Implied Dissent

Monday, May 29, 2006

Misapplied credit

I think I've been able to pinpoint exactly where the The Wages of Wins guys go wrong in valuing basketball players, or at least what their largest error is. Clearly they overvalue rebounds, but that's not my insight, it seems like everyone is pointing that out. My thoughts are on why they do so. Defensive rebounds are valuable not so much because of the rebound itself (as valuable as that is), but rather because they mean that the other team missed a shot. The player who grabs the board then receives all of the credit for this missed shot in the formula, when in fact credit would probably properly go to the on-the-ball defender first, then to the rebounder, then to team play (this exact order may be wrong, but the general idea is right), and of course discredit to the offense. There are other defensive issues, but in my opinion this is by far the biggest one. Yes, the numbers add up pretty much correctly in the formula 'predicting' a team's wins, but that makes sense; an opponent's possession will end in a steal, a block, a rebound, or points, but the apportioning of credit to individuals is not accurate.

As for their offensive evaluations, they are closer to correct than their defensive ones, but there are some issues there as well, such as no outlet passing credits, setting picks doesn't have any place in the formula, the pass to set up the pass is ignored, etc.

I understand that all these things that I want to include in the formula are extremely hard to measure, but that's partly the point. Just because they are hard to measure doesn't mean that they don't matter, and pretending that they don't matter leads to evaluations like that Garnett, an excellent but flawed player, is by far the best playing today.


  • The idea of introducting increased analysis into the basketball world is a positive. Unfortunately the authors of the book (I'm blanking on the title) have a long way to go. To argue Dennis Rodman was more valuable than Michael Jordan is to admit you are only using numbers with not making an attempt to watch the games. Rodman was a complete offensive liability who also missed multiple games a season. Rodman would often leave his man on D to get in better rebounding position. While scoring is not everything, you do need to score to win. Rodman was a very good player but need scorers around him as the right complement. Some shooters will shoot a high percentage due to getting open looks because someone else forced a double team. Not all shots are the same. You have to watch games to see who can create his own shot and who benefits from the right system. I look forward to better analysis in the future. This along with John Hollinger is a start.

    By Blogger Seth, at May 30, 2006, 4:25:00 AM  

  • Don't hate on objective analysis. Your eyes deceive you, just because Jordan looks great playing doesn't mean he's the best.

    By Anonymous Anonymous, at May 30, 2006, 3:25:00 PM  

  • The problem isn't with 'objective analysis', but incorrect analysis. The authors claim that their system works because the numbers add up. Of course they add up! Like I said, an opponents possession will (95%+ of the time) end in a steal, block, miss/rebound, or points, and all of these possibilities are accounted for. However, that doesn't mean that the right people get the credit they deserve. I'd guesstimate that the rebounder deserves about 25-35% of the credit, but these guys give it all to him. Whether or not my guess is good, it is much closer to the truth than 100% is. This isn't the only problem with their work, but it does seem to be the biggest.

    By Blogger Maestro, at May 31, 2006, 12:13:00 PM  

Post a Comment

<< Home