Milwaukee Bucks naked Win Numbers

Be careful about basketball win numbers produced by basketball metrics.  As Professor Berri has noted, as long as you tie them to offensive and defensive efficiency they will look accurate.  But that does not neccessarily mean they accurately explain wins.

Let me give you an example of how its done.  I simply took the 2009-10 Milwaukee Bucks offensive and defensive efficiency ratings posted on basketball-reference and converted them into wins, using Professor Berri’s efficiency formula and the corresponding winning percentage.  Its easy, and you come up with the same number I come up with using my Marginal Win Score (43.6 wins using the Naked Formula, 44.0 wins using MWS48).  The first set of numbers are the player’s Wins and Losses credited based on the numbers.  The second are MWS48.

Here’s what I got:

1.  Andrew Bogut……7.2__2.0  (7.9)
2. Luke Ridnour…….5.9__1.3 (5.1)
3. Ersan Ilyasova…….5.8__2.0  (4.2)
4. Luc RMB Moute…….5.8__1.9 (4.2)
5. Carlos Delfino…….4.6__4.7 (6.4)
6. Brandon Jennings……4.1__6.9 (5.1)
7. John Salmons…….3.8__0.8 (3.6)
8. Hakim Warrick…….2.5__1.5 (1.7)
9. Kurt Thomas……2.0__2.3 (1.8)
10. Charlie Bell…..2.0__4.6 (1.9)
11. Jerry Stackhouse……1.1__2.4 (1.4)
12. Michael Redd…..0.6__1.4 (0.3)
13. Dan Gadzuric……0.5__0.8 (0.1)
14. Jodie Meeks…..0.2__1.8 (0.3)

The point I’m trying to make is this.  Just because a metric adds up doesn’t mean its describing wins properly — and that includes my metric!  What I did above was simply give the player credit for producing wins simply based on his offensive efficiency and the team’s defensive efficiency while he was on the floor.  These are true win “shares”. 

But are they correctly identifying the players who took the actions that produced the wins.  I don’t think so. 

I generally like the Win Shares method.  Unlike other more popular metrics (PER), at least WS correlates somewhat with wins.  But as I showed above, that doesn’t mean its right. All it means is that they’ve wisely decided to disperse wins according to efficiency differential. 

What I don’t think they do is they do not credit those that get the ball so the scorers can score the ball.  I am convinced that Professor Berri is right about the value of possessions (not rebounds… as the Message Board Birds like to claim… its possessions that he recognizes have value).  Win Shares disagrees.  Win Shares follows an earlier, incredibly complex mathematical offensive rating formula developed by Dean Oliver.

As a result they rate players who relied heavily on possession creation to produce wins, players like Bill Russell, Oscar Robertson and Larry Bird, behind Dirk Nowitzki on their all time Win Share production list.  If you weight scoring too heavily, you get this result.

I know its illogical to argue that something is wrong simply because you disagree with the results it produces, I only bring those examples up for your consideration. 

But the point of this is, be discerning.  Be discerning with the metric I use, MWS48, be discerning with Win Shares, whatever.  Just because a metric can produce accurate win totals doesn’t necessarily mean its dead accurate.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: