APBRmetrics

Code: Select all

" thing, so I put them in this excel file > http://www.megaupload.com/?d=QCC7MFUN

Please download and open the file... ready? fine. As you can see in Page 1 the first 16 players of the index are from the Golden State Warriors, so this doesnt tell us much about the impact of one guy on his team. In order to correct that, I compare each Player Pace Index to his Team Pace. For example, Player A has a 95 PPI and his team a 92 Pace, so his real pace impact on his team would be +3.
You can find this last stat on "Page 2" of the Excel file.

For the lazy ones:

* Stephen Jackson (GSW) has the highest Player Pace Index with 102,3.
* Greg Oden (POR) the lowest at only 85,6.
* Travis Outlaw (POR) the highest impact on his team pace: +5,7
* Etan Thomas (OKC) the lowest: -4,7


IF YOU HAVE ANY TROUBLE DOWNLOADING THE FILE, PLASE TELL ME.

IM ALL OPEN TO SUGGESTIONS, CORRECTIONS, OPINIONS AND CRITICS. Smile Smile
Back to top 	
View user's profile Send private message Send e-mail 	 	
BobboFitos



Joined: 21 Feb 2009
Posts: 201
Location: Cambridge, MA
	
PostPosted: Mon Nov 08, 2010 11:19 am    Post subject: 	Reply with quote
I like the idea here. The big problem with pace is separating individual player contributions towards faster/slower from coaching, since pace seems to be one of the primary/few aspects that coaching can increase/decrease.
_________________
http://pointsperpossession.com/

@PPPBasketball
Back to top 	
View user's profile Send private message 	 	
Crow



Joined: 20 Jan 2009
Posts: 821

	
PostPosted: Mon Nov 08, 2010 4:19 pm    Post subject: 	Reply with quote
Thanks for the file. I don't think it has been summarized in public anywhere else. (Queencityhoops.com has player pace for the current season in each player writeup but no link to historical data that I know of and no summaries.)

Player pace index relative to team average is interesting though it is affected by what lineups you are used in / the spot in the rotation.

It is not necessarily the same as player impact on pace.

I would think it would be possible to use the Adjusted +/- methodology and find estimated Player Impact on pace. Every segment gets a pace scoring and you see how it moves depending on who all is on the court.


Rob (BobboFitos) makes a good point about coaching impact.

Though you could look at player pace effects across a range of years and it will in some cases go across different coaches and allow some speculation about coaching impact by looking at the raw player pace index variations across those seasons / roles / coaches.

Or you could perhaps do a multi-season Adjusted +/- style regression analysis and maybe include the coach as the 6th force present (in impact) on the court?

If the Adjusted +/- methodology is not used, there might still be somewhat useful information obtained by looking at player pace variation game to game when on the court against high, middle and low paced teams and play by play when on the court against high, middle and low "expected" paced lineups based on the net difference in team sums of player pace indexes. You could look at the correlations between the player pace indexes and actual pace for those subsets. (Or do the same things while using the Adjusted +/- methodology.)

Pace impact can be its partly its own thing (just plays taking more or less time, all else equal) but a lot of pace impact will be player impact on the 4 factors. More turnovers probably tend to quicken pace, offensive rebounds will really extend it, trips to the free throw line might be more likely to quicken it but it depends on a players distribution of when they get fouled on the shot clock. If the Adjusted +/- method is applied to pace, it would also be good to compare it to the Adjusted +/- for the 4 factors and see how much they explain of the total estimated pace impact.

I guess another part of all this is whether you want to look at game pace (offense and defense combined) or just offensive pace or perhaps defensive pace would be interesting, to a degree, also.

Last edited by Crow on Mon Nov 08, 2010 4:55 pm; edited 3 times in total
Back to top 	
View user's profile Send private message 	 	
acollard



Joined: 22 Sep 2010
Posts: 56
Location: MA
	
PostPosted: Mon Nov 08, 2010 4:34 pm    Post subject: 	Reply with quote
I think the need for adjusting these numbers is valid. But ignorning that, even if we could get a perfect number that showed exactly how a player affects his team's pace, what would that tell us? What further insight could be determined from this number?

I'm trying to brainstorm some and I'm not sure I can figure out any good ones. What I'm trying to say is I can't think of any practical applications of this data off the top of my head. Any ideas?

Maybe play similar PPI players together so they are compatible? Maybe play opposite PPI players together so they balance out?
Back to top 	
View user's profile Send private message 	 	
Crow



Joined: 20 Jan 2009
Posts: 821

	
PostPosted: Mon Nov 08, 2010 4:49 pm    Post subject: 	Reply with quote
You can play similar players together so they are compatible or play opposite players together so they balance out but I'd think you'd want to look at performance in the context of opponent / lineup match-ups. You might want custom paces for lineups and match-ups based on past results in similar circumstances instead of taking basically a one-choice strategy or just doing it or adjusting it by feel and immediate results within a game.


It is possible that player pace impacts will have some 4 factor impacts beyond what a static 4 factor or other analysis by themselves suggest.


Using the file about 6% of players have a +2 or more and about 6% of players have a -2 or less. Player pace index isn't separating them out much. Adjusted Pace estimates might be more dramatic or not.

Among those +2 or more it appears point guards are most highly represented, followed by bigs then wings. Among those -2 or less it appears not surprisingly that bigs are most highly represented, followed by point guards then wings. I think that fits with conventional visual based wisdom. Point guards can prefer to play fast or slow. Bigs pull it down a bit more often. Wings are not as often notably biased one way or another as the others.


If you wanted to get a sense of pace outside of turnover and offensive rebound impacts on offense you could calculate and just look at average time to a shot or foul.


If you are playing every angle for every drop of advantage I'd think that for any given set of average 4 factor performance for a team / lineup and its opponent (on average or preferably against relative similars) there is probably on paper, more favorable and less favorable paces. Not sure if there is truly optimal pace but you could look into the topic and see what you find and use it in some fashion and see if you do better than previously. Slightly more favorable pace might payoff slightly but that could in some cases have major impacts in playoff seeding or in the playoffs.


If you had time and reason you could perhaps also look at team and player pace impacts for home / away splits, or maybe more so impacts on later games in back to backs (and 3 in 4 days and more), impacts against superior overall quality teams vs inferior teams, by offense and defense ranking, and specifically shooting and shot defense quality.

Whatever you find, the main things in the end are what gets applied and the effectiveness.

How much research is enough? How much is too much? There will be different answers to those questions.

Last edited by Crow on Mon Nov 08, 2010 11:49 pm; edited 1 time in total
Back to top 	
View user's profile Send private message 	 	
ScoutingMachine



Joined: 02 Nov 2010
Posts: 7

	
PostPosted: Mon Nov 08, 2010 7:39 pm    Post subject: 	Reply with quote
Crow: First of all, thanks for taking your time for the responses. I really appreciate that.

I´m just a rookie in terms of advanced statistics, so if you can explain me how to use that +/- methodology it would be great and also where I can find that data.

What I can do is calculate Pace for each lineup of a team, but I dont know if this would be useful or if its already in the web. Please tell me if it is.

Anyway, if anyone wants to continue what I left and start using the +/- and the other splits you mention, feel free to do it (there is a 101% of chances you would do it better than me).
Back to top 	
View user's profile Send private message Send e-mail 	 	
Crow



Joined: 20 Jan 2009
Posts: 821

	
PostPosted: Mon Nov 08, 2010 8:07 pm    Post subject: 	Reply with quote
I am glad the responses were appreciated.


The +/- methodology is detailed here
http://www.countthebasket.com/blog/2008/06/01/calculating-adjusted-plus-minus/

The previous runs of Adjusted +/- estimates are mainly at basketballvalue.com and 82games.com.


I think pace for each lineup of a team would be useful to have available. I don't think it is already presented on the web anywhere for every lineup, but queencityhoops.com has it for each team's top 10 lineups for the current season

http://www.queencityhoops.com/teamPage.php?team=BOSTON

and basketballvalue.com has the time and the number of offensive and defensive possessions so it could be easily calculated from pages like this one

http://basketballvalue.com/teamunits.php?year=2010-2011&team=HOU

for all other lineups and for several seasons.

Perhaps pace could be added as a column there in the future, even if it will look way high or low for some lineups.

Last edited by Crow on Mon Nov 08, 2010 11:50 pm; edited 4 times in total
Back to top 	
View user's profile Send private message 	 	
ScoutingMachine



Joined: 02 Nov 2010
Posts: 7

	
PostPosted: Mon Nov 08, 2010 8:12 pm    Post subject: 	Reply with quote
Thanks for the links!

I will do the lineups pace between today and tomorrow. Lets see what we get from that...
Back to top 	
View user's profile Send private message Send e-mail 	 	
Mike G



Joined: 14 Jan 2005
Posts: 3612
Location: Hendersonville, NC
	
PostPosted: Wed Nov 10, 2010 6:33 am    Post subject: 	Reply with quote
acollard wrote:
... even if we could get a perfect number that showed exactly how a player affects his team's pace, what would that tell us?
... I can't think of any practical applications of this data off the top of my head...


If the Nets play 6% faster (poss/min) with Travis Outlaw on the floor,
and his Usage% is calculated based on the team's average pace,
then his estimated Usg% (16.0) is really 6% too high.

His adjusted Usg% would be right around 15.0 ?

This isn't exactly a major coaching decision-maker, but stat geeks want to know. There are almost twice as many players below 16% Usg, as there are below 15%.
_________________
`
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: Wed Nov 10, 2010 10:25 am    Post subject: 	Reply with quote
Yeah, that could be an issue, Mike. However, we could also calculate usage from PbP data anyway...
_________________
GodismyJudgeOK.com/DStats
Twitter.com/DSMok1
Back to top 	
View user's profile Send private message Send e-mail Visit poster's website 	 	
haralabob



Joined: 11 Apr 2007
Posts: 27

	
PostPosted: Wed Nov 10, 2010 12:21 pm    Post subject: 	Reply with quote
DSMok1 wrote:
Yeah, that could be an issue, Mike. However, we could also calculate usage from PbP data anyway...


Exactly.

If you want proper usage ratios you should be using Play by Play data to get a proper usage number.
Back to top 	
View user's profile Send private message 	 	
ScoutingMachine



Joined: 02 Nov 2010
Posts: 7

	
PostPosted: Mon Nov 15, 2010 3:37 pm    Post subject: 	Reply with quote
Here is the file with the lineups pace of last season:

http://www.megaupload.com/?d=GHD0IP60

Like in the Player Pace, I made one page to put the raw average of possesions per 48 minutes, while on the second I put the impact of that lineup in the team pace average.

The next step would be start using +/- to check in what kind of pace each teams/lineup plays best and worse.

Im open to hearn new ideas...
Back to top 	
View user's profile Send private message Send e-mail 	 	
kjb



Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
	
PostPosted: Mon Nov 15, 2010 7:39 pm    Post subject: 	Reply with quote
I did something like this using 05-06 on-off data from 82games. I didn't do everyone, just players that interested me -- 75 in total. Of that group, the Wizards changed most when Arenas was on the floor -- 6.4 possessions per 48 minutes faster when he was on the floor. The top 5 was Arenas (6.4), Iverson (5.7), Kobe (5.1), Smush Parker (5.0). Others whose teams got a lot faster when they were on the floor included: Sam Cassell, Chris Paul, Mike Bibby, Yao(!).

Teams slowed down most with Eric Snow (-4.2), Zydrunas (-2.7), Artest (-2.5), Kwame Brown (-2.3), Kirilenko (-2.2).