s.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
deepak
Joined: 26 Apr 2006
Posts: 665
PostPosted: Sun Oct 25, 2009 2:02 pm Post subject: Reply with quote
Thanks. Using that approach, could you not also estimate how many potential assists players had?
And I compared your efficiency numbers to the estimated ORTG from basketball-reference (only looked at players who weren't traded though).
It looks like the estimated ORTG tends to underrate guards:
Code:
Team Player min Ortg Usg% est_ORTG est_Usg% ORTG_diff Usg%_diff
1 ORL Jameer Nelson 1309 131 23.8 121 23.2 10 0.6
2 MIN Sebastian Telfair 2095 109 20.7 100 19.6 9 1.1
3 ATL Ronald Murray 1975 112 23.2 104 24.3 8 -1.1
4 DAL Antoine Wright 1552 110 14.8 102 15.6 8 -0.8
5 IND T.J. Ford 2258 115 24.9 108 23.5 7 1.4
6 LAC Baron Davis 2248 104 26.8 97 25.2 7 1.6
7 DEN Anthony Carter 1786 105 15.7 98 14.3 7 1.4
8 OKC Earl Watson 1776 100 18.5 93 16.3 7 2.2
9 DAL Jose Juan Barea 1600 113 21.9 106 20.2 7 1.7
10 ORL Anthony Johnson 1481 111 16.9 104 15.5 7 1.4
11 ATL Joe Johnson 3124 116 26.8 110 26.6 6 0.2
12 NOH Chris Paul 3002 130 30.7 124 27.5 6 3.2
13 CHI Derrick Rose 3000 114 23.1 108 22.6 6 0.5
14 WAS Caron Butler 2585 114 25.8 108 25.9 6 -0.1
15 DET Rodney Stuckey 2517 110 23.8 104 23 6 0.8
16 MIN Randy Foye 2494 112 22.7 106 22.6 6 0.1
17 MIL Ramon Sessions 2173 117 23.4 111 22.4 6 1
18 NJN Keyon Dooling 2074 118 17.7 112 17.4 6 0.3
19 PHI Lou Williams 1919 111 26.8 105 27.7 6 -0.9
20 CLE LeBron James 3054 127 34 122 33.8 5 0.2
And overrate bigs:
Code:
Team Player min Ortg Usg% est_ORTG est_Usg% ORTG_diff Usg%_diff
1 CHA Emeka Okafor 2691 104 18.7 112 18.9 -8 -0.2
2 CHI Joakim Noah 1938 115 13 123 12.3 -8 0.7
3 LAC Marcus Camby 1925 104 16 112 15.8 -8 0.2
4 MEM Darrell Arthur 1464 91 15.4 99 16.3 -8 -0.9
5 DET Jason Maxiell 1408 112 15 120 14.7 -8 0.3
6 LAL Sasha Vujacic 1293 107 15.2 115 16 -8 -0.8
7 ORL Tony Battie 1202 99 15.6 106 15.6 -7 0
8 PHI Reggie Evans 1140 95 13.7 102 13 -7 0.7
9 GSW Anthony Randolph 1129 95 21 102 21.7 -7 -0.7
10 MIA Udonis Haslem 2560 107 14.1 113 14.7 -6 -0.6
11 IND Troy Murphy 2482 114 16.8 120 16.9 -6 -0.1
12 CLE Anderson Varejao 2306 110 14.2 116 14.2 -6 0
13 MIL Charlie Villanueva 2095 100 27.8 106 28.5 -6 -0.7
14 MIN Kevin Love 2048 106 21 112 21 -6 0
15 PHI Samuel Dalembert 2036 97 13.5 103 13.6 -6 -0.1
16 POR Joel Przybilla 1952 118 10.5 124 10.2 -6 0.3
17 IND Jeff Foster 1828 109 12.8 115 12 -6 0.8
18 ATL Zaza Pachulia 1473 107 17.1 113 16.3 -6 0.8
19 NJN Yi Jianlian 1421 90 20.3 96 20.5 -6 -0.2
20 PHI Marreese Speights 1262 109 20.2 115 21.9 -6 -1.7
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 2:12 pm Post subject: Reply with quote
Quote:
... could you not also estimate how many potential assists players had?
I know there is a rule of thumb for potential assists (see here and here), so perhaps we can intelligently build this in to give assisters a better measure of failure instead of using the 1/2 multiplier for giving credit on assists.
Also, thanks for doing that comparison. I know B-R estimates things slightly differently than what Dean uses in BoP, so I want to at some point in the future compare these results with Dean's estimates to see 1) what differences there are and 2) see how much better one is at predicting than the other (if at all).
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
fundamentallysound
Joined: 18 Jul 2008
Posts: 25
Location: VA
PostPosted: Sun Oct 25, 2009 10:49 pm Post subject: Reply with quote
Ryan, this is great, great work. One quick thing that I just want to make sure I'm understanding properly. You, similarly to Eli Witus, found that there was a tradeoff for Usage and Efficiency. The magnitude of the trade off in your study was lower, though, than Eli's, is that correct?
So, if someone had previously been relying on the +/- 1.25 gain or drop in ORtg based on a decrease or increase in Usg% of 1%, they would be overvaluing high usage players relative to the numbers you came up with?
So, if someone were to use your numbers, instead, to standardize all ORtg's so that everyone was operating on 20% of possessions, high usage players would get a boost to their standardized ORtg of 1.11 pts/100 poss. for every 1% over 20% Usg they had?
Maybe it's improper to use these things in this way (i.e. the relationship between usg and efficiency might be more curved than linear), but I just wanted to make sure I was understanding the basic mechanics.
Sorry if that was confusing, sometimes I don't articulate my thoughts the best way, especially when it relates to this stuff where I am fairly novice.
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 11:27 pm Post subject: Reply with quote
I think your math is close. I estimate that each 1% increase over 100% in a lineup's usg% adds 0.11 points per hundred possessions. So we'd need a 9% increase in the lineup's usg% to expect a 1 point per hundred poss increase in efficiency for the lineup. Eli's estimate was 4% instead of 9%. The std error on his estimate is large enough to consider my estimate plausible. The bounds on my estimate are anywhere from 6% to 19%.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Mon Oct 26, 2009 9:39 am Post subject: Reply with quote
I believe Ryan's numbers suggest a 0.55 change in efficiency for every 1% change in usage as opposed to a 1.25 to 1 ratio that Eli's numbers suggested.
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Mon Oct 26, 2009 11:11 am Post subject: Reply with quote
I think that David was able to correctly articulate the estimated individual effect. Very Happy
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Fri Oct 30, 2009 4:41 pm Post subject: Reply with quote
I've now extracted the defensive efficiency ratings. The spreadsheet of the ratings can be found here. The data in the spreadsheet:
Code:
* Drtg: the player’s defensive efficiency rating
* Std Err: the standard error of the rating
* 95% CI: a 95% confidence interval for the rating
* Usg%: the percentage of possessions used by this player while on the court
* Total Used: the total number of possessions this player used
* %Shots: percentage of possessions used that were shots
* %Fouls: percentage of possessions used that were fouls
* %Drebs: percentage of possessions used that were defensive rebounds
* %Turnovers: percentage of possessions used that were turnovers
Full details can be found in the post.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
deepak
Joined: 26 Apr 2006
Posts: 665
PostPosted: Fri Oct 30, 2009 10:54 pm Post subject: Reply with quote
Ryan J. Parker wrote:
I've now extracted the defensive efficiency ratings. The spreadsheet of the ratings can be found here. The data in the spreadsheet:
Interesting stuff.
For assigning credit to shots, you may also consider basing it on floor location, or some combination of floor location and counterpart data. For instance, guards are more likely to be responsible for shots on the perimeter, and centers are more likely to be responsible for shots at the rim.
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 1:35 am Post subject: Individual Efficiency Ratings (BoP) from PBP Reply with quote
I would like to see what we can do with individual efficiency ratings (as Dean defines in BoP) that are extracted from the play-by-play.
My first result comes from individual offensive ratings. The spreadsheet of ratings can be found here. The data in this spreadsheet is:
Code:
* Ortg: the player’s offensive efficiency rating
* Usg%: the percentage of possessions used by this player while on the court
* Total Used: the total number of possessions this player used
* %Shots: percentage of possessions used that were shots
* %Free Throws: percentage of possessions used that were free throws
* %Assists: percentage of possessions used that were assists
* Assist eFG%: mean expected eFG% of assists
* %Oreb: percentage of possessions used that were offensive rebounds
* %Turnover: percentage of possessions used that were turnovers
I also used Eli's usg vs efficiency method and estimate there to be an effect, where as my estimate is 11 versus Eli's 25. Full details in the post.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Joe
Joined: 27 Sep 2009
Posts: 94
Location: Long Island, NY
PostPosted: Sun Oct 25, 2009 10:55 am Post subject: Reply with quote
Great stuff. Curious about one thing, though.
"A player uses a possession and produces points in one of the following ways:
* Making a field goal or free throw
* Assisting on a field goal
* Obtaining an offensive rebound that leads to a made field goal or free throw
* Missed field goals or free throws that are rebounded by the defense
* Turnovers"
I don't think you clarify anywhere, and I haven't read BoP in awhile, but how are possessions used affected on missed field goal attempts that are offensive rebounded? Does the shooter get credited for the used possession at all? Because it doesn't seem like that's the case here.
Looking forward to see the defensive ratings. This is really great stuff.
Back to top
View user's profile Send private message Visit poster's website
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 10:56 am Post subject: Reply with quote
If there is an offensive rebound then the guy that missed the shot gets a free pass. Very Happy
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
deepak
Joined: 26 Apr 2006
Posts: 665
PostPosted: Sun Oct 25, 2009 1:25 pm Post subject: Reply with quote
Nice work.
"Assist eFG%: mean expected eFG% of assists"
Can you elaborate on how you calculate this?
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 1:41 pm Post subject: Reply with quote
For each shot a player assists on I calculate an expected eFG% for that shot. So assist eFG% is simply the mean of all of these expected eFG%'s.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
deepak
Joined: 26 Apr 2006
Posts: 665
PostPosted: Sun Oct 25, 2009 2:02 pm Post subject: Reply with quote
Thanks. Using that approach, could you not also estimate how many potential assists players had?
And I compared your efficiency numbers to the estimated ORTG from basketball-reference (only looked at players who weren't traded though).
It looks like the estimated ORTG tends to underrate guards:
Code:
Team Player min Ortg Usg% est_ORTG est_Usg% ORTG_diff Usg%_diff
1 ORL Jameer Nelson 1309 131 23.8 121 23.2 10 0.6
2 MIN Sebastian Telfair 2095 109 20.7 100 19.6 9 1.1
3 ATL Ronald Murray 1975 112 23.2 104 24.3 8 -1.1
4 DAL Antoine Wright 1552 110 14.8 102 15.6 8 -0.8
5 IND T.J. Ford 2258 115 24.9 108 23.5 7 1.4
6 LAC Baron Davis 2248 104 26.8 97 25.2 7 1.6
7 DEN Anthony Carter 1786 105 15.7 98 14.3 7 1.4
8 OKC Earl Watson 1776 100 18.5 93 16.3 7 2.2
9 DAL Jose Juan Barea 1600 113 21.9 106 20.2 7 1.7
10 ORL Anthony Johnson 1481 111 16.9 104 15.5 7 1.4
11 ATL Joe Johnson 3124 116 26.8 110 26.6 6 0.2
12 NOH Chris Paul 3002 130 30.7 124 27.5 6 3.2
13 CHI Derrick Rose 3000 114 23.1 108 22.6 6 0.5
14 WAS Caron Butler 2585 114 25.8 108 25.9 6 -0.1
15 DET Rodney Stuckey 2517 110 23.8 104 23 6 0.8
16 MIN Randy Foye 2494 112 22.7 106 22.6 6 0.1
17 MIL Ramon Sessions 2173 117 23.4 111 22.4 6 1
18 NJN Keyon Dooling 2074 118 17.7 112 17.4 6 0.3
19 PHI Lou Williams 1919 111 26.8 105 27.7 6 -0.9
20 CLE LeBron James 3054 127 34 122 33.8 5 0.2
And overrate bigs:
Code:
Team Player min Ortg Usg% est_ORTG est_Usg% ORTG_diff Usg%_diff
1 CHA Emeka Okafor 2691 104 18.7 112 18.9 -8 -0.2
2 CHI Joakim Noah 1938 115 13 123 12.3 -8 0.7
3 LAC Marcus Camby 1925 104 16 112 15.8 -8 0.2
4 MEM Darrell Arthur 1464 91 15.4 99 16.3 -8 -0.9
5 DET Jason Maxiell 1408 112 15 120 14.7 -8 0.3
6 LAL Sasha Vujacic 1293 107 15.2 115 16 -8 -0.8
7 ORL Tony Battie 1202 99 15.6 106 15.6 -7 0
8 PHI Reggie Evans 1140 95 13.7 102 13 -7 0.7
9 GSW Anthony Randolph 1129 95 21 102 21.7 -7 -0.7
10 MIA Udonis Haslem 2560 107 14.1 113 14.7 -6 -0.6
11 IND Troy Murphy 2482 114 16.8 120 16.9 -6 -0.1
12 CLE Anderson Varejao 2306 110 14.2 116 14.2 -6 0
13 MIL Charlie Villanueva 2095 100 27.8 106 28.5 -6 -0.7
14 MIN Kevin Love 2048 106 21 112 21 -6 0
15 PHI Samuel Dalembert 2036 97 13.5 103 13.6 -6 -0.1
16 POR Joel Przybilla 1952 118 10.5 124 10.2 -6 0.3
17 IND Jeff Foster 1828 109 12.8 115 12 -6 0.8
18 ATL Zaza Pachulia 1473 107 17.1 113 16.3 -6 0.8
19 NJN Yi Jianlian 1421 90 20.3 96 20.5 -6 -0.2
20 PHI Marreese Speights 1262 109 20.2 115 21.9 -6 -1.7
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 2:12 pm Post subject: Reply with quote
Quote:
... could you not also estimate how many potential assists players had?
I know there is a rule of thumb for potential assists (see here and here), so perhaps we can intelligently build this in to give assisters a better measure of failure instead of using the 1/2 multiplier for giving credit on assists.
Also, thanks for doing that comparison. I know B-R estimates things slightly differently than what Dean uses in BoP, so I want to at some point in the future compare these results with Dean's estimates to see 1) what differences there are and 2) see how much better one is at predicting than the other (if at all).
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
fundamentallysound
Joined: 18 Jul 2008
Posts: 25
Location: VA
PostPosted: Sun Oct 25, 2009 10:49 pm Post subject: Reply with quote
Ryan, this is great, great work. One quick thing that I just want to make sure I'm understanding properly. You, similarly to Eli Witus, found that there was a tradeoff for Usage and Efficiency. The magnitude of the trade off in your study was lower, though, than Eli's, is that correct?
So, if someone had previously been relying on the +/- 1.25 gain or drop in ORtg based on a decrease or increase in Usg% of 1%, they would be overvaluing high usage players relative to the numbers you came up with?
So, if someone were to use your numbers, instead, to standardize all ORtg's so that everyone was operating on 20% of possessions, high usage players would get a boost to their standardized ORtg of 1.11 pts/100 poss. for every 1% over 20% Usg they had?
Maybe it's improper to use these things in this way (i.e. the relationship between usg and efficiency might be more curved than linear), but I just wanted to make sure I was understanding the basic mechanics.
Sorry if that was confusing, sometimes I don't articulate my thoughts the best way, especially when it relates to this stuff where I am fairly novice.
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Sun Oct 25, 2009 11:27 pm Post subject: Reply with quote
I think your math is close. I estimate that each 1% increase over 100% in a lineup's usg% adds 0.11 points per hundred possessions. So we'd need a 9% increase in the lineup's usg% to expect a 1 point per hundred poss increase in efficiency for the lineup. Eli's estimate was 4% instead of 9%. The std error on his estimate is large enough to consider my estimate plausible. The bounds on my estimate are anywhere from 6% to 19%.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Mon Oct 26, 2009 9:39 am Post subject: Reply with quote
I believe Ryan's numbers suggest a 0.55 change in efficiency for every 1% change in usage as opposed to a 1.25 to 1 ratio that Eli's numbers suggested.
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Mon Oct 26, 2009 11:11 am Post subject: Reply with quote
I think that David was able to correctly articulate the estimated individual effect. Very Happy
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Fri Oct 30, 2009 4:41 pm Post subject: Reply with quote
I've now extracted the defensive efficiency ratings. The spreadsheet of the ratings can be found here. The data in the spreadsheet:
Code:
* Drtg: the player’s defensive efficiency rating
* Std Err: the standard error of the rating
* 95% CI: a 95% confidence interval for the rating
* Usg%: the percentage of possessions used by this player while on the court
* Total Used: the total number of possessions this player used
* %Shots: percentage of possessions used that were shots
* %Fouls: percentage of possessions used that were fouls
* %Drebs: percentage of possessions used that were defensive rebounds
* %Turnovers: percentage of possessions used that were turnovers
Full details can be found in the post.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
deepak
Joined: 26 Apr 2006
Posts: 665
PostPosted: Fri Oct 30, 2009 10:54 pm Post subject: Reply with quote
Ryan J. Parker wrote:
I've now extracted the defensive efficiency ratings. The spreadsheet of the ratings can be found here. The data in the spreadsheet:
Interesting stuff.
For assigning credit to shots, you may also consider basing it on floor location, or some combination of floor location and counterpart data. For instance, guards are more likely to be responsible for shots on the perimeter, and centers are more likely to be responsible for shots at the rim.
Re: More recovered assist threads
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sun Mar 13, 2005 10:28 am Post subject: Assisted % - Creating Your Own Shot? Reply with quote
As brought up in the "How well would this team perform?" thread, I wonder about the validity of using assisted%, at least in small part, to find out how well someone can create their own relatively high percentage shot.
I think this would be something very interesting to look at, but really don't have a clue how to measure it. I don't think it should be done on a league-wide basis, but rather, on a player by player basis. I think there's something here, but I just can't quite put my finger on it. Hopefully you guys can help me out.
Let's take a look at Kobe Bryant, for example.
Year - Assisted% - FG% - eFG%
0203 - 40% - 45.1% - 48.2%
0304 - 47% - 44.1% - 47.1%
0405 - 30% - 42.6% - 46.7%
Personally, I think this can tell you something about Kobe. It seems that his eFG% has only taken a slight dip with the huge drop off in assisted%.
Now let's look at Dirk.
0203 - 66% - 46.3% - 51.4%
0304 - 72% - 46.2% - 50.0%
0405 - 53% - 45.2% - 48.4%
Dirk also had a huge drop off in assisted% this last year. In fact, his pattern followed about the same pattern as Kobe's over the last 3 years - bump up in asst'd%, then huge drop off, with a steadying decline in efficiency. Both are regarded as stars and have still been able to maintain a fairly efficient presence while creating more for themselves.
But now let's look at Richard Jefferson, who is clearly not on the same level as these two.
0203 - 68% - 50.1% - 50.4%
0304 - 70% - 49.8% - 52.1%
0405 - 54% - 42.2% - 44.9%
All these three players' assisted percentages dropped off by at least 15%. However, with the first two superstars, eFG% dropped by at most 1.6%. With Jefferson? 7.2%.
What I'm trying to say is that I think there's something there. I think you can come up with something similar to Dean's efficiency curves, player by player. Instead of usage, look at how often they are assisted.
I know there are problems with what I've presented above - different factors you could look at as to what the difference is between those players. But I still think that something can be taken out of this.
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Mon Mar 14, 2005 11:31 am Post subject: Reply with quote
this is something i'm very interested in investigating further, but i'm not sure if there's enough data out there to accurately look at it.
specifically, i would love to be able to come up with a chart for each player that shows efficiency vs assisted% (with assisted% as the independent variable)....i guess this would be similar to Ch 19 of DeanO's book.
Back to top
View user's profile Send private message Send e-mail AIM Address
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 4:38 pm Post subject: Reply with quote
I'm looking at cutting up the play-by-plays and analyzing this on a game-by-game basis. Graphing Assisted% vs. 2pt-FG% and Assisted% vs. 3ptFG% per game, and trying to get a "skill curve" type readout from that.
2 things:
a) Do you think this would be worthwhile? Ways to improve this study?
b) Does anyone have an already written out list of the NBA schedule (in excel of some kind of format that I can cut up) so that I can easily get a list of links to the play-by-play?
Thanks for the help.
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Tue Mar 15, 2005 4:54 pm Post subject: Reply with quote
FFSBasketball wrote:
"Insanity: doing the same thing over and over again and expecting different results." -Albert Einstein
Hmm, isn't that what basketball teams do 82 times a year? Wink
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Ed Küpfer
Joined: 30 Dec 2004
Posts: 784
Location: Toronto
PostPosted: Tue Mar 15, 2005 5:31 pm Post subject: Reply with quote
FFSBasketball wrote:
Does anyone have an already written out list of the NBA schedule (in excel of some kind of format that I can cut up) so that I can easily get a list of links to the play-by-play?
For this season, no, but it's not difficult for me to put together. If, however, you'd want to use previous seasons, I have those ready to go. I even have the actual logs in HTML and ascii, although the latter is messy. Tell me exactly what you want. NBA.com, ESPN, Sportsline?
_________________
ed
Back to top
View user's profile Send private message Send e-mail
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 6:14 pm Post subject: Reply with quote
Well, I can fairly easily write a parser to pick apart the play-by-play logs and find what I want - the difficulties lie in directing the parser to the right place.
I know that at least this year, the play-by-play on the NBA website is at the address of the following formula:
http://scores.nba.com/games/yyyymmdd/Aw ... Print.html
So the hard part that I was looking for help on was getting the schedule, so that I could have a list of games with their date to direct the parser to.
If you have the actual play-by-play logs from previous years, that would be great. Once again though, the schedule would be necessary (unless you have the files in a numbered order, or something) so that the parser would know the file name to go to.
If you can help with any of this, that would be great.
P.S. I would prefer the NBA logs, if you could e-mail them to FFSBasketball ]at[ gmail ]dot[ com. Thanks a bunch.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 7:16 pm Post subject: Reply with quote
Also, before embarking on this study, are there any suggestions of a better way to do it, or something to look into while doing it?
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 784
Location: Toronto
PostPosted: Tue Mar 15, 2005 7:32 pm Post subject: Reply with quote
FFSBasketball wrote:
Also, before embarking on this study, are there any suggestions of a better way to do it, or something to look into while doing it?
Variables to keep track of:
* Assister
* Assistee
* Type of shot (layup, dunk, 3pt, etc.)
* Maybe the score and time at the attempt
I don't think you really need much more than that. Once you have these, there are all kinds of cool things you can do. I am particularly interested in applying DanR's Potential Assists calculations. Player-player hookups would be the sexy study, but I'm not so sure how informative it would be due to small samples. Using type of shot as a proxy for area on the floor, you'll be able to tell where players tend to get assisted, and where they make their own shots. I'm sure others can come up with better ideas than mine.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 8:18 pm Post subject: Reply with quote
OK, well this is a slightly different approach to what I was thinking, but it might make more sense, because it's more general, which means there's more data possible.
The only problem with what you presented is how to record the data. Should it just be in a list, delimited by commas or something?
So a sample would be:
Time,Score,Assistee,type of shot,assister
3:40,33-40,Odom,Running Jump,Atkins
How exactly would this help? It would seem we're back with the same problem as in the play-by-plays. That information is right in the play-by-plays, the key is how to categorize it.
What I was going to do was for each player have a column in the database for assisted%, and one for FG%. Delimit each game in those columns by commas, then be able to paste it into excel and make a graph for anyone you want of their assisted% vs. FG%.
The parser could always be run again for 2ptFG%, 3ptFG%, eFG%, or whatever measure you wish.
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 784
Location: Toronto
PostPosted: Tue Mar 15, 2005 8:28 pm Post subject: Reply with quote
FFSBasketball wrote:
So a sample would be:
Time,Score,Assistee,type of shot,assister
3:40,33-40,Odom,Running Jump,Atkins
How about:
Code:
Quarter Time Score Assistee TypeOfShot Assister
1 3:40 -7 Odom Running Jump Atkins
Score would be presented as the difference between the scoring team and the opposing team. Comma delim is fine.
FFSBasketball wrote:
What I was going to do was for each player have a column in the database for assisted%, and one for FG%. Delimit each game in those columns by commas, then be able to paste it into excel and make a graph for anyone you want of their assisted% vs. FG%.
Excel only allows 255 columns, so you won't be able to put many games together on one sheet. Hmm....
_________________
ed
Back to top
View user's profile Send private message Send e-mail
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 8:37 pm Post subject: Reply with quote
There are many ways to get around excel, so that's not that big of a deal (for example, just rotate the table so columns become rows, and you can have unlimited rows - you can do this with Paste Special -> Transpose).
I think I'll do my original study first, see how it turns out, and then I'll edit the parser to take data that anyone else requests. That shouldn't be too hard. I'll let you know when I have results.
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Tue Mar 15, 2005 9:38 pm Post subject: Reply with quote
i had envisioned the same thing, where it's sort of like the skill curves. put assisted% on the x-axis, and efg% on the y-axis.
Back to top
View user's profile Send private message Send e-mail AIM Address
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 11:29 pm Post subject: Reply with quote
First off, thanks to Ed for the help.
Well, I got the parser working, and cut up the Hawks' PBPs from last year.
I'll post Terry's and Stephen Jackson's (added Reef, Sura and Ratliff) "skill curves" (we'll have to come up with another name), and you can take from them what you want.
The x-axis is assisted%, and the y-axis is FG% (straight, not eFG%).
Just a disclaimer: I compared my totals, and they look a little off of the actual NBA totals, so I'll investigate what the problem is later. For now though, they're pretty accurate.
Stephen Jackson
Jason Terry
Added in Shareef's 53 games with ATL:
More - Sura:
And Ratliff:
Obviously this is only 5 players, so it's hard to judge, but it seems to me you can take some general patterns from them.
All of the first three are relatively stable, and can create their own fairly good looks (I think this was somewhat expected with this caliber of player). But Reef and Jackson trend up slightly, suggesting they are still better set up by a passer, while Terry, as a PG, seems more comfortable creating his own shot, as he has a slight (very slight) downward trend.
Note: Terry's year-by-year data supports this finding - he's more efficient creating for himself. The last two years his ast'd% has been around 45%, and his FG% around 42%. This year, his ast'd% has gone down to about 35%, while his FG% has risen to over 50%.
As for the last 2, Sura is hard to judge, because of a fairly small sample size, and he's all over the place.
Ratliff is, as expected, trending upwards. His best looks are not set up by his moves, but by the passer.
Back to top
View user's profile Send private message
S.K.
Joined: 18 Feb 2005
Posts: 61
Location: Toronto
PostPosted: Wed Mar 16, 2005 2:37 pm Post subject: Reply with quote
Interesting stuff, FFS.
Are you making any allowances for offensive putbacks/tips? I would think that this would significantly affect the results for a guy like Ratliff - it would technically count as "creating" the shot, but reflects rebounding skill rather than shot creation ability.
_________________
No books - no articles - no website.
Just opinions.
Ill-informed opinions.
Back to top
View user's profile Send private message Send e-mail MSN Messenger
S.K.
Joined: 18 Feb 2005
Posts: 61
Location: Toronto
PostPosted: Wed Mar 16, 2005 2:43 pm Post subject: Reply with quote
Wait, that didn't make sense - you're trying to use assisted% only as an indicator. I should think things through before I post.
Well, anyway, I think it would be worthwhile to try subtracting 'tips' from the unassisted FG, since they'll never be assisted.
_________________
No books - no articles - no website.
Just opinions.
Ill-informed opinions.
Back to top
View user's profile Send private message Send e-mail MSN Messenger
Display posts from previous:
Post new topic Reply to topic APBRmetrics
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sun Mar 13, 2005 10:28 am Post subject: Assisted % - Creating Your Own Shot? Reply with quote
As brought up in the "How well would this team perform?" thread, I wonder about the validity of using assisted%, at least in small part, to find out how well someone can create their own relatively high percentage shot.
I think this would be something very interesting to look at, but really don't have a clue how to measure it. I don't think it should be done on a league-wide basis, but rather, on a player by player basis. I think there's something here, but I just can't quite put my finger on it. Hopefully you guys can help me out.
Let's take a look at Kobe Bryant, for example.
Year - Assisted% - FG% - eFG%
0203 - 40% - 45.1% - 48.2%
0304 - 47% - 44.1% - 47.1%
0405 - 30% - 42.6% - 46.7%
Personally, I think this can tell you something about Kobe. It seems that his eFG% has only taken a slight dip with the huge drop off in assisted%.
Now let's look at Dirk.
0203 - 66% - 46.3% - 51.4%
0304 - 72% - 46.2% - 50.0%
0405 - 53% - 45.2% - 48.4%
Dirk also had a huge drop off in assisted% this last year. In fact, his pattern followed about the same pattern as Kobe's over the last 3 years - bump up in asst'd%, then huge drop off, with a steadying decline in efficiency. Both are regarded as stars and have still been able to maintain a fairly efficient presence while creating more for themselves.
But now let's look at Richard Jefferson, who is clearly not on the same level as these two.
0203 - 68% - 50.1% - 50.4%
0304 - 70% - 49.8% - 52.1%
0405 - 54% - 42.2% - 44.9%
All these three players' assisted percentages dropped off by at least 15%. However, with the first two superstars, eFG% dropped by at most 1.6%. With Jefferson? 7.2%.
What I'm trying to say is that I think there's something there. I think you can come up with something similar to Dean's efficiency curves, player by player. Instead of usage, look at how often they are assisted.
I know there are problems with what I've presented above - different factors you could look at as to what the difference is between those players. But I still think that something can be taken out of this.
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Mon Mar 14, 2005 11:31 am Post subject: Reply with quote
this is something i'm very interested in investigating further, but i'm not sure if there's enough data out there to accurately look at it.
specifically, i would love to be able to come up with a chart for each player that shows efficiency vs assisted% (with assisted% as the independent variable)....i guess this would be similar to Ch 19 of DeanO's book.
Back to top
View user's profile Send private message Send e-mail AIM Address
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 4:38 pm Post subject: Reply with quote
I'm looking at cutting up the play-by-plays and analyzing this on a game-by-game basis. Graphing Assisted% vs. 2pt-FG% and Assisted% vs. 3ptFG% per game, and trying to get a "skill curve" type readout from that.
2 things:
a) Do you think this would be worthwhile? Ways to improve this study?
b) Does anyone have an already written out list of the NBA schedule (in excel of some kind of format that I can cut up) so that I can easily get a list of links to the play-by-play?
Thanks for the help.
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Tue Mar 15, 2005 4:54 pm Post subject: Reply with quote
FFSBasketball wrote:
"Insanity: doing the same thing over and over again and expecting different results." -Albert Einstein
Hmm, isn't that what basketball teams do 82 times a year? Wink
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Ed Küpfer
Joined: 30 Dec 2004
Posts: 784
Location: Toronto
PostPosted: Tue Mar 15, 2005 5:31 pm Post subject: Reply with quote
FFSBasketball wrote:
Does anyone have an already written out list of the NBA schedule (in excel of some kind of format that I can cut up) so that I can easily get a list of links to the play-by-play?
For this season, no, but it's not difficult for me to put together. If, however, you'd want to use previous seasons, I have those ready to go. I even have the actual logs in HTML and ascii, although the latter is messy. Tell me exactly what you want. NBA.com, ESPN, Sportsline?
_________________
ed
Back to top
View user's profile Send private message Send e-mail
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 6:14 pm Post subject: Reply with quote
Well, I can fairly easily write a parser to pick apart the play-by-play logs and find what I want - the difficulties lie in directing the parser to the right place.
I know that at least this year, the play-by-play on the NBA website is at the address of the following formula:
http://scores.nba.com/games/yyyymmdd/Aw ... Print.html
So the hard part that I was looking for help on was getting the schedule, so that I could have a list of games with their date to direct the parser to.
If you have the actual play-by-play logs from previous years, that would be great. Once again though, the schedule would be necessary (unless you have the files in a numbered order, or something) so that the parser would know the file name to go to.
If you can help with any of this, that would be great.
P.S. I would prefer the NBA logs, if you could e-mail them to FFSBasketball ]at[ gmail ]dot[ com. Thanks a bunch.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 7:16 pm Post subject: Reply with quote
Also, before embarking on this study, are there any suggestions of a better way to do it, or something to look into while doing it?
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 784
Location: Toronto
PostPosted: Tue Mar 15, 2005 7:32 pm Post subject: Reply with quote
FFSBasketball wrote:
Also, before embarking on this study, are there any suggestions of a better way to do it, or something to look into while doing it?
Variables to keep track of:
* Assister
* Assistee
* Type of shot (layup, dunk, 3pt, etc.)
* Maybe the score and time at the attempt
I don't think you really need much more than that. Once you have these, there are all kinds of cool things you can do. I am particularly interested in applying DanR's Potential Assists calculations. Player-player hookups would be the sexy study, but I'm not so sure how informative it would be due to small samples. Using type of shot as a proxy for area on the floor, you'll be able to tell where players tend to get assisted, and where they make their own shots. I'm sure others can come up with better ideas than mine.
_________________
ed
Back to top
View user's profile Send private message Send e-mail
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 8:18 pm Post subject: Reply with quote
OK, well this is a slightly different approach to what I was thinking, but it might make more sense, because it's more general, which means there's more data possible.
The only problem with what you presented is how to record the data. Should it just be in a list, delimited by commas or something?
So a sample would be:
Time,Score,Assistee,type of shot,assister
3:40,33-40,Odom,Running Jump,Atkins
How exactly would this help? It would seem we're back with the same problem as in the play-by-plays. That information is right in the play-by-plays, the key is how to categorize it.
What I was going to do was for each player have a column in the database for assisted%, and one for FG%. Delimit each game in those columns by commas, then be able to paste it into excel and make a graph for anyone you want of their assisted% vs. FG%.
The parser could always be run again for 2ptFG%, 3ptFG%, eFG%, or whatever measure you wish.
Back to top
View user's profile Send private message
Ed Küpfer
Joined: 30 Dec 2004
Posts: 784
Location: Toronto
PostPosted: Tue Mar 15, 2005 8:28 pm Post subject: Reply with quote
FFSBasketball wrote:
So a sample would be:
Time,Score,Assistee,type of shot,assister
3:40,33-40,Odom,Running Jump,Atkins
How about:
Code:
Quarter Time Score Assistee TypeOfShot Assister
1 3:40 -7 Odom Running Jump Atkins
Score would be presented as the difference between the scoring team and the opposing team. Comma delim is fine.
FFSBasketball wrote:
What I was going to do was for each player have a column in the database for assisted%, and one for FG%. Delimit each game in those columns by commas, then be able to paste it into excel and make a graph for anyone you want of their assisted% vs. FG%.
Excel only allows 255 columns, so you won't be able to put many games together on one sheet. Hmm....
_________________
ed
Back to top
View user's profile Send private message Send e-mail
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 8:37 pm Post subject: Reply with quote
There are many ways to get around excel, so that's not that big of a deal (for example, just rotate the table so columns become rows, and you can have unlimited rows - you can do this with Paste Special -> Transpose).
I think I'll do my original study first, see how it turns out, and then I'll edit the parser to take data that anyone else requests. That shouldn't be too hard. I'll let you know when I have results.
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Tue Mar 15, 2005 9:38 pm Post subject: Reply with quote
i had envisioned the same thing, where it's sort of like the skill curves. put assisted% on the x-axis, and efg% on the y-axis.
Back to top
View user's profile Send private message Send e-mail AIM Address
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Tue Mar 15, 2005 11:29 pm Post subject: Reply with quote
First off, thanks to Ed for the help.
Well, I got the parser working, and cut up the Hawks' PBPs from last year.
I'll post Terry's and Stephen Jackson's (added Reef, Sura and Ratliff) "skill curves" (we'll have to come up with another name), and you can take from them what you want.
The x-axis is assisted%, and the y-axis is FG% (straight, not eFG%).
Just a disclaimer: I compared my totals, and they look a little off of the actual NBA totals, so I'll investigate what the problem is later. For now though, they're pretty accurate.
Stephen Jackson
Jason Terry
Added in Shareef's 53 games with ATL:
More - Sura:
And Ratliff:
Obviously this is only 5 players, so it's hard to judge, but it seems to me you can take some general patterns from them.
All of the first three are relatively stable, and can create their own fairly good looks (I think this was somewhat expected with this caliber of player). But Reef and Jackson trend up slightly, suggesting they are still better set up by a passer, while Terry, as a PG, seems more comfortable creating his own shot, as he has a slight (very slight) downward trend.
Note: Terry's year-by-year data supports this finding - he's more efficient creating for himself. The last two years his ast'd% has been around 45%, and his FG% around 42%. This year, his ast'd% has gone down to about 35%, while his FG% has risen to over 50%.
As for the last 2, Sura is hard to judge, because of a fairly small sample size, and he's all over the place.
Ratliff is, as expected, trending upwards. His best looks are not set up by his moves, but by the passer.
Back to top
View user's profile Send private message
S.K.
Joined: 18 Feb 2005
Posts: 61
Location: Toronto
PostPosted: Wed Mar 16, 2005 2:37 pm Post subject: Reply with quote
Interesting stuff, FFS.
Are you making any allowances for offensive putbacks/tips? I would think that this would significantly affect the results for a guy like Ratliff - it would technically count as "creating" the shot, but reflects rebounding skill rather than shot creation ability.
_________________
No books - no articles - no website.
Just opinions.
Ill-informed opinions.
Back to top
View user's profile Send private message Send e-mail MSN Messenger
S.K.
Joined: 18 Feb 2005
Posts: 61
Location: Toronto
PostPosted: Wed Mar 16, 2005 2:43 pm Post subject: Reply with quote
Wait, that didn't make sense - you're trying to use assisted% only as an indicator. I should think things through before I post.
Well, anyway, I think it would be worthwhile to try subtracting 'tips' from the unassisted FG, since they'll never be assisted.
_________________
No books - no articles - no website.
Just opinions.
Ill-informed opinions.
Back to top
View user's profile Send private message Send e-mail MSN Messenger
Display posts from previous:
Post new topic Reply to topic APBRmetrics
Re: More recovered assist threads
Author Message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Sat Aug 06, 2005 11:31 am Post subject: passing statistics Reply with quote
I'm a fan who values passing as one of the most important, and often overlooked, skills in basketball. I was wondering if there are any good passing statistics out there. I think something as simple as adjusted assists for a player's usage rating could give a decent idea of how good players are at passing the ball which leads to a score. If there aren't any good statistics out there I might try to come up with a formula myself. Thanks.
Back to top
View user's profile Send private message
jeffpotts77
Joined: 18 Feb 2005
Posts: 150
Location: Cambridge, MA
PostPosted: Sat Aug 06, 2005 11:44 am Post subject: link to knickerblogger.net Reply with quote
You should check out Knickerblogger's Stats page. Below I've pasted the link to the page with all players sorted by Assist Ratio - a formula developed by John Hollinger, used to evaluate passing skills (40 best; 2 worst).
http://www.knickerblogger.net/stats/jh_ALL_AST.htm
Perhaps someone else will be able to elaborate on what the formula consists of.
Back to top
View user's profile Send private message Yahoo Messenger
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Sat Sep 03, 2005 5:30 pm Post subject: Reply with quote
Favors guards too much. Anyone else?
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Sat Sep 03, 2005 6:21 pm Post subject: Re: passing statistics Reply with quote
mateo82 wrote:
I'm a fan who values passing as one of the most important, and often overlooked, skills in basketball. I was wondering if there are any good passing statistics out there. I think something as simple as adjusted assists for a player's usage rating could give a decent idea of how good players are at passing the ball which leads to a score. If there aren't any good statistics out there I might try to come up with a formula myself. Thanks.
Pointing out the obvious, but Bob Chaikin's database has how often players pass the ball. It doesn't quantify how good the passes are, but he has done work to show that it's pretty accurate.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sat Sep 03, 2005 8:43 pm Post subject: Reply with quote
So I took the numbers I have from Bob's database from last year, and looked at percent of passes that ended up in assists.
The results are surprising, although perhaps not so telling.
The range of the entire dataset (limited to players playing more than 1000 minutes last year) was 15.16%-17.95%, with 177 of the 244 players being over 17%. In fact, take away Danny Fortson, and the range becomes 16.37%-17.95%.
Big men head the list, and someone like Nash ends up at #105 in the league, at 17.15%. So I'm not sure this is the greatest of metrics.
Top 10:
Code:
Eddy Curry - 0.179514051
Emeka Okafor - 0.178372352
Antonio Davis - 0.176453202
Marc Jackson - 0.176385124
Rasheed Wallace - 0.176333138
Erick Dampier - 0.176092811
Troy Murphy - 0.176043557
Donyell Marshall - 0.175356558
Tayshaun Prince - 0.175127636
Stromile Swift - 0.175104248
Back to top
View user's profile Send private message
bchaikin
Joined: 27 Jan 2005
Posts: 686
Location: cleveland, ohio
PostPosted: Sat Sep 03, 2005 10:45 pm Post subject: Reply with quote
So I took the numbers I have from Bob's database from last year, and looked at percent of passes that ended up in assists. The results are surprising, although perhaps not so telling.
not surprising at all - again in the stats database (and the simulation) a player's number of passes are calculated based on his assists, not vice versa. there are no players with higher percentages of assists per pass. any discrepencies you see between players are simply a factor of his team's game pace or some other small factor - there is very little difference between players ratios of assists to passes, as you can see from your calculations...
hopefully 82games.com will at some point show which players (if there are any) consistently show a higher (or lower) percentage of assists per pass, especially if many games are charted this upcoming season...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sat Sep 03, 2005 11:48 pm Post subject: Reply with quote
That would explain that.
Here's a hopefully more useful metric, combining data from 82games.com and Bob's database.
Top 15 Passing TOs/Pass:
Code:
Antonio Daniels 2.106251521
Matt Bonner 2.168063984
Jason Hart 2.592307439
Fred Hoiberg 2.607174945
PJ Brown 2.608071983
Aaron McKie 2.817962288
Elton Brand 2.848053498
Earl Boykins 2.854015803
Greg Buckner 2.942881234
Eric Snow 2.972836062
Kevin Garnett 2.975945433
Brendan Haywood 2.982163679
Tyronn Lue 2.995493736
Jeff McInnis 3.16151304
Brad Miller 3.162522656
Again, limited to players playing over 1000 minutes, and doesn't show that many surprises. Antonio Daniels is up there, I suppose PJ Brown, Elton Brand and Brendan Haywood could be called surprises, especially since they're big men.
If you wanted to do something to include the fact that someone might be taking a more risky pass if they're setting up a teammate for an assist, and so weight it for assists, we could do something like (1-passTO%)*(AST/MIN). Here's how that looks:
Code:
Steve Nash 32.2631855
Brevin Knight 29.42382862
Jason Kidd 21.56630353
Maurice Williams 20.56280182
Rick Brunson 20.31460208
Stephon Marbury 19.68436496
Jason Williams 19.65336655
Andre Miller 19.18777592
Sam Cassell 19.09896538
Earl Watson 19.04008207
Jason Hart 18.94468636
Jamaal Tinsley 18.68503728
Rafer Alston 18.20340527
Luke Ridnour 18.13249877
Gary Payton 17.86823422
No surprises here, I don't think. Almost all guards (that's who gets the assists), with the best big man checking in at #40 (KG). This really doesn't differentiate too much between itself and straight assists per minute (Steve Francis moves up a couple spots, but that's really the only dramatic movement).
What else could we include? How would you go about constructing a "passer rating"?
Back to top
View user's profile Send private message
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
PostPosted: Sun Sep 04, 2005 11:14 am Post subject: Reply with quote
This is an area that will be greatly helped by the game charting this season, but we also do post a 'passing rating' for players which is really just a placeholder until we get something better!
Pass Rtg = (3pt AST + 2pt Jump AST + (2*Close AST)+ (4*Dunk AST) - (2* Pass T/O) ) / (seconds played / 2880)
I think the formula was concocted in about 10 seconds..the object obviously being to reward the passer more heavily for close/dunk asssists under the theory that the pass was a bigger contributor to the made bucket.
Under this scheme last year's leaders
24.6 - Nash
17.9 - Knight
16.2 - A.Miller
14.1 - B.Davis (GS, he was 13.8 in NO minutes)
13.7 - Jason Williams (MEM)
13.3 - Maurice Williams (MIL)
12.9 - Arroyo (UTA, 11.9 with DET)
12.5 - Kidd
12.2 - Livingston (LAC)
12.2 - Watson
12.0 - Francis
11.7 - Brunson
11.3 - Stoudamire (POR)
11.3 - Duhon
11.3 - Snow
11.2 - Iverson
11.2 - Boykins
10.8 - Marbury
10.8 - Payton
10.7 - Claxton (NO, 15.3 in lim min with GS)
10.6 - Terry
10.6 - Jaric
10.5 - Tinsley
10.5 - Carter (MIN)
10.5 - Lue
10.4 - Hinrich
10.3 - Ridnour
10.3 - Cassell
10.2 - Wade
10.2 - Telfair
10.2 - Hart
10.0 - LeBron James
...and the worst rated
-1.0 - Collier
-1.0 - Dalembert
-0.6 - Henderson (DAL)
-0.3 - Gadzuric
-0.1 - Fortson
0.0 - Bradley (DAL)
0.1 - Mutombo
0.1 - Battie
0.1 - Curry
0.1 - Hunter
0.1 - Wilcox
0.2 - Mohammad (NY)
0.3 - Pollard
0.3 - Sweetney
0.3 - Jerome James
0.3 - Okafor
it would probably be worth ranking by position...
"PG"
24.6 - Nash
17.9 - Knight
16.2 - A.Miller
14.1 - B.Davis
13.7 - Williams (MEM)
...
4.8 - Vaughn
4.1 - Best
"PG/SG"
12.0 Francis
11.7 Brunson
11.3 Snow
11.2 Iverson
10.6 Terry
...
1.4 Evans (SAC)
1.3 Snyder (UTA)
"SG/SF"
10.0 James (CLE)
7.7 Hamilton
7.0 Christie (SAC)
6.5 Cater (TOR)
6.1 Bryant
...
1.1 J.Jones (IND)
0.7 Lewis (SEA)
"SF/PF"
9.2 Kukoc
7.4 Garnett
6.8 Webber (SAC)
5.5 Kirilenko
5.3 Walker (BOS)
...
0.3 Sweetney
0.1 Wilcox
"PF/C"
6.0 Brad Miller
4.8 Shaq
4.4 Collins (UTA)
4.4 Camby
4.0 Duncan
...
-1.0 Dalembert
-1.0 Collier
Back to top
View user's profile Send private message Visit poster's website
KD
Joined: 30 Jan 2005
Posts: 163
PostPosted: Sun Sep 04, 2005 4:15 pm Post subject: Off-topic, my apologies ... Reply with quote
From a Ric Bucher piece on Rick Brunson ...
Quote:
"His assists per minutes played is one of the best in the league," Sund said. "What we like is that there is no unknown with him. You know what you're getting. He knows who he is."
http://sports.espn.go.com/nba/columns/story?id=2149948
Not entirely true, but Brunson's assist ratio was third in the league last year.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sun Sep 04, 2005 9:21 pm Post subject: Reply with quote
Roland_Beech wrote:
the object obviously being to reward the passer more heavily for close/dunk asssists under the theory that the pass was a bigger contributor to the made bucket.
I'm slightly conflicted about this. On the one hand, I agree that an assist for a dunk is usually set up almost exclusively by the passer, and the finisher has to have very little ability to be able to finish. On the other hand, since an assist means that you have a guaranteed made shot, then shooting percentages are improved more by assists on jump shots.
That is, since the farther away from the basket you are, the harder the shot, then a great set up on the outside may be worth more (especially if it's for 3) than a great set up inside. Your chances have improved much more, relatively.
Back to top
View user's profile Send private message
kjb
Joined: 03 Jan 2005
Posts: 864
Location: Washington, DC
PostPosted: Mon Sep 05, 2005 10:21 am Post subject: Reply with quote
I also question giving a bonus for "close" assists. Why should a pass to cutter for an open dunk have more value in a rating system than say a big man passing out of a double team for an open 3? In the game, the open 3 is worth more, and the play (pass out and making the open 3) probably does more to change the way a defense plays. In this rating system, a pass-ahead to a wide-open teammate on the break, who makes an uncontested dunk has 4 times more value than that pass out of a double team. That doesn't seem right to me.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
PostPosted: Mon Sep 05, 2005 11:12 am Post subject: Reply with quote
Well as I mentioned I think the formula was about 10 seconds worth of thought so it's not without drawbacks...
However, I do think inside assists are worth more than outside shot assists (since we can't tell how open a guy is at this stage). Part of it is that often the degree of difficulty on the pass is much harder (threading through defenders, etc) and part of it requires more understanding I believe of player movement (eg anticipating both your own teammate and the defenders).
I agree that a big man passing to a wide open perimeter player out of a double team can be valuable, but even more valuable is often the pass to a cutter down the lane in such a situation.
Anyway I would completely agree that passing stats are currently inadequate...that's one reason why this coming season should be exciting.
Back to top
View user's profile Send private message Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Mon Sep 05, 2005 11:33 am Post subject: Reply with quote
Roland_Beech wrote:
Well as I mentioned I think the formula was about 10 seconds worth of thought so it's not without drawbacks...
However, I do think inside assists are worth more than outside shot assists (since we can't tell how open a guy is at this stage). Part of it is that often the degree of difficulty on the pass is much harder (threading through defenders, etc) and part of it requires more understanding I believe of player movement (eg anticipating both your own teammate and the defenders).
I agree that a big man passing to a wide open perimeter player out of a double team can be valuable, but even more valuable is often the pass to a cutter down the lane in such a situation.
Anyway I would completely agree that passing stats are currently inadequate...that's one reason why this coming season should be exciting.
Frankly, I thought Roland's passing stats were the most representative of passing ability that I'd seen.
And, though the weights he used were derived out of the ether, they do represent a concept that does generally make sense. Defenses take away the inside more than jump shots. There should be modification so that 3pt jump shots get more weight than 2 pt jump shots, but that's relatively minor.
In general, doing this much with assists is pretty impressive. Assists are a measure of teamwork and teamwork is what I still think is the toughest thing we have to deal with. If a big man passes to a 3pt shooter who is then helped on by another guard, but swings it to another guy for an open 3pt shot -- both the big man and that perimeter guy did important things. (I'm not advocating 2 assists because I think that normally 1 is enough.) We already know about passes to guys who get fouled for 2 FT, which could get an assist and modify our interpretation of players a fair amount. We know that Euros don't give as many assists and that rules for giving assists have changed in the NBA. Assists are hard. We can do a lot better through a project like that at 82games, but that may not help interpret the past until we've studied things for many years. And, a lot of times, you can't forecast the future without understanding the past. So Roland's megaproject is going to be very very useful in going forward. (I think Dan's work has the capability of explaining assists, too. But I don't know what weight he ends up with -- my own personal lack of knowledge -- and, further, I don't think it readily allows us to infer their value in other leagues, at different distances, or if they were to start giving assists on free throws, I don't think we can easily know how that would change the weight.)
Always lots of questions on teamwork stuff...
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Mon Sep 05, 2005 2:34 pm Post subject: Reply with quote
Here's my attempt at a similar rating to Roland's, except based on the premise that it's better to assist on a jump shot than a close shot, for the reasons I mentioned above:
"Since the farther away from the basket you are, the harder the shot, then a great set up on the outside may be worth more than a great set up inside. Your chances have improved much more, relatively."
I took data from Roland's "Is Outside Shooting a Lost Art?" article, and decided to look at what players shot from each range on unassisted shots.
So for example, we see that people shot 90.8% on dunks over the period of data Roland analyzed, and 75% of those were assisted. That means that of the 25% that were not assisted, approximately 71.2% went in. Therefore the difference between an assisted dunk's chances of going in and an un-assisted dunk's chances are 28.8%, since an assist means a guaranteed bucket. (I'm sure this is flawed in some way, but I think that we can still measure relative values of each assist this way).
I did that with Jump and Close shots, also. So my formula becomes (((1-j%)*(jump+(1.5*3PT)))+((1-c%)*close)+((1-d%)*dunk))*(1-passTO%), and then divide that all by minutes played, and multiply by 400, just to make it look nicer (also based on a per 40 min basis, although I'm not sure it's measuring anything that would need to be quantified per 40 min).
Edit: Added a multiplier to three pointers to count them more.
j% = unassisted jump shot chance
c% = unassisted close shot chance
d% = unassisted dunk shot chance
passTO% = percentage of passes that ended in TOs
I then compared Roland's results to mine. Here's how it looks.
Roland's Top 10:
Code:
Roland's Rank - My Rank - Player
1 - 1 - Steve Nash
2 - 2 - Brevin Knight
3 - 32 - Andre Miller
4 - 13 - Jason Williams
5 - 12 - Maurice Williams
6 - 7 - Tyronn Lue
7 - 3 - Jason Kidd
8 - 15 - Earl Watson
9 - 34 - Steve Francis
10 - 10 - Rick Brunson
My Top 10:
Code:
My Rank - Roland's Rank - Player
1 - 1 - Steve Nash
2 - 2 - Brevin Knight
3 - 7 - Jason Kidd
4 - 50 - Milt Palacio
5 - 22 - Luke Ridnour
6 - 16 - Stephon Marbury
7 - 6 - Tyronn Lue
8 - 26 - Jason Hart
9 - 36 - Rafer Alston
10 - 10 - Rick Brunson
As you can see, there are some fairly significant differences. Nothing extreme, but enough to mean they are measuring very different things.
Milt Palacio is the biggest outlier on these lists, and that's because he had 234 "outside" assists, as opposed to 45 "inside" ones.
And here are the biggest differences:
Code:
My Rank - Roland's Rank - Player
126 - 205 - Mickael Pietrus
147 - 71 - Carmelo Anthony
166 - 91 - DeShawn Stevenson
99 - 173 - Ben Gordon
81 - 140 - Richard Jefferson
136 - 189 - Tim Thomas
213 - 161 - Udonis Haslem
87 - 135 - Eddie Gill
152 - 104 - Nene
4 - 50 - Milt Palacio
Comments?
Back to top
View user's profile Send private message
THWilson
Joined: 19 Jul 2005
Posts: 164
Location: phoenix
PostPosted: Mon Sep 05, 2005 4:53 pm Post subject: Reply with quote
FFSBasketball:
This doesn't make sense to me. I think the problem is that you are ignoring all the potential assists that didn't result in made baskets. For example here are some dummy values:
Assisted Dunk FG % = 80%
Unassisted Dunk FG % = 80%
Assisted Jump FG % = 40%
Unassited Jump FG % = 40%
In this scenario, a jump shot assist is worth three-times a dunk assist. But, potential assists don't help shooting percentage at all for jump shots or dunks. That doesn't make sense. Your model also seems to assume that it's equally easy to get an unassisted or assisted dunk. That rings false.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Mon Sep 05, 2005 5:16 pm Post subject: Reply with quote
All good points. I of course couldn't include potential assists since we don't have data on that.
Note my comment that "I'm sure this is flawed in some way". How would you go about fixing it using similar data and a similar premise?
Back to top
View user's profile Send private message
Carlos
Joined: 21 Jan 2005
Posts: 64
Location: Montevideo, Uruguay
PostPosted: Mon Sep 05, 2005 9:12 pm Post subject: Reply with quote
FFSBasketball, I think the flaw may be that you are assuming that the "alternative shot" for an assisted dunk would be an unassisted dunk. In fact, dunks are relatively rare shots and the most probable alternative shot for an assisted dunk would be a jump shot. The rating could look better if you assumed that the FG% of the alternative shot to every assisted shot (dunk, close or jump shot) is the FG% of ALL unassisted shots.
Back to top
View user's profile Send private message
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Tue Sep 06, 2005 4:28 am Post subject: Reply with quote
FFSBasketball wrote:
So for example, we see that people shot 90.8% on dunks over the period of data Roland analyzed, and 75% of those were assisted. That means that of the 25% that were not assisted, approximately 71.2% went in. Therefore the difference between an assisted dunk's chances of going in and an un-assisted dunk's chances are 28.8%, since an assist means a guaranteed bucket. (I'm sure this is flawed in some way, but I think that we can still measure relative values of each assist this way).
Echoing THWilson's comment, there's a big thing missing here: yes, all assists imply that a bucket was scored. But you're ignoring the passes that were made which led to shots which were MISSED. Which meant no assist. But if the shot had been made, there would've been an assist.
In other words, you're giving a free lunch to the passer who throws the ball to, say, Iverson who's 25 feet from the basket. Iverson lets fly, if the shot goes in, the passer gets an assist and you're assuming that his pass led to a "guaranteed bucket".
But that bucket was not guaranteed. In fact, Iverson misses more often than he connects, and you're not counting the passes thrown to Iverson which led to missed shots rather than assists.
To apply this to your dunk example: you're assigning "blame" for all the missed dunks to the unassisted shooters, and assuming that the recipients of assists had 100% success and no misses. But some of those missing dunkers (an unknown percent) were the recipients of passes, passes which would've led to an assist if the dunker hadn't missed.
With dunks this may not be a big deal because most dunks are made anyway. But outside jumpshots and especially 3-pointers have a high rate of misses.
Iverson averages 31% on his 3-pointers. Some of them are assisted, some are unassisted, but we can't assume that the assisted ones were "guaranteed buckets". The key questions are these two: what was his 3pt FG% on unassisted attempts? And what was his 3pt FG% on attempts which *could have* led to assists? The difference between those two is the measure which tells us how much value added the assister gave.
Unfortunately, due to inadequate stats-keeping by the NBA, we don't know Iverson's unassisted and assisted FG%s are. But I can tell you for sure, they're not 100% and it's unsound to assume that they are.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Wed Sep 07, 2005 9:42 am Post subject: Reply with quote
I can't tell if it's been said in roundabout ways already, but the FG% of a "potential assist" is built in to the assist total.
A pass to a guy within dunking distance is going to be an assist about twice as likely as a pass to an open 3-pt shooter. So there's no real point (that I can see) in doubling credit for the passer in the former instance.
Assigning weights for the closeness of the shots is probably just distinguishing teams that put more dunkers or shooters on the floor.
Back to top
View user's profile Send private message Send e-mail
HoopStudies
Joined: 30 Dec 2004
Posts: 706
Location: Near Philadelphia, PA
PostPosted: Wed Sep 07, 2005 11:23 am Post subject: Reply with quote
Mike G wrote:
I can't tell if it's been said in roundabout ways already, but the FG% of a "potential assist" is built in to the assist total.
A pass to a guy within dunking distance is going to be an assist about twice as likely as a pass to an open 3-pt shooter. So there's no real point (that I can see) in doubling credit for the passer in the former instance.
Assigning weights for the closeness of the shots is probably just distinguishing teams that put more dunkers or shooters on the floor.
Not true, actually.
Yes, a higher percentage of dunks will be assisted on that mid-range js. But it is a lot harder to get an assist to a player dunking the ball because the D takes that away first.
Passers take bigger risks getting the ball to a guy near the basket than to a jump shooter.
We give bigger weight to people who do well on their SAT than on their 4th grade addition test. The SAT is tougher. If someone aces 200 4th grade addition tests, but gets a 700 on their SAT, how do we feel he ranks relative to a guy who did ok on his 4th grade addition tests, but then got a 1500 on their SAT?
We say "nice pass" when it's to a player in great scoring position through traffic, not when it's a floater to a wide open big man at the 3pt line. The pass is increasing the likelihood of scoring in the first case, but not in the 2nd. Because of the bias in not counting potential assists, you may say it balances out. But it doesn't. It may take one player 300 passes to get 100 perimeter assists. It may take another 150 passes to get 100 inside assists. We don't know from the data. But we know it should take fewer to get the inside assists. Giving conceptual credit as Roland has (or as I do in my work) gets us closer to measuring an efficiency of passing.
It's a philosophical question, which means that this thread could go on waaaayyy too long for me. But I am comfortable that more weight should go on the pass for close shots.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Wed Sep 07, 2005 12:40 pm Post subject: Reply with quote
HoopStudies wrote:
...
We say "nice pass" when it's to a player in great scoring position through traffic, not when it's a floater to a wide open big man at the 3pt line. .... It may take one player 300 passes to get 100 perimeter assists. It may take another 150 passes to get 100 inside assists. ...
Well, I think that was exactly what I was saying. Or to use different numbers, 300 passes to the outside may lead to 100 assists, while 300 passes to the inside may get you 200 assists.
These number make sense, intuitively and statistically. Neglecting for the moment that more of the 'outside' passes will lead to 3 points , and that more of the inside passes will lead to turnovers, the inside-passing player gets twice as many assists for the same number of passes; and that seems to be just about what his credit should be, relatively speaking.
Inside players who are a scoring threat are the ones who open up the outside players; if they also find them with timely passes, they get assists on something less than half their passes.
It seems a point guard who runs with a bunch of dunkers is going to get a good % of his assists with dunks, but I don't see why he should be credited with more than the actual number of baskets created. If Steve Nash gets 12 assists by setting up lots of open jumpshots, that's just as good (I think) as 12 assisted dunks.
I'm pretty sure the value of a basketball play is in how many points it produces, and not in how difficult it is.
Back to top
View user's profile Send private message Send e-mail
HoopStudies
Joined: 30 Dec 2004
Posts: 706
Location: Near Philadelphia, PA
PostPosted: Wed Sep 07, 2005 12:58 pm Post subject: Reply with quote
Mike G wrote:
HoopStudies wrote:
...
We say "nice pass" when it's to a player in great scoring position through traffic, not when it's a floater to a wide open big man at the 3pt line. .... It may take one player 300 passes to get 100 perimeter assists. It may take another 150 passes to get 100 inside assists. ...
Well, I think that was exactly what I was saying. Or to use different numbers, 300 passes to the outside may lead to 100 assists, while 300 passes to the inside may get you 200 assists.
These number make sense, intuitively and statistically. Neglecting for the moment that more of the 'outside' passes will lead to 3 points , and that more of the inside passes will lead to turnovers, the inside-passing player gets twice as many assists for the same number of passes; and that seems to be just about what his credit should be, relatively speaking.
Inside players who are a scoring threat are the ones who open up the outside players; if they also find them with timely passes, they get assists on something less than half their passes.
It seems a point guard who runs with a bunch of dunkers is going to get a good % of his assists with dunks, but I don't see why he should be credited with more than the actual number of baskets created. If Steve Nash gets 12 assists by setting up lots of open jumpshots, that's just as good (I think) as 12 assisted dunks.
I'm pretty sure the value of a basketball play is in how many points it produces, and not in how difficult it is.
Not how difficult the whole play is, but how difficult the pass is relative to the shot. It's a lot easier to make a dunk than a jumper. You should give more credit to the passer on a dunk because the pass is harder than on a jumper where the pass (and creation) is easier.
Conceptually you can watch any play and identify what are key points in making the play work. That's what analysts try to do. Sometime it is just making the shot, creating your own shot. An assist is an attempt to show that there may be other keys. That's why we talk about measuring picks, too, because they may be key. But some picks and some assists are worth relatively more in generating that score. As assist rules have loosened up, what I'm saying is that the credit per assist has gone down because their job to get an assist has gotten easier. It's a theoretical framework that works in interpreting assists from other leagues and from the past. Getting the data to parameterize it is more difficult.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Wed Sep 28, 2005 4:55 pm Post subject: Reply with quote
FFSBasketball wrote:
No surprises here, I don't think. Almost all guards (that's who gets the assists), with the best big man checking in at #40 (KG). This really doesn't differentiate too much between itself and straight assists per minute (Steve Francis moves up a couple spots, but that's really the only dramatic movement).
What else could we include? How would you go about constructing a "passer rating"?
Yeah, that's the problem. Taking assists into account assumes that #1 your teammate can catch the ball and #2 he can score the ball. A big man could be an excellent passer, but if his guards can't make jumpshots he doesn't get the assists. I'm looking for something that doesn't take teammates into account at all, if possible. The search continues.
Back to top
View user's profile Send private message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Wed Sep 28, 2005 5:03 pm Post subject: Reply with quote
HoopStudies wrote:
In general, doing this much with assists is pretty impressive. Assists are a measure of teamwork and teamwork is what I still think is the toughest thing we have to deal with. If a big man passes to a 3pt shooter who is then helped on by another guard, but swings it to another guy for an open 3pt shot -- both the big man and that perimeter guy did important things.
This is exactly my point. Plenty of times someone created a play with a good pass to a player who then passes it to someone else who is wide open. Only the second player gets credit with the assist, although the 1st player might have been the one who made the play happen. In the triangle offense this happens often.
This is why I don't like the assist approach. A pass that immediately leads to a score is only 1 of many types of good passes.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Thu Sep 29, 2005 8:30 am Post subject: Reply with quote
Here again, passer #1 may or may not get the assist; passer #2 has the option of shooting or passing (potential for points or assist, but not both). It doesn't seem reasonable that over the course of a season passer#1 would predominantly pass to guys who don't shoot. In the long run, everyone gets the # of assists they should get, based on the # of good passes they make.
That's relative to the team, of course. If you are on a poor-shooting team, or with weak finishers around the basket, you won't get as many assists (Mateo's point). Magic had great shooters and finishers. Stockton was just a great passer. Or are those just subjective notions?
Subtracting Stockton's FG/FGA from his team's FG/FGA would give you 'everyone else's' FG%. How much of the Jazz' high FG% was due to the great passing of Stockton? How much of Stock's assist rate was due to his great finishers?
Players who bounce around from team to team may see their assist rates bounce, too. The Brevin Knight instance has been pointed out as a case of "system assists". Whether it's by coaching design or by teammate strengths, statistical reward tends to be team-specific. Playoff appearances, championships, etc, will always be beyond the control of the individual. Some stats will be, as well.
Back to top
View user's profile Send private message Send e-mail
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Thu Sep 29, 2005 3:24 pm Post subject: Reply with quote
Mike G wrote:
Here again, passer #1 may or may not get the assist; passer #2 has the option of shooting or passing (potential for points or assist, but not both). It doesn't seem reasonable that over the course of a season passer#1 would predominantly pass to guys who don't shoot. In the long run, everyone gets the # of assists they should get, based on the # of good passes they make.
Not necessarily. This seems like a guard-centric statement Wink . The player who "makes the play" so to speak, or the player who draws the defense is going to get more assists because he has given a teammate an easier shot. That doesn't mean he's a better passer though, just a better playmaker. Forwards and centers are always going to be at a disadvantage when you use assists as your starting point. They're already close to the basket, there isn't a lot they can do, in quick movements, to draw in the opposing defense. Except for cutters, the big men are more likely to pass out to someone who's open. Sometimes that player will take the shot, sometimes that player will make the extra pass to someone who's even more open. I can't think of a whole lot of circumstances where guards make passes that lead to extra passes. It's not a regular feature of the game.
I think this is why the attempts in this thread to construct a list of the best passers has been so guardcentric. Although I think guards are in general the better passers, not in this dominant of a degree. I'd take Kevin Garnett and probably a couple of other big men's passing over the vast majority of guards.
Back to top
View user's profile Send private message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Thu Sep 29, 2005 3:25 pm Post subject: Reply with quote
I think the best way to resolve this is to wait out this season. Then we'll have more data and we can use turnovers caused by bad passes as a better gauge (or at least part of a bigger formula) for passing ability.
Back to top
View user's profile Send private message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Wed Nov 16, 2005 7:11 pm Post subject: Re: passing statistics Reply with quote
HoopStudies wrote:
Pointing out the obvious, but Bob Chaikin's database has how often players pass the ball. It doesn't quantify how good the passes are, but he has done work to show that it's pretty accurate.
Do you know where I can find (or learn about) this database? I would love to see this. We can combine this with the usage rating and get a "ball hog" stat. Smile
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Thu Nov 17, 2005 7:12 am Post subject: Reply with quote
Bob's thing is in the Links thread, up top.
You might deduce "passes" by subtracting from "touches" several events (like FGA, TO, Ast, etc.) But I think Bob's 'touches' are merely a multiple of assists, added to those other events. In other words, a guess based on league averages.
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Thu Nov 17, 2005 7:14 am Post subject: Reply with quote
Bob's thing is in the 'Sticky: Linkage' thread, up top.
You might deduce 'passes' by subtracting from 'touches' several events (like FGA, TO, Ast, etc.) But I think Bob's 'touches' are merely a multiple of assists, added to those other events. In other words, a guess based on league averages.
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Sat Aug 06, 2005 11:31 am Post subject: passing statistics Reply with quote
I'm a fan who values passing as one of the most important, and often overlooked, skills in basketball. I was wondering if there are any good passing statistics out there. I think something as simple as adjusted assists for a player's usage rating could give a decent idea of how good players are at passing the ball which leads to a score. If there aren't any good statistics out there I might try to come up with a formula myself. Thanks.
Back to top
View user's profile Send private message
jeffpotts77
Joined: 18 Feb 2005
Posts: 150
Location: Cambridge, MA
PostPosted: Sat Aug 06, 2005 11:44 am Post subject: link to knickerblogger.net Reply with quote
You should check out Knickerblogger's Stats page. Below I've pasted the link to the page with all players sorted by Assist Ratio - a formula developed by John Hollinger, used to evaluate passing skills (40 best; 2 worst).
http://www.knickerblogger.net/stats/jh_ALL_AST.htm
Perhaps someone else will be able to elaborate on what the formula consists of.
Back to top
View user's profile Send private message Yahoo Messenger
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Sat Sep 03, 2005 5:30 pm Post subject: Reply with quote
Favors guards too much. Anyone else?
Back to top
View user's profile Send private message
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Sat Sep 03, 2005 6:21 pm Post subject: Re: passing statistics Reply with quote
mateo82 wrote:
I'm a fan who values passing as one of the most important, and often overlooked, skills in basketball. I was wondering if there are any good passing statistics out there. I think something as simple as adjusted assists for a player's usage rating could give a decent idea of how good players are at passing the ball which leads to a score. If there aren't any good statistics out there I might try to come up with a formula myself. Thanks.
Pointing out the obvious, but Bob Chaikin's database has how often players pass the ball. It doesn't quantify how good the passes are, but he has done work to show that it's pretty accurate.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sat Sep 03, 2005 8:43 pm Post subject: Reply with quote
So I took the numbers I have from Bob's database from last year, and looked at percent of passes that ended up in assists.
The results are surprising, although perhaps not so telling.
The range of the entire dataset (limited to players playing more than 1000 minutes last year) was 15.16%-17.95%, with 177 of the 244 players being over 17%. In fact, take away Danny Fortson, and the range becomes 16.37%-17.95%.
Big men head the list, and someone like Nash ends up at #105 in the league, at 17.15%. So I'm not sure this is the greatest of metrics.
Top 10:
Code:
Eddy Curry - 0.179514051
Emeka Okafor - 0.178372352
Antonio Davis - 0.176453202
Marc Jackson - 0.176385124
Rasheed Wallace - 0.176333138
Erick Dampier - 0.176092811
Troy Murphy - 0.176043557
Donyell Marshall - 0.175356558
Tayshaun Prince - 0.175127636
Stromile Swift - 0.175104248
Back to top
View user's profile Send private message
bchaikin
Joined: 27 Jan 2005
Posts: 686
Location: cleveland, ohio
PostPosted: Sat Sep 03, 2005 10:45 pm Post subject: Reply with quote
So I took the numbers I have from Bob's database from last year, and looked at percent of passes that ended up in assists. The results are surprising, although perhaps not so telling.
not surprising at all - again in the stats database (and the simulation) a player's number of passes are calculated based on his assists, not vice versa. there are no players with higher percentages of assists per pass. any discrepencies you see between players are simply a factor of his team's game pace or some other small factor - there is very little difference between players ratios of assists to passes, as you can see from your calculations...
hopefully 82games.com will at some point show which players (if there are any) consistently show a higher (or lower) percentage of assists per pass, especially if many games are charted this upcoming season...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sat Sep 03, 2005 11:48 pm Post subject: Reply with quote
That would explain that.
Here's a hopefully more useful metric, combining data from 82games.com and Bob's database.
Top 15 Passing TOs/Pass:
Code:
Antonio Daniels 2.106251521
Matt Bonner 2.168063984
Jason Hart 2.592307439
Fred Hoiberg 2.607174945
PJ Brown 2.608071983
Aaron McKie 2.817962288
Elton Brand 2.848053498
Earl Boykins 2.854015803
Greg Buckner 2.942881234
Eric Snow 2.972836062
Kevin Garnett 2.975945433
Brendan Haywood 2.982163679
Tyronn Lue 2.995493736
Jeff McInnis 3.16151304
Brad Miller 3.162522656
Again, limited to players playing over 1000 minutes, and doesn't show that many surprises. Antonio Daniels is up there, I suppose PJ Brown, Elton Brand and Brendan Haywood could be called surprises, especially since they're big men.
If you wanted to do something to include the fact that someone might be taking a more risky pass if they're setting up a teammate for an assist, and so weight it for assists, we could do something like (1-passTO%)*(AST/MIN). Here's how that looks:
Code:
Steve Nash 32.2631855
Brevin Knight 29.42382862
Jason Kidd 21.56630353
Maurice Williams 20.56280182
Rick Brunson 20.31460208
Stephon Marbury 19.68436496
Jason Williams 19.65336655
Andre Miller 19.18777592
Sam Cassell 19.09896538
Earl Watson 19.04008207
Jason Hart 18.94468636
Jamaal Tinsley 18.68503728
Rafer Alston 18.20340527
Luke Ridnour 18.13249877
Gary Payton 17.86823422
No surprises here, I don't think. Almost all guards (that's who gets the assists), with the best big man checking in at #40 (KG). This really doesn't differentiate too much between itself and straight assists per minute (Steve Francis moves up a couple spots, but that's really the only dramatic movement).
What else could we include? How would you go about constructing a "passer rating"?
Back to top
View user's profile Send private message
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
PostPosted: Sun Sep 04, 2005 11:14 am Post subject: Reply with quote
This is an area that will be greatly helped by the game charting this season, but we also do post a 'passing rating' for players which is really just a placeholder until we get something better!
Pass Rtg = (3pt AST + 2pt Jump AST + (2*Close AST)+ (4*Dunk AST) - (2* Pass T/O) ) / (seconds played / 2880)
I think the formula was concocted in about 10 seconds..the object obviously being to reward the passer more heavily for close/dunk asssists under the theory that the pass was a bigger contributor to the made bucket.
Under this scheme last year's leaders
24.6 - Nash
17.9 - Knight
16.2 - A.Miller
14.1 - B.Davis (GS, he was 13.8 in NO minutes)
13.7 - Jason Williams (MEM)
13.3 - Maurice Williams (MIL)
12.9 - Arroyo (UTA, 11.9 with DET)
12.5 - Kidd
12.2 - Livingston (LAC)
12.2 - Watson
12.0 - Francis
11.7 - Brunson
11.3 - Stoudamire (POR)
11.3 - Duhon
11.3 - Snow
11.2 - Iverson
11.2 - Boykins
10.8 - Marbury
10.8 - Payton
10.7 - Claxton (NO, 15.3 in lim min with GS)
10.6 - Terry
10.6 - Jaric
10.5 - Tinsley
10.5 - Carter (MIN)
10.5 - Lue
10.4 - Hinrich
10.3 - Ridnour
10.3 - Cassell
10.2 - Wade
10.2 - Telfair
10.2 - Hart
10.0 - LeBron James
...and the worst rated
-1.0 - Collier
-1.0 - Dalembert
-0.6 - Henderson (DAL)
-0.3 - Gadzuric
-0.1 - Fortson
0.0 - Bradley (DAL)
0.1 - Mutombo
0.1 - Battie
0.1 - Curry
0.1 - Hunter
0.1 - Wilcox
0.2 - Mohammad (NY)
0.3 - Pollard
0.3 - Sweetney
0.3 - Jerome James
0.3 - Okafor
it would probably be worth ranking by position...
"PG"
24.6 - Nash
17.9 - Knight
16.2 - A.Miller
14.1 - B.Davis
13.7 - Williams (MEM)
...
4.8 - Vaughn
4.1 - Best
"PG/SG"
12.0 Francis
11.7 Brunson
11.3 Snow
11.2 Iverson
10.6 Terry
...
1.4 Evans (SAC)
1.3 Snyder (UTA)
"SG/SF"
10.0 James (CLE)
7.7 Hamilton
7.0 Christie (SAC)
6.5 Cater (TOR)
6.1 Bryant
...
1.1 J.Jones (IND)
0.7 Lewis (SEA)
"SF/PF"
9.2 Kukoc
7.4 Garnett
6.8 Webber (SAC)
5.5 Kirilenko
5.3 Walker (BOS)
...
0.3 Sweetney
0.1 Wilcox
"PF/C"
6.0 Brad Miller
4.8 Shaq
4.4 Collins (UTA)
4.4 Camby
4.0 Duncan
...
-1.0 Dalembert
-1.0 Collier
Back to top
View user's profile Send private message Visit poster's website
KD
Joined: 30 Jan 2005
Posts: 163
PostPosted: Sun Sep 04, 2005 4:15 pm Post subject: Off-topic, my apologies ... Reply with quote
From a Ric Bucher piece on Rick Brunson ...
Quote:
"His assists per minutes played is one of the best in the league," Sund said. "What we like is that there is no unknown with him. You know what you're getting. He knows who he is."
http://sports.espn.go.com/nba/columns/story?id=2149948
Not entirely true, but Brunson's assist ratio was third in the league last year.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Sun Sep 04, 2005 9:21 pm Post subject: Reply with quote
Roland_Beech wrote:
the object obviously being to reward the passer more heavily for close/dunk asssists under the theory that the pass was a bigger contributor to the made bucket.
I'm slightly conflicted about this. On the one hand, I agree that an assist for a dunk is usually set up almost exclusively by the passer, and the finisher has to have very little ability to be able to finish. On the other hand, since an assist means that you have a guaranteed made shot, then shooting percentages are improved more by assists on jump shots.
That is, since the farther away from the basket you are, the harder the shot, then a great set up on the outside may be worth more (especially if it's for 3) than a great set up inside. Your chances have improved much more, relatively.
Back to top
View user's profile Send private message
kjb
Joined: 03 Jan 2005
Posts: 864
Location: Washington, DC
PostPosted: Mon Sep 05, 2005 10:21 am Post subject: Reply with quote
I also question giving a bonus for "close" assists. Why should a pass to cutter for an open dunk have more value in a rating system than say a big man passing out of a double team for an open 3? In the game, the open 3 is worth more, and the play (pass out and making the open 3) probably does more to change the way a defense plays. In this rating system, a pass-ahead to a wide-open teammate on the break, who makes an uncontested dunk has 4 times more value than that pass out of a double team. That doesn't seem right to me.
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Roland_Beech
Joined: 14 Jan 2005
Posts: 43
PostPosted: Mon Sep 05, 2005 11:12 am Post subject: Reply with quote
Well as I mentioned I think the formula was about 10 seconds worth of thought so it's not without drawbacks...
However, I do think inside assists are worth more than outside shot assists (since we can't tell how open a guy is at this stage). Part of it is that often the degree of difficulty on the pass is much harder (threading through defenders, etc) and part of it requires more understanding I believe of player movement (eg anticipating both your own teammate and the defenders).
I agree that a big man passing to a wide open perimeter player out of a double team can be valuable, but even more valuable is often the pass to a cutter down the lane in such a situation.
Anyway I would completely agree that passing stats are currently inadequate...that's one reason why this coming season should be exciting.
Back to top
View user's profile Send private message Visit poster's website
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
PostPosted: Mon Sep 05, 2005 11:33 am Post subject: Reply with quote
Roland_Beech wrote:
Well as I mentioned I think the formula was about 10 seconds worth of thought so it's not without drawbacks...
However, I do think inside assists are worth more than outside shot assists (since we can't tell how open a guy is at this stage). Part of it is that often the degree of difficulty on the pass is much harder (threading through defenders, etc) and part of it requires more understanding I believe of player movement (eg anticipating both your own teammate and the defenders).
I agree that a big man passing to a wide open perimeter player out of a double team can be valuable, but even more valuable is often the pass to a cutter down the lane in such a situation.
Anyway I would completely agree that passing stats are currently inadequate...that's one reason why this coming season should be exciting.
Frankly, I thought Roland's passing stats were the most representative of passing ability that I'd seen.
And, though the weights he used were derived out of the ether, they do represent a concept that does generally make sense. Defenses take away the inside more than jump shots. There should be modification so that 3pt jump shots get more weight than 2 pt jump shots, but that's relatively minor.
In general, doing this much with assists is pretty impressive. Assists are a measure of teamwork and teamwork is what I still think is the toughest thing we have to deal with. If a big man passes to a 3pt shooter who is then helped on by another guard, but swings it to another guy for an open 3pt shot -- both the big man and that perimeter guy did important things. (I'm not advocating 2 assists because I think that normally 1 is enough.) We already know about passes to guys who get fouled for 2 FT, which could get an assist and modify our interpretation of players a fair amount. We know that Euros don't give as many assists and that rules for giving assists have changed in the NBA. Assists are hard. We can do a lot better through a project like that at 82games, but that may not help interpret the past until we've studied things for many years. And, a lot of times, you can't forecast the future without understanding the past. So Roland's megaproject is going to be very very useful in going forward. (I think Dan's work has the capability of explaining assists, too. But I don't know what weight he ends up with -- my own personal lack of knowledge -- and, further, I don't think it readily allows us to infer their value in other leagues, at different distances, or if they were to start giving assists on free throws, I don't think we can easily know how that would change the weight.)
Always lots of questions on teamwork stuff...
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Mon Sep 05, 2005 2:34 pm Post subject: Reply with quote
Here's my attempt at a similar rating to Roland's, except based on the premise that it's better to assist on a jump shot than a close shot, for the reasons I mentioned above:
"Since the farther away from the basket you are, the harder the shot, then a great set up on the outside may be worth more than a great set up inside. Your chances have improved much more, relatively."
I took data from Roland's "Is Outside Shooting a Lost Art?" article, and decided to look at what players shot from each range on unassisted shots.
So for example, we see that people shot 90.8% on dunks over the period of data Roland analyzed, and 75% of those were assisted. That means that of the 25% that were not assisted, approximately 71.2% went in. Therefore the difference between an assisted dunk's chances of going in and an un-assisted dunk's chances are 28.8%, since an assist means a guaranteed bucket. (I'm sure this is flawed in some way, but I think that we can still measure relative values of each assist this way).
I did that with Jump and Close shots, also. So my formula becomes (((1-j%)*(jump+(1.5*3PT)))+((1-c%)*close)+((1-d%)*dunk))*(1-passTO%), and then divide that all by minutes played, and multiply by 400, just to make it look nicer (also based on a per 40 min basis, although I'm not sure it's measuring anything that would need to be quantified per 40 min).
Edit: Added a multiplier to three pointers to count them more.
j% = unassisted jump shot chance
c% = unassisted close shot chance
d% = unassisted dunk shot chance
passTO% = percentage of passes that ended in TOs
I then compared Roland's results to mine. Here's how it looks.
Roland's Top 10:
Code:
Roland's Rank - My Rank - Player
1 - 1 - Steve Nash
2 - 2 - Brevin Knight
3 - 32 - Andre Miller
4 - 13 - Jason Williams
5 - 12 - Maurice Williams
6 - 7 - Tyronn Lue
7 - 3 - Jason Kidd
8 - 15 - Earl Watson
9 - 34 - Steve Francis
10 - 10 - Rick Brunson
My Top 10:
Code:
My Rank - Roland's Rank - Player
1 - 1 - Steve Nash
2 - 2 - Brevin Knight
3 - 7 - Jason Kidd
4 - 50 - Milt Palacio
5 - 22 - Luke Ridnour
6 - 16 - Stephon Marbury
7 - 6 - Tyronn Lue
8 - 26 - Jason Hart
9 - 36 - Rafer Alston
10 - 10 - Rick Brunson
As you can see, there are some fairly significant differences. Nothing extreme, but enough to mean they are measuring very different things.
Milt Palacio is the biggest outlier on these lists, and that's because he had 234 "outside" assists, as opposed to 45 "inside" ones.
And here are the biggest differences:
Code:
My Rank - Roland's Rank - Player
126 - 205 - Mickael Pietrus
147 - 71 - Carmelo Anthony
166 - 91 - DeShawn Stevenson
99 - 173 - Ben Gordon
81 - 140 - Richard Jefferson
136 - 189 - Tim Thomas
213 - 161 - Udonis Haslem
87 - 135 - Eddie Gill
152 - 104 - Nene
4 - 50 - Milt Palacio
Comments?
Back to top
View user's profile Send private message
THWilson
Joined: 19 Jul 2005
Posts: 164
Location: phoenix
PostPosted: Mon Sep 05, 2005 4:53 pm Post subject: Reply with quote
FFSBasketball:
This doesn't make sense to me. I think the problem is that you are ignoring all the potential assists that didn't result in made baskets. For example here are some dummy values:
Assisted Dunk FG % = 80%
Unassisted Dunk FG % = 80%
Assisted Jump FG % = 40%
Unassited Jump FG % = 40%
In this scenario, a jump shot assist is worth three-times a dunk assist. But, potential assists don't help shooting percentage at all for jump shots or dunks. That doesn't make sense. Your model also seems to assume that it's equally easy to get an unassisted or assisted dunk. That rings false.
Back to top
View user's profile Send private message
Ben F.
Joined: 07 Mar 2005
Posts: 391
PostPosted: Mon Sep 05, 2005 5:16 pm Post subject: Reply with quote
All good points. I of course couldn't include potential assists since we don't have data on that.
Note my comment that "I'm sure this is flawed in some way". How would you go about fixing it using similar data and a similar premise?
Back to top
View user's profile Send private message
Carlos
Joined: 21 Jan 2005
Posts: 64
Location: Montevideo, Uruguay
PostPosted: Mon Sep 05, 2005 9:12 pm Post subject: Reply with quote
FFSBasketball, I think the flaw may be that you are assuming that the "alternative shot" for an assisted dunk would be an unassisted dunk. In fact, dunks are relatively rare shots and the most probable alternative shot for an assisted dunk would be a jump shot. The rating could look better if you assumed that the FG% of the alternative shot to every assisted shot (dunk, close or jump shot) is the FG% of ALL unassisted shots.
Back to top
View user's profile Send private message
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Tue Sep 06, 2005 4:28 am Post subject: Reply with quote
FFSBasketball wrote:
So for example, we see that people shot 90.8% on dunks over the period of data Roland analyzed, and 75% of those were assisted. That means that of the 25% that were not assisted, approximately 71.2% went in. Therefore the difference between an assisted dunk's chances of going in and an un-assisted dunk's chances are 28.8%, since an assist means a guaranteed bucket. (I'm sure this is flawed in some way, but I think that we can still measure relative values of each assist this way).
Echoing THWilson's comment, there's a big thing missing here: yes, all assists imply that a bucket was scored. But you're ignoring the passes that were made which led to shots which were MISSED. Which meant no assist. But if the shot had been made, there would've been an assist.
In other words, you're giving a free lunch to the passer who throws the ball to, say, Iverson who's 25 feet from the basket. Iverson lets fly, if the shot goes in, the passer gets an assist and you're assuming that his pass led to a "guaranteed bucket".
But that bucket was not guaranteed. In fact, Iverson misses more often than he connects, and you're not counting the passes thrown to Iverson which led to missed shots rather than assists.
To apply this to your dunk example: you're assigning "blame" for all the missed dunks to the unassisted shooters, and assuming that the recipients of assists had 100% success and no misses. But some of those missing dunkers (an unknown percent) were the recipients of passes, passes which would've led to an assist if the dunker hadn't missed.
With dunks this may not be a big deal because most dunks are made anyway. But outside jumpshots and especially 3-pointers have a high rate of misses.
Iverson averages 31% on his 3-pointers. Some of them are assisted, some are unassisted, but we can't assume that the assisted ones were "guaranteed buckets". The key questions are these two: what was his 3pt FG% on unassisted attempts? And what was his 3pt FG% on attempts which *could have* led to assists? The difference between those two is the measure which tells us how much value added the assister gave.
Unfortunately, due to inadequate stats-keeping by the NBA, we don't know Iverson's unassisted and assisted FG%s are. But I can tell you for sure, they're not 100% and it's unsound to assume that they are.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Wed Sep 07, 2005 9:42 am Post subject: Reply with quote
I can't tell if it's been said in roundabout ways already, but the FG% of a "potential assist" is built in to the assist total.
A pass to a guy within dunking distance is going to be an assist about twice as likely as a pass to an open 3-pt shooter. So there's no real point (that I can see) in doubling credit for the passer in the former instance.
Assigning weights for the closeness of the shots is probably just distinguishing teams that put more dunkers or shooters on the floor.
Back to top
View user's profile Send private message Send e-mail
HoopStudies
Joined: 30 Dec 2004
Posts: 706
Location: Near Philadelphia, PA
PostPosted: Wed Sep 07, 2005 11:23 am Post subject: Reply with quote
Mike G wrote:
I can't tell if it's been said in roundabout ways already, but the FG% of a "potential assist" is built in to the assist total.
A pass to a guy within dunking distance is going to be an assist about twice as likely as a pass to an open 3-pt shooter. So there's no real point (that I can see) in doubling credit for the passer in the former instance.
Assigning weights for the closeness of the shots is probably just distinguishing teams that put more dunkers or shooters on the floor.
Not true, actually.
Yes, a higher percentage of dunks will be assisted on that mid-range js. But it is a lot harder to get an assist to a player dunking the ball because the D takes that away first.
Passers take bigger risks getting the ball to a guy near the basket than to a jump shooter.
We give bigger weight to people who do well on their SAT than on their 4th grade addition test. The SAT is tougher. If someone aces 200 4th grade addition tests, but gets a 700 on their SAT, how do we feel he ranks relative to a guy who did ok on his 4th grade addition tests, but then got a 1500 on their SAT?
We say "nice pass" when it's to a player in great scoring position through traffic, not when it's a floater to a wide open big man at the 3pt line. The pass is increasing the likelihood of scoring in the first case, but not in the 2nd. Because of the bias in not counting potential assists, you may say it balances out. But it doesn't. It may take one player 300 passes to get 100 perimeter assists. It may take another 150 passes to get 100 inside assists. We don't know from the data. But we know it should take fewer to get the inside assists. Giving conceptual credit as Roland has (or as I do in my work) gets us closer to measuring an efficiency of passing.
It's a philosophical question, which means that this thread could go on waaaayyy too long for me. But I am comfortable that more weight should go on the pass for close shots.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Wed Sep 07, 2005 12:40 pm Post subject: Reply with quote
HoopStudies wrote:
...
We say "nice pass" when it's to a player in great scoring position through traffic, not when it's a floater to a wide open big man at the 3pt line. .... It may take one player 300 passes to get 100 perimeter assists. It may take another 150 passes to get 100 inside assists. ...
Well, I think that was exactly what I was saying. Or to use different numbers, 300 passes to the outside may lead to 100 assists, while 300 passes to the inside may get you 200 assists.
These number make sense, intuitively and statistically. Neglecting for the moment that more of the 'outside' passes will lead to 3 points , and that more of the inside passes will lead to turnovers, the inside-passing player gets twice as many assists for the same number of passes; and that seems to be just about what his credit should be, relatively speaking.
Inside players who are a scoring threat are the ones who open up the outside players; if they also find them with timely passes, they get assists on something less than half their passes.
It seems a point guard who runs with a bunch of dunkers is going to get a good % of his assists with dunks, but I don't see why he should be credited with more than the actual number of baskets created. If Steve Nash gets 12 assists by setting up lots of open jumpshots, that's just as good (I think) as 12 assisted dunks.
I'm pretty sure the value of a basketball play is in how many points it produces, and not in how difficult it is.
Back to top
View user's profile Send private message Send e-mail
HoopStudies
Joined: 30 Dec 2004
Posts: 706
Location: Near Philadelphia, PA
PostPosted: Wed Sep 07, 2005 12:58 pm Post subject: Reply with quote
Mike G wrote:
HoopStudies wrote:
...
We say "nice pass" when it's to a player in great scoring position through traffic, not when it's a floater to a wide open big man at the 3pt line. .... It may take one player 300 passes to get 100 perimeter assists. It may take another 150 passes to get 100 inside assists. ...
Well, I think that was exactly what I was saying. Or to use different numbers, 300 passes to the outside may lead to 100 assists, while 300 passes to the inside may get you 200 assists.
These number make sense, intuitively and statistically. Neglecting for the moment that more of the 'outside' passes will lead to 3 points , and that more of the inside passes will lead to turnovers, the inside-passing player gets twice as many assists for the same number of passes; and that seems to be just about what his credit should be, relatively speaking.
Inside players who are a scoring threat are the ones who open up the outside players; if they also find them with timely passes, they get assists on something less than half their passes.
It seems a point guard who runs with a bunch of dunkers is going to get a good % of his assists with dunks, but I don't see why he should be credited with more than the actual number of baskets created. If Steve Nash gets 12 assists by setting up lots of open jumpshots, that's just as good (I think) as 12 assisted dunks.
I'm pretty sure the value of a basketball play is in how many points it produces, and not in how difficult it is.
Not how difficult the whole play is, but how difficult the pass is relative to the shot. It's a lot easier to make a dunk than a jumper. You should give more credit to the passer on a dunk because the pass is harder than on a jumper where the pass (and creation) is easier.
Conceptually you can watch any play and identify what are key points in making the play work. That's what analysts try to do. Sometime it is just making the shot, creating your own shot. An assist is an attempt to show that there may be other keys. That's why we talk about measuring picks, too, because they may be key. But some picks and some assists are worth relatively more in generating that score. As assist rules have loosened up, what I'm saying is that the credit per assist has gone down because their job to get an assist has gotten easier. It's a theoretical framework that works in interpreting assists from other leagues and from the past. Getting the data to parameterize it is more difficult.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers.
Back to top
View user's profile Send private message Visit poster's website
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Wed Sep 28, 2005 4:55 pm Post subject: Reply with quote
FFSBasketball wrote:
No surprises here, I don't think. Almost all guards (that's who gets the assists), with the best big man checking in at #40 (KG). This really doesn't differentiate too much between itself and straight assists per minute (Steve Francis moves up a couple spots, but that's really the only dramatic movement).
What else could we include? How would you go about constructing a "passer rating"?
Yeah, that's the problem. Taking assists into account assumes that #1 your teammate can catch the ball and #2 he can score the ball. A big man could be an excellent passer, but if his guards can't make jumpshots he doesn't get the assists. I'm looking for something that doesn't take teammates into account at all, if possible. The search continues.
Back to top
View user's profile Send private message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Wed Sep 28, 2005 5:03 pm Post subject: Reply with quote
HoopStudies wrote:
In general, doing this much with assists is pretty impressive. Assists are a measure of teamwork and teamwork is what I still think is the toughest thing we have to deal with. If a big man passes to a 3pt shooter who is then helped on by another guard, but swings it to another guy for an open 3pt shot -- both the big man and that perimeter guy did important things.
This is exactly my point. Plenty of times someone created a play with a good pass to a player who then passes it to someone else who is wide open. Only the second player gets credit with the assist, although the 1st player might have been the one who made the play happen. In the triangle offense this happens often.
This is why I don't like the assist approach. A pass that immediately leads to a score is only 1 of many types of good passes.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Thu Sep 29, 2005 8:30 am Post subject: Reply with quote
Here again, passer #1 may or may not get the assist; passer #2 has the option of shooting or passing (potential for points or assist, but not both). It doesn't seem reasonable that over the course of a season passer#1 would predominantly pass to guys who don't shoot. In the long run, everyone gets the # of assists they should get, based on the # of good passes they make.
That's relative to the team, of course. If you are on a poor-shooting team, or with weak finishers around the basket, you won't get as many assists (Mateo's point). Magic had great shooters and finishers. Stockton was just a great passer. Or are those just subjective notions?
Subtracting Stockton's FG/FGA from his team's FG/FGA would give you 'everyone else's' FG%. How much of the Jazz' high FG% was due to the great passing of Stockton? How much of Stock's assist rate was due to his great finishers?
Players who bounce around from team to team may see their assist rates bounce, too. The Brevin Knight instance has been pointed out as a case of "system assists". Whether it's by coaching design or by teammate strengths, statistical reward tends to be team-specific. Playoff appearances, championships, etc, will always be beyond the control of the individual. Some stats will be, as well.
Back to top
View user's profile Send private message Send e-mail
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Thu Sep 29, 2005 3:24 pm Post subject: Reply with quote
Mike G wrote:
Here again, passer #1 may or may not get the assist; passer #2 has the option of shooting or passing (potential for points or assist, but not both). It doesn't seem reasonable that over the course of a season passer#1 would predominantly pass to guys who don't shoot. In the long run, everyone gets the # of assists they should get, based on the # of good passes they make.
Not necessarily. This seems like a guard-centric statement Wink . The player who "makes the play" so to speak, or the player who draws the defense is going to get more assists because he has given a teammate an easier shot. That doesn't mean he's a better passer though, just a better playmaker. Forwards and centers are always going to be at a disadvantage when you use assists as your starting point. They're already close to the basket, there isn't a lot they can do, in quick movements, to draw in the opposing defense. Except for cutters, the big men are more likely to pass out to someone who's open. Sometimes that player will take the shot, sometimes that player will make the extra pass to someone who's even more open. I can't think of a whole lot of circumstances where guards make passes that lead to extra passes. It's not a regular feature of the game.
I think this is why the attempts in this thread to construct a list of the best passers has been so guardcentric. Although I think guards are in general the better passers, not in this dominant of a degree. I'd take Kevin Garnett and probably a couple of other big men's passing over the vast majority of guards.
Back to top
View user's profile Send private message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Thu Sep 29, 2005 3:25 pm Post subject: Reply with quote
I think the best way to resolve this is to wait out this season. Then we'll have more data and we can use turnovers caused by bad passes as a better gauge (or at least part of a bigger formula) for passing ability.
Back to top
View user's profile Send private message
mateo82
Joined: 06 Aug 2005
Posts: 211
PostPosted: Wed Nov 16, 2005 7:11 pm Post subject: Re: passing statistics Reply with quote
HoopStudies wrote:
Pointing out the obvious, but Bob Chaikin's database has how often players pass the ball. It doesn't quantify how good the passes are, but he has done work to show that it's pretty accurate.
Do you know where I can find (or learn about) this database? I would love to see this. We can combine this with the usage rating and get a "ball hog" stat. Smile
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Thu Nov 17, 2005 7:12 am Post subject: Reply with quote
Bob's thing is in the Links thread, up top.
You might deduce "passes" by subtracting from "touches" several events (like FGA, TO, Ast, etc.) But I think Bob's 'touches' are merely a multiple of assists, added to those other events. In other words, a guess based on league averages.
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3626
Location: Hendersonville, NC
PostPosted: Thu Nov 17, 2005 7:14 am Post subject: Reply with quote
Bob's thing is in the 'Sticky: Linkage' thread, up top.
You might deduce 'passes' by subtracting from 'touches' several events (like FGA, TO, Ast, etc.) But I think Bob's 'touches' are merely a multiple of assists, added to those other events. In other words, a guess based on league averages.
Re: More recovered assist threads
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Wed Jan 17, 2007 5:03 pm Post subject: Michael Redd - all-time leader in turnover rate Reply with quote
I was looking at basketball-reference and noticed that Michael Redd is the all-time leader in turnover rate (meaning he has the lowest rate of turnovers, not the highest). For the most part the leaders on this list are the shooting specialists who camp on the perimeter. The all-time NBA top 50 list includes guys like John Paxson, Steve Kerr, Wesley Person and Vinny Del Negro. But the top 50 is also full of all-star players: Jordan, McGrady, Marion, Jamison, Glenn Rice and Redd.
Turnover rate is defined as "estimated turnovers per 100 possessions." It takes FGA + 0.44*FTA + AST + TO as the measure of possessions: the player either shoots, gets to the line, makes an assisting pass or turns it over.
What exactly does turnover rate mean? Is it one of those stats like Ortg where the Fred Hoiberg-esque specialists who hit a lot of threes appear at the top? As far as I can tell, hitting threes wouldn't help your turnover rate (it would just be a FGA), but camping on the perimeter and never dribbling would certainly help your rate just by minimizing the chance of any turnovers.
I remember when Redd signed the max contract a year ago and there was a lot of skepticism because of a PER around 18 (good but not great) and a perceived weakness on defense. I was among the skeptics: he seemed like a chucker on a bad team who didn't come close to meriting 10+ million a year. He looked a lot like an Allan Houston and we all know how _that_ big contract worked out.
Does this mean Redd is a superlative player because he manages to have the ball in his hands all the time yet almost never turn it over? Or is there something missing in this calculation?
Here's what I can't figure out: if a player passes the ball to another player and the second player makes the shot, that helps the first player's turnover rate (lowers it). If the second player misses the shot, it has not effect on the turnover rate (since it's not a FGA, FTA, AST or TO). So it makes sense that players who both scored a lot and got a lot of assists would have a low turnover rate. However, Redd has never really been known as a big assists guy.
Is he really a great guard who hangs onto the ball unlike any other, or will "extreme charting" (possession-by-possession a la 82games) reveal that Redd is just average in this category? What's missing from the turnover rate calculation that could make it better?
Back to top
View user's profile Send private message Visit poster's website
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Wed Jan 17, 2007 6:03 pm Post subject: Re: Michael Redd - all-time leader in turnover rate Reply with quote
The two main ways to get turnovers are to dribble the ball or pass the ball. So players that seldom do this will avoid turnovers. That's why catch and shoot guys have good turnover rates. Of course, just making safe dribbles/passes is almost as good. And then you have the guys that are so talented that they can dribble and pass in traffic and still not commit many turnovers.
Unfortunately, I haven't watched Redd enough to know which one he is. His Assist Rate is really low though, so it looks like he doesn't pass much -- or only makes safe passes that don't immediately lead to shots.
AJax wrote:
What's missing from the turnover rate calculation that could make it better?
Off the top of my head it would be nice to track touches and passes that don't lead to any other boxscore statistic. And it would be nice to track how many times a player initiates a dribble. I'm not sure these would make turnover rate better, but it could give us a fuller picture.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Thu Jan 18, 2007 6:50 am Post subject: Reply with quote
'Turnover rate' is another of those stats that's hard to figure a use for. 'Used possessions' are in the denominator, along with assists. 'Gained possessions' (DefReb) are not.
Using a formula for 'expected TO' based on other stats, I've generated a ratio of expected/actual for current players with substantial careers. Redd is quite close to the best:
Code:
ac/ex player TO exp
.592 LaFrentz 1.2 2.1
.595 BWallace 1.2 2.0
.628 Marion 1.6 2.5
.637 Nowitzki 1.9 3.0
.668 Redd 1.6 2.4
.701 Dale Davis 1.3 1.9
.718 Stojakovic 1.7 2.3
.736 Camby 1.7 2.3
.752 PJBrown 1.4 1.8
.754 EJones 1.7 2.3
.755 McGrady 2.4 3.2
.755 RWallace 1.8 2.4
Redd does in fact look like the best guard. Further down the list is a better mix of bigs, midsizers, and even PG -- Billups and Payton being foremost.
Such a list isn't dominated by catch-and-shoot players, or any other 'type'. Rebounders tend to get more TO than non-rebounders because they can be stripped or otherwise lose a ball they haven't moved to a safe place.
Here's the bottom of the list. The rates are per36:
Code:
1.153 AWalker 3.0 2.6
1.153 Dampier 2.2 1.9
1.158 JWilliams 2.8 2.4
1.174 Wells 2.8 2.4
1.206 RDavis 2.8 2.3
1.222 Harrington 2.5 2.1
1.225 Francis 3.6 2.9
1.307 Stackhouse 3.3 2.5
Stack has tended to get 31% more turnovers than he should, considering what else he does. These are career rates, and he used to be even worse.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Neil Paine
Joined: 13 Oct 2005
Posts: 774
Location: Atlanta, GA
PostPosted: Thu Jan 18, 2007 8:48 am Post subject: Reply with quote
That's really cool, Mike. If you can say, what's the formula for "expected TO"? Did you run a regression on the other boxscore stats?
Back to top
View user's profile Send private message Visit poster's website
cherokee_ACB
Joined: 22 Mar 2006
Posts: 157
PostPosted: Thu Jan 18, 2007 9:01 am Post subject: Reply with quote
Mike G wrote:
Such a list isn't dominated by catch-and-shoot players, or any other 'type'. Rebounders tend to get more TO than non-rebounders because they can be stripped or otherwise lose a ball they haven't moved to a safe place.
Play by play analysis reveals that, out of 50481 rebounds so far this year, a turnover of the rebounder followed, within 5 seconds, in 1021 cases, or roughly 2% of rebounds. Not really significant. I'm pretty sure the putback to turnover ratio is higher than the assist to bad pass ratio. Big men have more problems with offensive fouls than with post-rebound turnovers.
Actually, because potential assists are not tracked, it's passers who are misrepresented the most with the turnover ratio.
Back to top
View user's profile Send private message
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Thu Jan 18, 2007 4:42 pm Post subject: Reply with quote
Speaking of offensive fouls, I wonder if that isn't the main source of excellence for some of these players. Redd is much more of a pull-up shooter than Iverson or Arenas, so I would bet he commits many fewer charges than the average 25+ ppg scorer. Even though he's not a great passer, he avoids the crash-through-the-lane drives that turn the ball over one way or another (stripped, charge, bad pass). What seems unusual about Redd compared to big-time scorers like Iverson, McGrady, Kobe, Carmelo and the others is that he's not a particular wizard off the dribble, and doesn't even have greats hops, so he's really just shooting a _lot_ of jump shots.
The low "actual to expected turnover rate" makes sense. One of the pieces of praise I've heard for Nowitzki is that for a player who dribbles and shoots a lot, he's very good at not turning it over.
A lot of the big men on that "best" list very rarely dribble, and so make their living on putbacks and alley-oops. 82games pointed out that Marion is one of the most-assisted players in basketball (~80% of fgm coming from an assist), and though I don't see him play often, he is rarely beating anyone one-on-one off the dribble. I'm pretty sure the same is true for Ben Wallace, PJ Brown, and Camby. Guess who's #1 on that most-assisted list? Mr. LaFrentz, the #1 guy on Mike G's list.
Interesting note from the linked article is that Redd is nowhere on the list of least-assisted players, which is dominated by the blue-chip guards like Iverson, Nash, Baron Davis, Billups, Wade, Arenas. So maybe his turnover rate is low because he actually gets many more assisted passes than other high-scoring small players.
I'm also curious about how "expected TOs" was calculated - is it an estimate based on how many shots and assists the player gets per game, or some other heuristic?
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Thu Jan 18, 2007 9:48 pm Post subject: Reply with quote
My Expected TO is:
.16*Ast + .10*Blk + .08*Sco + .07*Reb + .05*Stl - .005*Min
There was a thread some time ago about whether blocks should be correlated with TO. I don't know why it should be so large, unless shotblockers just tend to be 'swatters', as opposed to 'grabbers' of the ball.
LaFrentz has been a shotblocker; so if this relation is overstated, that could be part of his low ratio. But the %-unassisted-FG thing is newer than this formula, is highly intriguing, and also is estimable from other stats. So that is an inviting avenue for improvement.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Fri Jan 19, 2007 9:57 am Post subject: Reply with quote
OK, here's the thread from a year ago on the subject:
http://sonicscentral.com/apbrmetrics/vi ... c&start=15
Dunno why that URL is so long.
On this year's stats, and using a few more of them, here's an 'expected TO' formula:
EDITED formula:
ExpTO = .026*2FGA - .007*3FGA + .162*FTA + .090*OReb -.051*DReb + .167*Ast + .036*Stl + .018*Blk + .176*PF + .0167*Pts
EDIT: The edited formula (see later post in this thread) doesn't much change the order of players in these mini-tables.
From this formula, the most turnover-proof players this year (>200 minutes played):
Code:
a/e ex to player pos team
.32 37 12 ross,quinton SF lac
.40 20 8 henderson,alan PF phi
.51 49 25 varejao,anderson PF cle
.51 14 7 johnson,linton SF nor
.51 16 8 jones,james SF pho
.52 44 23 daniels,antonio PG was
.52 25 13 roberts,lawrence PF mem
.52 25 13 aldridge,lamarcu C por
.58 24 14 brezec,primoz C cha
.58 14 8 snyder,kirk SG hou
a/e is 'actual/expected' turnovers. QRoss gets less than a third of what he should.
If I limit this to just the 100 biggest producers this year, it's:
Code:
a/e ex to player pos team
.68 63 43 jamison,antawn PF was
.68 61 42 artest,ron SF sac
.69 90 62 billups,chauncey PG det
.70 88 61 hamilton,richard SG det
.71 97 69 redd,michael SG mil
.71 73 52 atkins,chucky PG mem
.73 74 54 martin,kevin SG sac
.76 63 48 boykins,earl PG den
.78 72 56 terry,jason PG dal
.78 67 52 wilcox,chris PF sea
There's our man Redd. In both lists, a good mix of positions.
Here's the worst of the top 100:
Code:
a/e ex to player pos team
1.51 51 77 odom,lamar PF lal
1.47 56 82 harrington,al PF ind
1.45 68 99 mason,desmond SG nor
1.40 80 112 ellis,monta PG gsw
1.39 89 124 howard,dwight PF orl
1.36 48 65 blount,mark C min
1.33 71 95 johnson,joe SG atl
1.31 64 84 miller,mike SF mem
1.28 91 117 nash,steve PG pho
1.26 64 80 bogut,andrew C mil
Nashy is near the opposite end from Billups. Well, while this measure gives us a closer fit between 'real' and 'expected', it actually misses whether a guy is productive with his shot attempts. Nash and Howard are very efficient, others less so. Just using 'points' (instead of FGA and FTA) would give us a better grasp of 'tradeoff' with TO.
From the 200-minute set:
Code:
a/e ex to player pos team
1.96 7 13 croshere,austin PF dal
1.92 8 15 johnson,dermarr SG den
1.90 15 28 wells,bonzi SG hou
1.67 35 58 walker,antoine PF mia
1.63 13 21 tsakalidis,jake C mem
1.63 32 52 williams,marcus PG njn
1.58 41 65 head,luther PG hou
1.55 14 22 hudson,troy PG min
1.55 10 16 griffin,adrian SG chi
1.51 51 77 odom,lamar PF lal
All these guys have turned it over >50% more than it seems they should.
_________________
`
36% of all statistics are wrong
Last edited by Mike G on Mon Jan 22, 2007 2:25 pm; edited 2 times in total
Back to top
View user's profile Send private message Send e-mail
jkubatko
Joined: 05 Jan 2005
Posts: 702
Location: Columbus, OH
PostPosted: Fri Jan 19, 2007 10:55 am Post subject: Reply with quote
Mike G wrote:
My Expected TO is:
.16*Ast + .10*Blk + .08*Sco + .07*Reb + .05*Stl - .005*Min
There was a thread some time ago about whether blocks should be correlated with TO. I don't know why it should be so large, unless shotblockers just tend to be 'swatters', as opposed to 'grabbers' of the ball.
The coefficient of blocked shots is larger than the coefficient of points scored because blocked shots are much more rare. Imagine two players with identical statlines. Now increase one player's blocked shots by 20%, say from 100 to 120. Using your formula, this leads to an expected increase of (120 - 100)*0.10 = 2 turnovers, which is insignificant given the error in your estimates. Now instead of increasing blocks by 20%, increase scoring by 20%, say from 1000 to 1200. This leads to an expected increase of (1200 - 1000)*0.08 = 16 turnovers, a much larger change.
_________________
Regards,
Justin Kubatko
Basketball-Reference.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
jkubatko
Joined: 05 Jan 2005
Posts: 702
Location: Columbus, OH
PostPosted: Fri Jan 19, 2007 11:08 am Post subject: Reply with quote
I meant to mention above that the size of the estimated coefficient for a particular variable does not tell you much about the importance of that variable. You need to consider the error in the estimated coefficient. For example, my guess is that if we constructed a 95% confidence interval for the true coefficient of blocked shots, that interval would contain zero. In other words, the true coefficient of blocked shots is not statistically significantly different from zero at the 5% level of significance. A 95% confidence interval for the true coefficient of points scored, however, would *not* contain zero.
_________________
Regards,
Justin Kubatko
Basketball-Reference.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Fri Jan 19, 2007 11:26 am Post subject: Reply with quote
I would guess that minutes are also not significant at the 95% level of confidence. Partly because the coefficient is small (-.005), and partly because it seems unlikely that turnovers decrease with more minutes played.
Mike, do you have the standard errors for these estimates from the regression?
It seems like the best predictor of "expected turnovers" would be touches, and maybe an interaction variable of touches and position (G/F/C). Without full data on touches, the linear regression model is probably best.
One additional question I have about turnovers: if a player shoots an airball and the ball falls out of bounds, is that a turnover? Similarly, on shots where the ball bounces over the backboard, does that get recorded as a turnover? In general, is a missed shot with no rebound = 1 turnover?
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Mon Jan 22, 2007 2:39 pm Post subject: Reply with quote
Mike G wrote:
ExpTO = .026*2FGA - .007*3FGA + .162*FTA + .090*OReb -.051*DReb + .167*Ast + .036*Stl + .018*Blk + .176*PF + .0167*Pts
... it actually misses whether a guy is productive with his shot attempts. Nash and Howard are very efficient, others less so. Just using 'points' (instead of FGA and FTA) would give us a better grasp of 'tradeoff' with TO...
.
OK, it occurred to me that I had included 3FGA with (all) FGA; so I've separated 2FGA from 3FGA. Also, as quoted above, I've added Points (while retaining FGA) so as to separate shot frequency and shot efficiency.
Most likely, there's a correlation between a given player's FG% and his TO rate, in that a low-% bomb or fallaway has less chance of a TO than, say, a drive to the hoop in traffic.
With the adjustments, Steve Nash is the one 'top ten' TO guy to fall out. Rather than having 28% more TO than expected, he's merely at 121%.
I don't know about regressions. I just have player stats on a spreadsheet, multiply them by factors at the tops of my columns and add them up (calling this 'expected TO'); another column is the absolute value of the difference (TO-ExpTO).
I just adjust the variables up/down, back and forth, getting closest fit between player total Actual-Expected diff. I could get just about as good a fit if I throw out the 3FGA, Stl, and Blk terms, and further fine-tune the others. As Justin points out, they're small.
_________________
`
36% of all statistics are wrong
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Wed Jan 17, 2007 5:03 pm Post subject: Michael Redd - all-time leader in turnover rate Reply with quote
I was looking at basketball-reference and noticed that Michael Redd is the all-time leader in turnover rate (meaning he has the lowest rate of turnovers, not the highest). For the most part the leaders on this list are the shooting specialists who camp on the perimeter. The all-time NBA top 50 list includes guys like John Paxson, Steve Kerr, Wesley Person and Vinny Del Negro. But the top 50 is also full of all-star players: Jordan, McGrady, Marion, Jamison, Glenn Rice and Redd.
Turnover rate is defined as "estimated turnovers per 100 possessions." It takes FGA + 0.44*FTA + AST + TO as the measure of possessions: the player either shoots, gets to the line, makes an assisting pass or turns it over.
What exactly does turnover rate mean? Is it one of those stats like Ortg where the Fred Hoiberg-esque specialists who hit a lot of threes appear at the top? As far as I can tell, hitting threes wouldn't help your turnover rate (it would just be a FGA), but camping on the perimeter and never dribbling would certainly help your rate just by minimizing the chance of any turnovers.
I remember when Redd signed the max contract a year ago and there was a lot of skepticism because of a PER around 18 (good but not great) and a perceived weakness on defense. I was among the skeptics: he seemed like a chucker on a bad team who didn't come close to meriting 10+ million a year. He looked a lot like an Allan Houston and we all know how _that_ big contract worked out.
Does this mean Redd is a superlative player because he manages to have the ball in his hands all the time yet almost never turn it over? Or is there something missing in this calculation?
Here's what I can't figure out: if a player passes the ball to another player and the second player makes the shot, that helps the first player's turnover rate (lowers it). If the second player misses the shot, it has not effect on the turnover rate (since it's not a FGA, FTA, AST or TO). So it makes sense that players who both scored a lot and got a lot of assists would have a low turnover rate. However, Redd has never really been known as a big assists guy.
Is he really a great guard who hangs onto the ball unlike any other, or will "extreme charting" (possession-by-possession a la 82games) reveal that Redd is just average in this category? What's missing from the turnover rate calculation that could make it better?
Back to top
View user's profile Send private message Visit poster's website
asimpkins
Joined: 30 Apr 2006
Posts: 245
Location: Pleasanton, CA
PostPosted: Wed Jan 17, 2007 6:03 pm Post subject: Re: Michael Redd - all-time leader in turnover rate Reply with quote
The two main ways to get turnovers are to dribble the ball or pass the ball. So players that seldom do this will avoid turnovers. That's why catch and shoot guys have good turnover rates. Of course, just making safe dribbles/passes is almost as good. And then you have the guys that are so talented that they can dribble and pass in traffic and still not commit many turnovers.
Unfortunately, I haven't watched Redd enough to know which one he is. His Assist Rate is really low though, so it looks like he doesn't pass much -- or only makes safe passes that don't immediately lead to shots.
AJax wrote:
What's missing from the turnover rate calculation that could make it better?
Off the top of my head it would be nice to track touches and passes that don't lead to any other boxscore statistic. And it would be nice to track how many times a player initiates a dribble. I'm not sure these would make turnover rate better, but it could give us a fuller picture.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Thu Jan 18, 2007 6:50 am Post subject: Reply with quote
'Turnover rate' is another of those stats that's hard to figure a use for. 'Used possessions' are in the denominator, along with assists. 'Gained possessions' (DefReb) are not.
Using a formula for 'expected TO' based on other stats, I've generated a ratio of expected/actual for current players with substantial careers. Redd is quite close to the best:
Code:
ac/ex player TO exp
.592 LaFrentz 1.2 2.1
.595 BWallace 1.2 2.0
.628 Marion 1.6 2.5
.637 Nowitzki 1.9 3.0
.668 Redd 1.6 2.4
.701 Dale Davis 1.3 1.9
.718 Stojakovic 1.7 2.3
.736 Camby 1.7 2.3
.752 PJBrown 1.4 1.8
.754 EJones 1.7 2.3
.755 McGrady 2.4 3.2
.755 RWallace 1.8 2.4
Redd does in fact look like the best guard. Further down the list is a better mix of bigs, midsizers, and even PG -- Billups and Payton being foremost.
Such a list isn't dominated by catch-and-shoot players, or any other 'type'. Rebounders tend to get more TO than non-rebounders because they can be stripped or otherwise lose a ball they haven't moved to a safe place.
Here's the bottom of the list. The rates are per36:
Code:
1.153 AWalker 3.0 2.6
1.153 Dampier 2.2 1.9
1.158 JWilliams 2.8 2.4
1.174 Wells 2.8 2.4
1.206 RDavis 2.8 2.3
1.222 Harrington 2.5 2.1
1.225 Francis 3.6 2.9
1.307 Stackhouse 3.3 2.5
Stack has tended to get 31% more turnovers than he should, considering what else he does. These are career rates, and he used to be even worse.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Neil Paine
Joined: 13 Oct 2005
Posts: 774
Location: Atlanta, GA
PostPosted: Thu Jan 18, 2007 8:48 am Post subject: Reply with quote
That's really cool, Mike. If you can say, what's the formula for "expected TO"? Did you run a regression on the other boxscore stats?
Back to top
View user's profile Send private message Visit poster's website
cherokee_ACB
Joined: 22 Mar 2006
Posts: 157
PostPosted: Thu Jan 18, 2007 9:01 am Post subject: Reply with quote
Mike G wrote:
Such a list isn't dominated by catch-and-shoot players, or any other 'type'. Rebounders tend to get more TO than non-rebounders because they can be stripped or otherwise lose a ball they haven't moved to a safe place.
Play by play analysis reveals that, out of 50481 rebounds so far this year, a turnover of the rebounder followed, within 5 seconds, in 1021 cases, or roughly 2% of rebounds. Not really significant. I'm pretty sure the putback to turnover ratio is higher than the assist to bad pass ratio. Big men have more problems with offensive fouls than with post-rebound turnovers.
Actually, because potential assists are not tracked, it's passers who are misrepresented the most with the turnover ratio.
Back to top
View user's profile Send private message
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Thu Jan 18, 2007 4:42 pm Post subject: Reply with quote
Speaking of offensive fouls, I wonder if that isn't the main source of excellence for some of these players. Redd is much more of a pull-up shooter than Iverson or Arenas, so I would bet he commits many fewer charges than the average 25+ ppg scorer. Even though he's not a great passer, he avoids the crash-through-the-lane drives that turn the ball over one way or another (stripped, charge, bad pass). What seems unusual about Redd compared to big-time scorers like Iverson, McGrady, Kobe, Carmelo and the others is that he's not a particular wizard off the dribble, and doesn't even have greats hops, so he's really just shooting a _lot_ of jump shots.
The low "actual to expected turnover rate" makes sense. One of the pieces of praise I've heard for Nowitzki is that for a player who dribbles and shoots a lot, he's very good at not turning it over.
A lot of the big men on that "best" list very rarely dribble, and so make their living on putbacks and alley-oops. 82games pointed out that Marion is one of the most-assisted players in basketball (~80% of fgm coming from an assist), and though I don't see him play often, he is rarely beating anyone one-on-one off the dribble. I'm pretty sure the same is true for Ben Wallace, PJ Brown, and Camby. Guess who's #1 on that most-assisted list? Mr. LaFrentz, the #1 guy on Mike G's list.
Interesting note from the linked article is that Redd is nowhere on the list of least-assisted players, which is dominated by the blue-chip guards like Iverson, Nash, Baron Davis, Billups, Wade, Arenas. So maybe his turnover rate is low because he actually gets many more assisted passes than other high-scoring small players.
I'm also curious about how "expected TOs" was calculated - is it an estimate based on how many shots and assists the player gets per game, or some other heuristic?
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Thu Jan 18, 2007 9:48 pm Post subject: Reply with quote
My Expected TO is:
.16*Ast + .10*Blk + .08*Sco + .07*Reb + .05*Stl - .005*Min
There was a thread some time ago about whether blocks should be correlated with TO. I don't know why it should be so large, unless shotblockers just tend to be 'swatters', as opposed to 'grabbers' of the ball.
LaFrentz has been a shotblocker; so if this relation is overstated, that could be part of his low ratio. But the %-unassisted-FG thing is newer than this formula, is highly intriguing, and also is estimable from other stats. So that is an inviting avenue for improvement.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Fri Jan 19, 2007 9:57 am Post subject: Reply with quote
OK, here's the thread from a year ago on the subject:
http://sonicscentral.com/apbrmetrics/vi ... c&start=15
Dunno why that URL is so long.
On this year's stats, and using a few more of them, here's an 'expected TO' formula:
EDITED formula:
ExpTO = .026*2FGA - .007*3FGA + .162*FTA + .090*OReb -.051*DReb + .167*Ast + .036*Stl + .018*Blk + .176*PF + .0167*Pts
EDIT: The edited formula (see later post in this thread) doesn't much change the order of players in these mini-tables.
From this formula, the most turnover-proof players this year (>200 minutes played):
Code:
a/e ex to player pos team
.32 37 12 ross,quinton SF lac
.40 20 8 henderson,alan PF phi
.51 49 25 varejao,anderson PF cle
.51 14 7 johnson,linton SF nor
.51 16 8 jones,james SF pho
.52 44 23 daniels,antonio PG was
.52 25 13 roberts,lawrence PF mem
.52 25 13 aldridge,lamarcu C por
.58 24 14 brezec,primoz C cha
.58 14 8 snyder,kirk SG hou
a/e is 'actual/expected' turnovers. QRoss gets less than a third of what he should.
If I limit this to just the 100 biggest producers this year, it's:
Code:
a/e ex to player pos team
.68 63 43 jamison,antawn PF was
.68 61 42 artest,ron SF sac
.69 90 62 billups,chauncey PG det
.70 88 61 hamilton,richard SG det
.71 97 69 redd,michael SG mil
.71 73 52 atkins,chucky PG mem
.73 74 54 martin,kevin SG sac
.76 63 48 boykins,earl PG den
.78 72 56 terry,jason PG dal
.78 67 52 wilcox,chris PF sea
There's our man Redd. In both lists, a good mix of positions.
Here's the worst of the top 100:
Code:
a/e ex to player pos team
1.51 51 77 odom,lamar PF lal
1.47 56 82 harrington,al PF ind
1.45 68 99 mason,desmond SG nor
1.40 80 112 ellis,monta PG gsw
1.39 89 124 howard,dwight PF orl
1.36 48 65 blount,mark C min
1.33 71 95 johnson,joe SG atl
1.31 64 84 miller,mike SF mem
1.28 91 117 nash,steve PG pho
1.26 64 80 bogut,andrew C mil
Nashy is near the opposite end from Billups. Well, while this measure gives us a closer fit between 'real' and 'expected', it actually misses whether a guy is productive with his shot attempts. Nash and Howard are very efficient, others less so. Just using 'points' (instead of FGA and FTA) would give us a better grasp of 'tradeoff' with TO.
From the 200-minute set:
Code:
a/e ex to player pos team
1.96 7 13 croshere,austin PF dal
1.92 8 15 johnson,dermarr SG den
1.90 15 28 wells,bonzi SG hou
1.67 35 58 walker,antoine PF mia
1.63 13 21 tsakalidis,jake C mem
1.63 32 52 williams,marcus PG njn
1.58 41 65 head,luther PG hou
1.55 14 22 hudson,troy PG min
1.55 10 16 griffin,adrian SG chi
1.51 51 77 odom,lamar PF lal
All these guys have turned it over >50% more than it seems they should.
_________________
`
36% of all statistics are wrong
Last edited by Mike G on Mon Jan 22, 2007 2:25 pm; edited 2 times in total
Back to top
View user's profile Send private message Send e-mail
jkubatko
Joined: 05 Jan 2005
Posts: 702
Location: Columbus, OH
PostPosted: Fri Jan 19, 2007 10:55 am Post subject: Reply with quote
Mike G wrote:
My Expected TO is:
.16*Ast + .10*Blk + .08*Sco + .07*Reb + .05*Stl - .005*Min
There was a thread some time ago about whether blocks should be correlated with TO. I don't know why it should be so large, unless shotblockers just tend to be 'swatters', as opposed to 'grabbers' of the ball.
The coefficient of blocked shots is larger than the coefficient of points scored because blocked shots are much more rare. Imagine two players with identical statlines. Now increase one player's blocked shots by 20%, say from 100 to 120. Using your formula, this leads to an expected increase of (120 - 100)*0.10 = 2 turnovers, which is insignificant given the error in your estimates. Now instead of increasing blocks by 20%, increase scoring by 20%, say from 1000 to 1200. This leads to an expected increase of (1200 - 1000)*0.08 = 16 turnovers, a much larger change.
_________________
Regards,
Justin Kubatko
Basketball-Reference.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
jkubatko
Joined: 05 Jan 2005
Posts: 702
Location: Columbus, OH
PostPosted: Fri Jan 19, 2007 11:08 am Post subject: Reply with quote
I meant to mention above that the size of the estimated coefficient for a particular variable does not tell you much about the importance of that variable. You need to consider the error in the estimated coefficient. For example, my guess is that if we constructed a 95% confidence interval for the true coefficient of blocked shots, that interval would contain zero. In other words, the true coefficient of blocked shots is not statistically significantly different from zero at the 5% level of significance. A 95% confidence interval for the true coefficient of points scored, however, would *not* contain zero.
_________________
Regards,
Justin Kubatko
Basketball-Reference.com
Back to top
View user's profile Send private message Send e-mail Visit poster's website
AJax
Joined: 22 Jan 2006
Posts: 49
Location: The Prairies of the Great Middle West
PostPosted: Fri Jan 19, 2007 11:26 am Post subject: Reply with quote
I would guess that minutes are also not significant at the 95% level of confidence. Partly because the coefficient is small (-.005), and partly because it seems unlikely that turnovers decrease with more minutes played.
Mike, do you have the standard errors for these estimates from the regression?
It seems like the best predictor of "expected turnovers" would be touches, and maybe an interaction variable of touches and position (G/F/C). Without full data on touches, the linear regression model is probably best.
One additional question I have about turnovers: if a player shoots an airball and the ball falls out of bounds, is that a turnover? Similarly, on shots where the ball bounces over the backboard, does that get recorded as a turnover? In general, is a missed shot with no rebound = 1 turnover?
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3590
Location: Hendersonville, NC
PostPosted: Mon Jan 22, 2007 2:39 pm Post subject: Reply with quote
Mike G wrote:
ExpTO = .026*2FGA - .007*3FGA + .162*FTA + .090*OReb -.051*DReb + .167*Ast + .036*Stl + .018*Blk + .176*PF + .0167*Pts
... it actually misses whether a guy is productive with his shot attempts. Nash and Howard are very efficient, others less so. Just using 'points' (instead of FGA and FTA) would give us a better grasp of 'tradeoff' with TO...
.
OK, it occurred to me that I had included 3FGA with (all) FGA; so I've separated 2FGA from 3FGA. Also, as quoted above, I've added Points (while retaining FGA) so as to separate shot frequency and shot efficiency.
Most likely, there's a correlation between a given player's FG% and his TO rate, in that a low-% bomb or fallaway has less chance of a TO than, say, a drive to the hoop in traffic.
With the adjustments, Steve Nash is the one 'top ten' TO guy to fall out. Rather than having 28% more TO than expected, he's merely at 121%.
I don't know about regressions. I just have player stats on a spreadsheet, multiply them by factors at the tops of my columns and add them up (calling this 'expected TO'); another column is the absolute value of the difference (TO-ExpTO).
I just adjust the variables up/down, back and forth, getting closest fit between player total Actual-Expected diff. I could get just about as good a fit if I throw out the 3FGA, Stl, and Blk terms, and further fine-tune the others. As Justin points out, they're small.
_________________
`
36% of all statistics are wrong