Player Impact Ratings

Most of you are probably aware of the "adjusted +/-" stats that are becoming more popular of late. Simply put, the idea is to try to guage a player's impact based on what his team does with respect to the opposing team while he's on the floor, taking into account other 9 guys sharing the court. In theory, its nice because its not always clear how individual statistics translate to team success, and there are also a lot of holes with tracked individual stats.

For example, if a player sets a good screen or him being a threat from distance stretches the defense enough for other players to get good looks, that can lead to improved team statistics. But he doesn't necessarily get credit for it with an individual stat. The advantages of looking at things this way (based on what the team as a whole does when he's on the floor) are even more obvious on the defensive side. There are myriad ways a player can help or hurt his team defensively, and steals/blocks/rebounds/fouls don't come close to capturing it all.

The limitation, of course, is that do be able to tease out which players are actually making a positive or negative difference from team stats, you can't just look at data over a few games. But with enough possessions to look at, the good players start to separate from the not-so-good players. This should be evident below.

And the cool thing is you don't just have to look at player impact on the scoring margin (that's what traditional adjusted +/- does). Using similar techniques, you can break it down into impact on offense and impact on defense. This corresponds to point-differential, offensive efficiency, and defensive efficiency for team stats, respecitively. And just as you could break down team stats even further, you can also try to analyze player impact at a more detailed level.

HoopNumbers earlier published adjusted +/- ratings, broken down into offense and defensive components, based on the 06/07-08/09 seasons. These numbers were pretty interesting to me, but recently they have created ratings based on "four factors" impact. For those that don't know what I'm talking about, the "Four Factors" of basketball (coined by Dean Oliver, now an analyst for the Nuggets) refer to 4 components of any good offense or defense:

(1) efficiency from the field (EFG%, DEFG%)
(2) rebounding (OREB%, DREB%)
(3) turnover rate (TOV%, DTOV%)
(4) made free throws per possession (FTM-r, DFTM-r)

So, you want to know which players have the most impact on their team's field goal efficiency? Or which players make their teams more/less turnover prone? When putting together player combinations these are the types of things a coach is interested in after all, aren't they?

With that lengthy introduction, here are the recent results from Hoop Numbers (right now, based on an equal weighting of games from 06/07-08/09). Next to each "impact rating" is a percentile rank for the player (in parentheses).

Offense:
http://hoopnumbers.com/allAnalysisV...n=False&leaders=True&year=2009_multi_year_off

Best


Worst



Defense:
http://hoopnumbers.com/allAnalysisV...n=False&leaders=True&year=2009_multi_year_def

Best


Worst


Hoop Numbers also has pages which ranks players impact along all the four factors. So you can see which players have particularly strong impact on rebounding (both good and bad), for instance. I think think its fairly interesting to go through.

 

Here's one way in which the above type of analysis is instructive. Chris Kaman being amongst the worst offensive players, according to adjusted +/-, sort of jumped out at me. But when its broken down into the 4 factors components, one can try to make more sense out of it.

His impact on field goal efficiency for his team is actually positive (59th percentile). He doesn't have a strong impact on the offensive rebounds however (only 40th percentile), which seems a little strange. Its the other two factors where his impact is very negative, though. With him on the floor, his team does not get to the line well. For a guy who posts up as often as he does, Kaman doesn't get a lot of free throws, and him clogging the paint could perhaps limit dribble penetration and foul drawing opportunities for his teammates (e.g., Corey Maggette?). And then there's the turnovers. The Clippers are apparently much more turnover-prone when playing with Kaman. (Yao, incidentally, is another skilled post player who does poorly by this TOV% metric -- he's at the bottom percentile).

So, put it altogether, and these results suggest that the Clippers need to rethink how they're using Kaman. That, or get players that fit better around him.

 

Another interesting case is Dwyane Wade, who didn't show up as a bottom 20 defensive player, but nevertheless rated at 296 out of 388 players over the prior 3 seasons (again, equally weighted). One looks at his steals and certain athletic defensive plays he makes, and they may get the impression that he's a very good defender. Coaches apparently did, and they made him second team All-Defense.

But based on this analysis, he's at 15th percentile in limiting field goal efficiency, 6th percentile in limiting free throws, and only 14th percentile in positively impacting defensive rebounding. Where he excels, as expected, is in forcing turnovers. He's in the 93rd percentile in impacting forced turnovers. The rebounding is a little strange, because usually we think of Wade as a superb rebounder at the guard position. Is it possible that his gambling puts him out of position so much that his teammates get out of position due to covering for him, thus yielding more offensive boards? It might be something to look for.

 

He didn't. He only played 1300 minutes in the first of the three years this is looking at (06/07). It appears he may have had a very positive defensive impact on a poor defensive team that season, playing off the bench.

 

Another surprising rating for me was Matt Barnes. He has a good defensive reputation. I decided to take a look at Hollinger's take from his Insider scouting report, and his observation looks to provide a possible explanation:

[rquoter]
Scouting report: Barnes is only 6-foot-7, 226 pounds but has played the majority of his minutes the past three seasons as an undersized power forward. That has made him a problematic matchup at the offensive end because of his ability to shoot from distance and set up teammates off the dribble. But it came at a price. Barnes was completely overmatched physically at the defensive end. Though he's tough and competes, he doesn't have the athleticism to make up for such gross disparities, and bigger frontcourt players took advantage of him on the blocks and under the glass.

At small forward, it's more of a fair fight. He can guard most 3s and do a decent job, and he should move back into that role this season with the Magic.
[/rquoter]

 

He wrote that he's hoping to have results weighted more towards last year soon.

http://hoopnumbers.com/allAnalysisView?analysis=fourFactorsRAPM&discussion=True