APBRmetrics

The discussion of the analysis of basketball through objective evidence, especially basketball statistics.
It is currently Wed Jul 30, 2014 5:10 pm

All times are UTC




Post new topic Reply to topic  [ 70 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next

Who is 2012 Rookie of the Year?
Marshon Brooks, NJN 0%  0%  [ 0 ]
Kyrie Irving, Cle 56%  56%  [ 5 ]
Brandon Knight, Det 0%  0%  [ 0 ]
Kawhi Leonard, SAS 0%  0%  [ 0 ]
Klay Thompson, GSW 11%  11%  [ 1 ]
Isaiah Thomas, Sac 0%  0%  [ 0 ]
Chandler Parsons, Hou 0%  0%  [ 0 ]
Ricky Rubio, Min 11%  11%  [ 1 ]
Kenneth Faried, Den 22%  22%  [ 2 ]
Kemba Walker, Cha 0%  0%  [ 0 ]
Total votes : 9
Author Message
 Post subject: Re: 2011-12 Rookies
PostPosted: Thu Jan 05, 2012 9:30 pm 
Offline

Joined: Thu Apr 14, 2011 10:41 pm
Posts: 814
Location: Hotlanta
mystic wrote:
Evan, USG is bascially included via points created. I calculate a value I call OFEF, which are the created points per 100 offensive possessions. Irving has 32.98 right now and is listed 10th in the league so far. That value is multiplied by EF%, which is just the combination of TS% and To-R. Irving has 0.461 and is listed 143th in the league. The combined value goes into the formula as offensive value, for Irving it is 15.21, 25th in the league (32.98*0.461 ~ 15.21). The leader in offensive value right now is Manu Ginobili with 24.78. The highest value for a whole season since 1985 has LeBron James in 2010 with 23.26. He has also the 2nd highest with 22.99 in 2009. After that Chris Paul in 2009 with 22.58 and Dwyane Wade in 2009 with 22.38. Followed by Michael Jordan with 21.95 in 1990.


Thanks, interesting.

_________________
The City: A Golden State Warriors-Centric NBA Blog
NBA WOWY: Find statistics of your favorite team with any arbitrary combination of players on or off the court.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Thu Jan 05, 2012 10:26 pm 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2210
Location: Asheville, NC
Based on several years' previous drafts, early projections of eWins on the season minus historic averages (scaled to 66 games).
Code:
steals of the 2011 draft ?   measured           expected
eW+    rookies          tm   e484  eW66  draft  eW66  e484
3.3   Marshon Brooks   NJN   1.25   4.0    25    .7   .52
3.0   Norris Cole      Mia   1.14   3.5    28    .5   .47
2.1   Kyrie Irving     Cle   2.10   7.5     1   5.3  1.35
2.1   Ricky Rubio      Min   1.12   4.2     5   2.2   .75
1.9   Greg Stiemsma    Bos   1.19   1.9         
1.7   Jeremy Pargo     Mem    .56   1.7         
1.7   Markieff Morris  Phx   1.07   2.9    13   1.2   .63
1.4   Jon Leuer        Mil   1.04   1.9    40    .5   .47
1.2   Kawhi Leonard    SAS    .93   2.1    15    .9   .59
1.1   Isaiah Thomas    Sac    .82   1.3    60    .2   .36
0.8   Chandler Parsons Hou    .96   1.3    38    .5   .47
0.8   Kemba Walker     Cha    .98   2.5     9   1.6   .68
0.8   Iman Shumpert    NYK   1.30   1.5    17    .7   .52
Topping the list, a couple of late 1st round picks. Players in this stratum of the draft typically play 7-10 minutes per game as rookies, and not nearly this well.
Stiemsma and Pargo were undrafted, and I took this to mean they'd be expected to do nothing in the NBA.

Code:
busts of the 2011 draft ?       measured           expected
eW+     rookies          tm    e484   eW66 draft  eW66  e484
-1.1   Marcus Morris     Hou   -.24   -0.1   14   1.1    .62
-1.4   Enes Kanter       Uta   1.01    2.0    3   3.4   1.00
-1.4   Bismack Biyombo   Cha    .20    0.4    7   1.8    .70
-1.5   Chris Singleton   Was   -.30   -0.8   18    .7    .52
-1.8   Klay Thompson     GSW   -.11   -0.2   10   1.5    .67
-2.0   Jan Vesely        Was                  6   2.0    .71
-2.2   Jonas Valanciunas Tor                  5   2.2    .75
-2.7   Derrick Williams  Min    .65    1.6    2   4.3   1.15
I went ahead and included a couple of high picks who haven't appeared. And others who won't get many more minutes if they continue in the negative. And still, Mr. Williams blows them away.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Thu Jan 05, 2012 10:31 pm 
Offline

Joined: Thu Apr 14, 2011 10:41 pm
Posts: 814
Location: Hotlanta
I got to learn how to pronounce that dude's name. Steams-muh?

_________________
The City: A Golden State Warriors-Centric NBA Blog
NBA WOWY: Find statistics of your favorite team with any arbitrary combination of players on or off the court.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 9:24 am 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 318
Mike G wrote:
I went ahead and included a couple of high picks who haven't appeared. And others who won't get many more minutes if they continue in the negative. And still, Mr. Williams blows them away.


Isn't such an evaluation a bit unfair? It is not like Derrick Williams is playing bad, in fact he is above average overall. He doesn't get the minutes, because Beasley and Love are getting the majority of minutes at his position. In that case, given the fact that Beasley is worse than Williams, it is more due to coaching than "Williams blows...".

In my rookie model based on the drafts from 1999 to 2010 I get him with -0.18 SPM as expectation. He has +0.49 so far. For all qualified 1st round draft picks:

Code:
        Name          Team     SPM      Exp      Dif
Ricky Rubio            MIN    3.43     -1.07    4.50
Iman Shumpert          NYK    2.01     -2.31    4.32
Alec Burks             UTA    1.87     -2.17    4.04
Markieff Morris        PHO    1.45     -2.24    3.69
Norris Cole            MIA    1.62     -1.50    3.12
Kyrie Irving           CLE    3.29     0.18     3.11
Kawhi Leonard          SAS    0.75     -2.31    3.06
Kemba Walker           CHA    -0.59    -1.85    1.26
Enes Kanter            UTA    0.70     -0.51    1.21
Marshon Brooks         NJN    -1.03    -1.81    0.78
Derrick Williams       MIN    0.49     -0.18    0.67
Tristan Thompson       CLE    -0.18    -0.80    0.62
Brandon Knight         DET    -1.48    -1.70    0.22
Jimmer Fredette        SAC    -3.47    -1.98    -1.49
Klay Thompson          GSW    -6.35    -2.09    -4.26
Bismack Biyombo        CHA    -5.82    -1.51    -4.31
Chris Singleton        WAS    -7.17    -2.29    -4.88


13 are performing above expectations, only 4 below. That is a really good result so far. Obviously, only a few games are played, thus it can still change a lot until the end of the season.

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 11:50 am 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2210
Location: Asheville, NC
Last year at Arizona, Derrick Williams made 75% of his FT, and in the NBA he's currently at 42% (5 of 12). He never gets an assist (1 per 36 min.) and he's not been aggressive getting shots (11.2 FGA/36); yet he gets turnovers (3/36).

He's an adequate rebounder, for a SF. I guess Love doesn't miss many.
Why is Beasley playing so much? Maybe they're boosting his totals to put him on the trading block?
Maybe earnestly giving him the chance to lock up the starting job?

Williams isn't doing that badly. It's more that there are a dozen or so who are overachieving expectations by a lot.
His ws/48 is just avg, his PER is below (13.9), and I don't see what his strengths are. I fully expect him to grow into his role, and probably the depth at his position is creating some frustration.

In fact, I figured a #2 pick gets 27 mpg; that would be 50% more minutes and eWins, just enough to move him off the bottom of the list above.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 12:20 pm 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 318
First of all: Williams is PF, and he gets the majority of his minutes as PF behind Love. When Williams is on the court as PF the Timberwolves are at +14.1 per 48 minutes, with Love as PF they are at -17.5. He still doesn't get the minutes. He scores at 56 TS%, while the league average is 52 TS%. He is not getting the shots, because he doesn't get the opportunities. His FT% is weak so far, but we also only saw him taking 12 FTA, I suspect an issue with the sample size much more than with his ability to convert FTA at a higher rate.

Your ew484 has him with 1.15, which is above average. That is basically similar above average as my SPM having him with +0.49.

The non-prior informed RAPM has Derrick Williams with +0.3, also slightly above average at this point. We can conclude that the high PM at PF is not just against garbage opponents in his case. Williams is solid, just isn't getting the minutes, because he is behind Love and Beasley in the rotation. If I would be the Timberwolves, I would try to trade Beasley to free up the minutes for Derrick Williams. That could give Williams 11 minutes more as PF, enough to push him to 4.5 eW66, a bit better than expectations. And I have little doubt that Williams could keep his performance level with increased minutes.
Williams' performance level in his given minutes is acceptable. The Timberwolves overall are a better offensive and defensive team with him on the court, in fact they are at +7.6 per 100 possession with him and -0.9 with him on the bench. What more would you want from a rookie?

I personally think that your way of judging a player by comparing his minutes to the expected minutes is wrong. The minute distribution is not just based upon the performance level of the specific player, but also determined by coaching and the depth situation on the respective team. On another team he would probably get more minutes.

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 1:58 pm 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2210
Location: Asheville, NC
Maybe others besides you and I will make note of the fact Williams could be getting some of Beasley's minutes, and the Wolves' coach will make that adjustment, and then it won't seem so "wrong" to point it out?

In doing these assessments for several years, a regularly noted trend is for rookies to start slowly, both in production and in minutes, and to finish stronger. We agree this rookie crop is off to an amazingly fast start. It'll be odd if they get even better, as a group, and it'll be odd if they slow down. It's an odd year.

I expected Derrick Williams, as the #2 pick, to get 26 mpg (downgraded due to crowded frontcourt) and 1.15 eW/484; I've measured him at .65, which is significantly down.

Here are the Wolves' starters and reserves, their per36 rates adjusted to team and opponent averages, estimated % vs starters, and lots of other stuff.
Code:
Wolves   pos Min  eff%   %Sta   Sco   Reb   Ast   PF   Stl    TO   Blk   e484
Love      F  39   .595   .73   26.2  14.5   1.9   2.2   1.1   3.5   .7   2.45
Beasley   F  32   .424   .78   16.2   7.3   1.4   3.5    .6   3.3   .4    .78
Ridnour   G  27   .630   .81   16.9   2.1   5.2   3.9   1.4   2.7   .2    .92
Johnson   G  22   .473   .85   10.4   3.8   1.8   2.3    .6   2.9   .5    .11
Milicic   C  20   .513   .86   13.3  11.0    .7   4.7    .6   2.5   .9    .85

Rubio     G  28   .622   .52   12.5   4.8   8.8   2.7   1.2   3.7   .2   1.12
Tolliver  C  24   .516   .49    8.0   6.8    .5   2.8    .0   1.4   .4    .07
Williams  F  18   .551   .46   13.8   7.7    .9   3.3   1.5   2.7   .0    .65
Ellington G  11   .360   .45    8.5   4.4    .5   2.9   1.4   1.0   .0    .02
Barea     G  11   .443   .48   17.0   4.3   4.3   2.6    .5   1.6   .0   1.07
Randolph  F   9   .583   .43   18.0   5.5    .0   3.6   1.2   4.2  1.6    .72
I do have Beasley ranked higher than Williams at this point.
It's odd lineup management, in which the starters have gotten just 59% of all minutes. Historic average is about 2/3.
I don't generally know why a player does/doesn't get minutes, or shots, or assists. It may be coaches, teammates, assertiveness, any combo.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 2:22 pm 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 318
Ah, I misread your table, I obviously thought your metric would measure out Williams as 1.15 ew484 player. Well, I'm a bit shocked to see a metric thinking that Beasley so far performed better than Williams.

http://bkref.com/tiny/XfHZ4

Beasley is basically worse in everything. Can you explain why your metric thinks Beasley is better?

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 3:39 pm 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2210
Location: Asheville, NC
As you see in that link, Beasley scores 15 points per 36, vs 14.6 for Williams.
Because Williams has a higher TS%, his 'equivalent' scoring rate would be higher, if all else is equal.
But Beasley starts every game, and he not only gets 34% more shots per minute, he gets them against (and alongside) about 78% starting players; Williams gets fewer shots -- backing up Love, you say -- when only 46% of those around him have been starters.

Probably they should reverse roles, and Beasley should be the instant offense guy off the bench. Beasley seems to take lots of shots, regardless of who else is in the game. Williams so far is pretty selective.

Because of their %Sta (see table above), I've multiplied Beasley's numbers by 1.03 and Williams' by .93 . This is a workable standard around the league, and maybe the Wolves are just too different, at least right now.

Beasley apparently gets rebounds against starters (and also against Love) a lot more than Williams does. Without this consideration, Williams' Reb% is 23% better; with the factors shown above, it's only 8% better.
You may think of this as 'quality' of rebound -- as a high% scorer gets points of better quality, vs a high-volume, low% scorer, like Beasley.

Because 'shot creation' has value (even if some metrics don't think so), that's Beasley's strength. Williams makes up about half the difference by being slightly better in Stl, TO, Reb.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 4:15 pm 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 318
Mike G wrote:
Because 'shot creation' has value (even if some metrics don't think so), that's Beasley's strength. Williams makes up about half the difference by being slightly better in Stl, TO, Reb.


Shot creation has value, but hardly enough value to overcome such a big difference in scoring efficiency. Williams nearly equals Beasley's scoring per minute while taking far less shots. There is a difference between creating a shot and just taking a shot. Beasley way below league average in terms of efficiency, overall he is ranked in my EF% as 264th among 292 qualified players. Going by your explanation it would mean that Beasley should even take more shots in order to help his team more, that doesn't sound like a good idea.

The starters vs. non-starters it's not really that plausible. Playing against better players makes it tougher, but at the same time playing with better players usually makes it also easier. Is there any research that really gives any evidence that there is an effect on rotation players at all? If we exclude garbage minutes, I suspect not much of a difference here. Where did you get the 1.03 and the 0.93 from?

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 5:06 pm 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2210
Location: Asheville, NC
mystic wrote:
Shot creation has value, but hardly enough value to overcome such a big difference in scoring efficiency. Williams nearly equals Beasley's scoring per minute while taking far less shots. ...
Going by your explanation it would mean that Beasley should even take more shots in order to help his team more, that doesn't sound like a good idea.
I won't argue with the first part of this comment. For whatever reason, around the league, shooting% are not so highly correlated with winning. Usually there's stronger correlation. Points are scarce (teams getting 5 ppg less than last year), and shooting% are all over the place.

Beasley wouldn't help his team by taking more shots, but he wouldn't necessarily be helping by taking fewer. Maybe he's taking a lot of bail-out shots. A 40% shot is worse than a 50% shot, but it's a lot better than no shot.
Quote:
Playing against better players makes it tougher, but at the same time playing with better players usually makes it also easier. Is there any research that really gives any evidence that there is an effect on rotation players at all?
Yes, studies have appeared here; though they'd be hard to find, after the forum meltdowns.

Among players who both start and sub a significant amount, assists and shooting% tend to be better for starters. Rebounds, steals, and blocks are easier off the bench -- presumably vs a greater % of bench players on the opposing team.

The formula estimating % of minutes vs starters (%Sta) is a bit complicated. The factor I use to compensate is:
Fac = (%Sta *3/2)^n
The constant 3/2 normalizes for the average 2/3 of minutes played by starters.
The exponent n varies but is generally around .2 (in fact exactly that right now).

This factor gives the biggest boost to players who start but play few minutes -- their opponents are almost all starters. And it most reduces those who don't start and play few minutes -- garbage players.

If you should study players' stats starting vs subbing, consider that players tend to be starters because they're playing well. Consider Greg Monroe last year:
http://www.basketball-reference.com/pla ... lits/2011/

He produced much better numbers as a starter. But he basically improved every month, before and after he became a starter in late December.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 5:45 pm 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 318
Mike G wrote:
For whatever reason, around the league, shooting% are not so highly correlated with winning. Usually there's stronger correlation.


What exactly do you want to say? Does it has a strong correlation or not?

Over the last 4 seasons linear correlation coefficient to win%:

TS%: 0.71
To-R: -0.44
EF%: 0.76

FGA per min: -0.27
FGA+0.44*FTA per min: -0.19

The correlation for TS% is much bigger than for FGA.

Mike G wrote:
Among players who both start and sub a significant amount, assists and shooting% tend to be better for starters. Rebounds, steals, and blocks are easier off the bench -- presumably vs a greater % of bench players on the opposing team.


So, basically it is just assumed that those players are achieving it against a certain type of player? And in essence it is expected that Beasley should actually have better efficiency?

Mike G wrote:
The formula estimating % of minutes vs starters (%Sta) is a bit complicated. The factor I use to compensate is:
Fac = (%Sta *3/2)^n
The constant 3/2 normalizes for the average 2/3 of minutes played by starters.
The exponent n varies but is generally around .2 (in fact exactly that right now).


Can you back that estimation up with any real data?

Mike G wrote:
If you should study players' stats starting vs subbing, consider that players tend to be starters because they're playing well. Consider Greg Monroe last year:
http://www.basketball-reference.com/pla ... lits/2011/

He produced much better numbers as a starter. But he basically improved every month, before and after he became a starter in late December.


Do you want to say that a starter has per se better boxscore stats anyway? Why would you make an adjustment in such a case? If coming from the bench will likely decrease the numbers, it seems to be useful to actually do it the other way around? If you have specific estimates for different boxscore entries, it makes more sense to adjust those boxscore entries before calculating anything else, rather than using a somehow arbitrary coefficient to adjust that, especially when you are assuming how often a player is in against starters/subs anyway.

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 6:10 pm 
Offline

Joined: Thu Apr 14, 2011 10:41 pm
Posts: 814
Location: Hotlanta
Mike G wrote:

Beasley wouldn't help his team by taking more shots, but he wouldn't necessarily be helping by taking fewer. Maybe he's taking a lot of bail-out shots. A 40% shot is worse than a 50% shot, but it's a lot better than no shot.


This might be a good argument for Monta Ellis, who takes 17% of his shots with 5 seconds to go on the shot clock. Or Carmelo who takes 22% of his shots with 5 s left. Kobe and Dirk take 13% of their shots in similar situations. Beasley only takes 5% of his shots at the end of the clock. So, I'm not sure the "bail out" excuse works for him. The data are according to 82games.

_________________
The City: A Golden State Warriors-Centric NBA Blog
NBA WOWY: Find statistics of your favorite team with any arbitrary combination of players on or off the court.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Fri Jan 06, 2012 10:13 pm 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2210
Location: Asheville, NC
mystic wrote:
Do you want to say that a starter has per se better boxscore stats anyway? Why would you make an adjustment in such a case? If coming from the bench will likely decrease the numbers,..

Starters tend to be better than non-starters. That's not to say putting a random bench guy in the starting lineup will make him have better stats. He'll probably have worse stats.

Intuitively, his FG% may improve, but he'll get fewer shots. And his scoring output is likely diminished.

Conversely, a starter who is demoted to the bench because his game is slipping may continue to have declining numbers. But the causality is that he was benched because his numbers declined, not the other way around.

No team has a 2nd unit that's better than their starting 5. And as such, if you're competing against mostly bench players, your competition is less than if you're playing vs mostly starters.

This is analogous to the argument that getting more minutes causes a player to have better (per minute) numbers. This is an inversion of causality. In truth, after he has played better, he's given more minutes.


Top
 Profile  
 
 Post subject: Re: 2011-12 Rookies
PostPosted: Sat Jan 07, 2012 11:02 am 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 318
Well, that sounds to me as if the two factors you are using for bench and starting players are on a very weak foundation.

Anyway, having Beasley as a better performer than Williams on a per minute basis would be an alarming signal for me in terms of quality of the metric overall. Especially when your explanation doesn't sound very solid (at least to me).

In case of Beasley for example, you would expect him to be even less effecient, if he would be coming off the bench. That is hardly an argument for me to give his rating a boost via those starter/sub adjustment.

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 70 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next

All times are UTC


Who is online

Users browsing this forum: Bing [Bot] and 6 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group