82games.com has suddenly reintroduced their “Counterpart Opponent” statistics after a bizarre episode where they pulled them completely (I’ve been warned though and am actively looking to produce a better version).
So as long as they are there for now, I used them to calculate this season’s most valuable rookie using Win Contribution (see Pages on the right of this blog for full explanation).
Its not Brandon Jennings. And its not anyone you think it is unless you’ve read this blog dating back to last summer’s draft. Its San Antonio Spurs rookie DaJuan Blair. He won going away. First a couple of caveats.
Candidate Pool Limited to BasketballProspectus All-Rookies
I didn’t want to spend all night calculating so the pool of candidates was restricted to those rookies that were named to either BasketballProspectus’ All-Rookie first or second team. There may be other rookies with higher Win Contributions than some of the bottom of the rankings, but these ten are probably the most deserving. Which brings me to my second point.
The Limits of Win Contribution
This kind of ranking illustrates the functional limit of Win Contribution. It fails as a comparative around +0.000, in this sense. Any positive marginal win score trumps any negative marginal win score. Thus, Darren Collison has a slightly negative Marginal Win Score and thus has an overall negative win contribution. Meaning, his overall contributions made the Hornets a slightly below .500 team.
The problem is, some rookie could come off the bench and play 300 minutes, and if he finishes with anything positive, he appears more valuable than Darren Collison, whether he is or isn’t.
The problem would be solved with a proper “replacement level” for basketball. That way the “zero” mark is set lower. Thus, in theory, you would be comparing the performances against what you could get from any old schmoo (because in sports, as in most things, “average” is not synonymous with “the exact middle”).
Here’s the problem. What the hell is replacement level in basketball? In baseball and football its obvious… the guy on the bench waiting to replace the guy in the game.
In basketball, though, a “bench” player is not the equivalent of a baseball “bench” player. A basketball bench player could be a second stringer like, say, the Milwaukee Bucks Ersan Ilyasova, who virtually shares time with the starter. That’s not traditional “replacement”. So where do you set it?
If you set it below second string, then you have the following problem Its no longer true “replacement” level. When Bucks center Andrew Bogut got hurt, he wasn’t replaced by Dan Gadzuric, he was replaced by Kurt Thomas, who in turn was replaced by Dan Gadzuric.
So you have the definitional problem. But lets say you solve that one. Then the further problem is the unique values that correspond to the different positions in basketball.
As the Bucks are finding out, there’s a monumental difference between the “replacement value” at the center position and the “replacement value” at shooting guard. So you’d have to account for that.
But I’ve gotten way off the track. The central point is that I’m sure there are rookies with higher “Win Contributions” than Brandon Jennings, but I don’t think they belong ahead of him in the Rookie Rankings just because of it.
Wrong about Curry
I thought Stephen Curry would struggle because of his build and because his college numbers were pumped up by weak competition. Neither was the case. He had an excellent season.
That said, I wouldn’t want him as anything other than a Vinnie Johnson type off the bench. You cannot build a decent defensive team without some size at the 2 guard (the Bucks found out what a difference that can make) and Curry simply isn’t a point guard.
Three that I got right
The three players I liked going into last summer’s NBA Draft finished One, Three, and Two in this spring’s NBA Rookie of the Year rankings.
Harden had the size and game you just crave in a shooting guard. He was willing to take it to the basket, and not settle for jumpers. His play, along with that of Serge Ibaka’s play and the continued ascenscion of Kevin Durant all combined to make Oklahoma City a surprise playoff team in the Western Conference.
Both Lawson and Blair fell well further down in the draft than their resumes warranted. Thus each went to an established playoff team where their time was limited. Nevertheless each made the top 3 by unsurprisingly continuing the stellar production they showed in college.
Blair continued to be an absolute beast on the backboards. His demolition of the Los Angeles Lakers on the glass in one of TNT’s final regular season telecasts brought an audible gasp from color commentator Doug Collins. He is undersized for the center position, and his contribution to team defense is a bit uncertain because of it, but his personal production cannot be questioned.
As for Lawson, he was the strangest NBA draft prospect I can remember. How could his college resume have been better? What more could he have done? And he almost dominated the Final Four. Yet he slipped so effortlessly down to Denver in the middle of the first round.
Making it even more strange is his strong suit is efficient scoring, something normally craved by NBA general managers. His secondary production, excluding steals, was about break even across the board. He excelled in his scoring efficiency, posting an eFG% of over 55%, outstanding for a point guard.
I still don’t understand how he fell like he did. Maybe it was his lack of knee ligaments. Oh no, that was Blair’s supposed problem.
For more on how NBA GM’s struggle with rational decisions regarding the NBA draft, check out Stumbling on Wins.
Footnote: Bucks fans may notice there is a discrepancy between the numbers for Brandon Jennings and the numbers posted for him on the Milwaukee Bucks Win Chart. He has about 0.4 more wins than he does on the Bucks chart.
That’s because the Bucks Win Chart is about 45% “Level One Marginal Win Score”, meaning 45% of the statistics are taken directly from NBA transcripts, and 55% “Level Two Marginal Win Score” meaning 55% of the statistics for him come from 82games.com.
Since every other candidate’s numbers are pure “Level Two”, I used pure “Level Two” for Jennings as well. It still didn’t get him to ROY.