The disappointing TrailBlazers and the myth that Win Score is rebound-centric

Bucks fans may be disappointed in their team’s performance so far, but think how Trail Blazer fans must feel.  The Trail Blazers were picked by many to be a dark horse contender in the West, but so far the only thing they are contending for is a seat on the panel of losers that annually gathers in Secaucus, New Jersey.

What the hell?  Portland returned virtually the same team that won 54 games.  Why then are they under .500%?  Who is to blame?

Here is the Win Chart so far:

2010-11
Portland Trailblazers
(12-14)
Pos
Marginal
Win Score
Win__Loss
Credits
Win Contribution
Player Winning
Percentage
M Camby
5.0 +4.62 4.0__(-0.9) +0.551 1.286%
A Miller
1.0 +1.78 2.8__0.7 +0.240 .804%
B Roy
3.1 +0.24 1.8__1.6 +0.031 .543%
R Fernandez
1.8 +0.19 0.9__0.8 +0.012 .534%
W Matthews
2.6 -1.24 0.9__2.2 -0.150 .292%
L Alridge
4.3
-1.65
0.9__3.2 -0.261
.222%
N Batum
3.3 -1.52 0.7__2.2 -0.171 .244%
A Johnson
1.0 +1.18 0.6__0.2 +0.037 .703%
J Pryzbilla
5.0 +2.83 0.3__0.0 +0.035 .983%
D Cunningham
4.0 -2.69 0.1__1.7 -0.190 .046%
P Mills
1.0 -3.31 0.0__0.5 -0.056 (-.059)%
S Marks
5.0 -3.60 0.0__0.3 -0.047 (-.108)%
F Oberto
4.8 -6.75 (-0.1)__0.3 -0.048 (-.643)%


MWS Record: 12.9__13.1 Total Win Contribution
-0.017



Pythagorean
Record:
12.1__13.9 Projected Record
(41.0__41.0)

As you can see, many of the team’s mainstays are having down seasons, save for point guard Andre Miller and center Marcus Camby.  Miller’s numbers are up, and Camby’s numbers are virtually identical.

To isolate who deserves the most blame for the team’s reversal of fortune, I took each player’s Marginal Win Score from last season and punched it into the Win Contribution formula for this season.  Then I compared that number to the particular player’s actual Win Contribution from this season.  As a team, the Blazers total Win Contribution is down by -0.901, which over an entire season is roughly equivalent to the difference between a 53.7 win team and an average team (a 41 win team).

Here are the biggest comparative WC declines so far:

1. N Batum (-0.524)
2. L Alridge (-0.335)
3. B Roy (-0.181)
4. W Matthews (-0.163)
5. D Cunningham (-0.159)
6. R Fernandez (-0.117)
7. M Camby (-0.031)
8. J Pryzbilla (+0.015)
9. A Miller (|+0.102)

As you can see, while many people are primarily blaming the decline in performance on G/F Brandon Roy, MWS argues that a great deal more blame should actually fall on SF Nicolas Batum and PF LaMarcus Alridge.  Which brings me to my subtopic.

I need to stop now and make a point here.   Many people wrongly argue that the Win Score basketball metric overvalues rebounding, or favors rebounders, and somehow punishes scorers, especially “high usage” scorers.  That is simply not the case.

Win Score penalizes inefficient scorers, and especially high usage inefficient scorers because they burn possessions without producing points.  But with its other hand it rewards efficient scorers, and none more than high usage efficient scorers.  Thus Win Score is agnostic on the issue of usage rates.  What Win Score judges players on is whether they can put the ball into the basket at a regular rate.  It simply penalizes those who cannot, and it especially penalizes those who cannot and yet keep shooting nonetheless.

The declining Blazers make this point perfectly.  As noted, the blame here is placed to a larger extent on Batum and Alridge, while Roy is blamed to a lesser extent.  Yet if you compare career statistics for each of the three player you will notice that the secondary statistics (non-scoring related statistics) for Batum and Alridge are virtually the same as last season, particularly each player’s rebounding rate. At the same time, Brandon Roy’s rebounding and assist rates have each declined, with his rebounding having declined substantially.  Yet, according to Win Score, Roy does not shoulder as much of the blame for the Blazers slow start.

The reason is simple.  Comparative shooting efficiencies.  Roy’s shooting efficiency is down, to be sure, but not nearly so much as Alridge’s, and not even close to Batum’s.  Batum’s shooting accuracy is way off.  And Win Score penalizes him harshly for it, this despite his low usage.

Thus the idea that Win Score is a rebounding myopic statistic is pure urban legend, born from what logician’s refer to as the “exceptional case” fallacy (or, the “accident fallacy”).  Critics will often try to build their cases against Win Score on the backs of well-known “role players” like Ben Wallace and Dennis Rodman who happen to produce very high wins under the Win Score metrics   This, the critics will argue, is proof that the metric produces absurd results that any knowledgable critic can recognize as foolish, and therefore the metric cannot be considered reliable. (the classic straw man argument against Win Score seen repeatedly on-line  begins with a critic pointing out that Win Score credited Rodman with more wins in 1997 than Michael Jordan and then ends with the critic comparing the two players strictly on the basis of their basketball skills rather than their statistical production in that given season.)

But, as I pointed out, these critics point make their  cases on the backs of players whose “one dimension” is far above average.  That’s because those are the only kind of players who can produce elite level wins without being well-rounded players.

The fact is, in general, Win Score punishes players who are one-dimensional and who cannot shoot accurately.  In fact, unless the player is on par with a Rodman or Wallace in one particular area, or a Kidd in several non-scoring areas, he will never qualify as a high win producer.  And now back to the main topic.

With that aside, we return to the main topic, the Blazers.  I wrap this discussion by pointing out that the shooting problem that has crippled the Blazers could turn out well for the Blazers.  No other statistic in basketball is quite as erratic (which is why the Bucks were stupid to overprice scorer John Salmons on the open market, and ditto for the Pistons and Ben Gordon).  That means Batum and Alridge and Roy and Fernandez and even Wes Matthews could suddenly get hot just as quickly as they went cold.  And that, of course, would change things in a hurry for the once promising 2010-11 Portland Trailblazers.

Advertisements

5 Responses to “The disappointing TrailBlazers and the myth that Win Score is rebound-centric”

  1. Chicago Tim Says:

    I think it’s hard to convince people that high-usage players should not be any less efficient than low-usage players; i.e., that Fields is responsible for more wins than Stoudemire, or that Brewer is responsible for almost as many wins as Rose. It seems like Win Score rewards players who score only when they have easy shots, and somehow that seems unfair — except that the truly great players have lots of easy shots because of how great they play.

  2. Guy Says:

    Ty: Win score says that a player with 2 FG in 4 FGA per 48 minutes has the same productivity (as a FG shooter) as an 8-for-16 shooter. Or 12-for-24. Or 15-for-30. Is that what you mean by being “agnostic” on usage? And do you think those performances all have equal value in the NBA?

    And the argument that win score or WP overvalues rebounds does not in any way rest on extreme players. It rests on the fact that players’ WP is very highly correlated with rebounding, much more than with shooting efficiency. So WP says that rebounding is much, much more important than efficiency in distinguishing productive from unproductive players. However, rebounds are much, much LESS important than shooting efficiency when it comes to actually winning games in the NBA — the main difference between winning and losing teams is shooting efficiency, not rebounding (and team WP confirms the same story). THAT is why people believe win score overvalues rebounds — because it says that rebounds are the most important source of player productivity, while we know that is not how teams mainly win games. I’m afraid that the notion that this claim is based only on extreme players like Rodman is, well, an urban legend.

    • Alvy Says:

      ” * Points per field goal attempt: 5.2%
      * Rebounds: 3.2%
      * Free throw percentage: 1.2%
      * Personal fouls: -1.1%
      * Assists: 1%
      * Turnovers: -0.9%
      * Steals: 0.7%
      * Blocked shots: 0.2%

      Rebounding certainly matters. After all, getting and keeping possession of the ball is important; and rebounds are the primary way a team gains possession (without letting the other team score). But WP48 is more “responsive” to shooting efficiency from the field. A 1% change in points per field goal attempt (or adjusted field goal percentage times two) leads to a 5.2% change in WP48. ”

      http://dberri.wordpress.com/2010/12/08/how-about-those-jazz-and-answers-to-other-questions/

    • Guy Says:

      Alvy:
      These elasticities tell us very little about how big a role rebounds play in determining players’ WP48. For that purpose, one uses standardized coefficients not elasticities. And standardized coefficients tell us that rebounds play a much, much larger role than efficiency in determining players’ WP48. If I told you every players’ Reb48, and I knew their eFG% (or any efficiency metric you want), and we both tried to predict players’ WP48, you would kick my butt. Knowing REb48 tells you vastly more about a player’s WP48.

      But for teams, the reverse is true. Knowing rebounds tells you only a little about who wins and who loses. Right now, the worst 10 teams in the NBA are out-rebounding the best 10 teams! That never happens with shooting efficiency, which is what actually drives wins in the NBA. Do you have an explanation for this massive discrepancy between the story told by WP48 and what actually produces wins?

      And I have to say it’s a bit unkind of you to reproduce this particular post of Dr. Berri’s, which makes him look like a statistical illiterate. I’m sure he’s rather embarrassed about it at this point. So I think we should just forget about this mistaken use of elasticities (and also his confusion about when to use coefficient of variation), and focus on the real issues.

  3. Austin Says:

    Let’s compare scoring efficiency statistics from 09-10 and 10-11.

    Roy: TS% of .568 to .501
    Aldridge: TS% of .535 to .512
    Batum: TS% of .646 to .521

    So clearly Batum is slumping, regressing strongly towards the mean, or both. But Roy’s scoring efficiency has dropped far more than Aldridge’s has, along with some other stats. In fact Roy’s WP48 has also dropped by way more than Aldridge’s. So as someone who doesn’t regularly read this blog, I am extremely confused by this analysis.

    Also, anyone watching the Blazers can tell you that Aldridge has actually been marginally better at defense this year, whereas Roy has been a complete liability – so if counterpart production is the issue here, then this analysis is failing to take into account Roy’s contributions (or lack thereof) on defense (perhaps complicated by the fact that Batum and Matthews always guard the other team’s best wing scorers).

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s


%d bloggers like this: