In a previous post, I deconstructed the MVP race for fun using a consolidated concensus methodology of advanced stats (Editor Arturo’s note- try saying that three times fast) .
Today, after being oh so serious for a week, I reverted to form and decided to have a bit of fun with that method. I’ll do that by ranking every player for every team in the NBA for the 2010-2011 season. The method combines my three favorite metrics for player evaluation:
- The Classic Wins Produced stats courtesy of Prof. Dave Berri (explained here).
- My own opponent adjusted Wins Produced based stat Wins Produced Combined (explained here).
- Basketball Reference‘s Win Shares (credit Justin Kubatko and Dean Oliver)
- The average of all three “Win” metrics. (add them together and divide by three)
I chose these metrics because
- They all have a high correlation to wins at the team level.
- We want to measure wins when evaluating players.
- It’s fun to see how players do in different advanced metrics.
Everyone then is ranked by Average Wins, players are shown for their team at seasons end and only player’s with 400 minutes or more are shown.
Let’s get to that creamy center shall we?
I was going to give you a lot more here but a heavy editorial hand came down. You’ll have to wait.
Much like NBA players, I am being oppressed by the system.
Thankfully, waiting is good for the soul.
-Arturo
Editor Dre’s Note. I love Arturo’s posters but they get a bit long for my liking in the blog section. After a long debate with Arturo I’ve decided to make a spot on the Wages of Wins Journal for just cool posters:
http://wagesofwins.net/cool-posters/
To make things fun though if any of you out there have cool posters you can forward them along to us and if we think they’re cool we’ll post them in the Cool Posters area and talk about them. Sound good? I hope so!
Wages of Wins Journal 
brgulker
July 12, 2011
Just FYI,
Link is broken.
(Editor Arturo’s note- Try again Ben. It works for me)
dsal1951
July 12, 2011
Hey Arturo, this is my first time every posting here and I really enjoy/appreciate all of the work that you guys do. I was wondering if you had any thoughts about why Wins Produced was consistently higher than Win Shares for the best player on each team (I quickly counted 27 of 30 teams)? The most dramatic example that I noticed was Kevin Love (25 to 11.4), but there were several others that had a sizable difference. I know that Win Share isn’t your stat of choice, but I am interested in your thoughts.
Arturo Galletti
July 12, 2011
dsal1951,
Good question.
One of the pieces we’ve got on tap is a decontruction on Win Shares that explains how it’s calculated. To give you the short version, Win Shares has a much bigger functional team adjustment. What that means is that you get a bigger credit/penalty for the team you play on. That means player’s cluster around a tighter mean with a smaller distribution. This approach assumes a larger role for intangibles and attempts to capture it thru that adjustment.
greyberger
July 12, 2011
Intangible defense, I would say, or hard-to-tang defense (Editor’s Note: Interesting choice of words there, I like it!)
Win shares is based on offensive and defensive ratings for players from the Dean Oliver book Basketball on Paper. With defensive ratings, players share credit/blame for being on a good/bad defensive team. A player starts with the team’s defensive efficiency and that number is adjusted up or down based on defensive box stats.
Wins Produced does it the other way around – box score stats are the core of how a player is evaluated on defense, and team performance adjusts that number up or down. For the most part, though, WP is more reflective of a player’s box score contributions on defense and Defensive Ratings and Defensive Win Shares are more based on team performance.
(Editor’s Note: This is a really great explanation of the difference between Wins Produced and Win Shares, thanks Grey!)
ges05a
July 12, 2011
I’m stoked about the “Cool Posters” section! I will be following that section as closely as the actual blog posts in future.
Also,it’s interesting to note how many of the players which WP rates differently than does the conventional wisdom also come out looking pretty good with Win Shares. I notice especially the fact that Tyson Chandler’s seismic impact on the Mavs’ defense is well-reflected in Win Shares, and that several accurate shooters who are poor defenders still come out pretty well (Kevin Love and David Lee stand out).
Parker Flowers
July 13, 2011
I know that if I were more astute I should be able to figure this out on my own, but I have been tracing back the links for a long time and cannot seem to figure out the definition of losses produced by opponent. You say that it is, “Opponent WP48 for each player by game based on time by position (this is Wins Produced by team based on deviation of the opponent from average). Used opponent WP48 to generate losses Produced by opponent.”
So if losses produced by opponent is the deviation of each opposing player from average, shouldn’t bad defensive players have negative opponent losses produced? I know that this way of conceiving Losses Produced by Opponent HAS to be wrong, but I can’t determine what it actually means or why averaging it with Wins Produced results in a valid combined statistic. Could you describe it for perhaps a few sentences?
Parker Flowers
July 13, 2011
Oh.. nevermind. I stupidly didn’t think this out. The average player has produced something like .1 losses per 48 minutes so a player would have to be particularly bad defensively to have a negative losses produced by opponent (LPO) value. Do you have any sense of how much more dependent a player’s LPO48 is on the team context than WP48—that is, is a player’s LP048 more likely to change than WP48 after a trade? Do you have this chart available as an excel file?