Bucks Diary

Monday, December 15, 2008

Introducing new statistic called "Championship Percentage"


I've been wrestling with how I can better express my basic economic argument against Michael Redd. Its not that I think he's a bad player. Its just that he doesn't justify the investment of resources we've made in him if our ultimate goal is a championship. I think I've come up with a way to express what I mean.

I call it a player's "Championship Percentage". Its based on the following. In order to win a championship, at least in the last 18 non-strike seasons, a team has had to average 58.8 wins. In terms of cumulative Total Win Contribution, that means you must post around +1.250 as a team.

Now, if a team fielded a squad of replacement level players, in other words, scrubs they could readily acquire at low or no cost, that team could expect to win about 21.4 games. The cumulative Total Win Contribution of a replacement level team is therefore -1.500. (And remember, the replacement costs at each position vary, with SG being the lowest, then SF, then PG, then PF, then C being the most expensive.)

To have a legitimate chance to win a championship, a team must therefore accumulate a squad of players whose Total Win Contribution is +1.750 above replacement level. Each player's contribution toward that goal is what I refer to as the player's "Championship Percentage". I would argue that a team should not sink maximum money into any player who has not shown the consistent ability to achieve a Championship Percentage of at least 15%.

This of course places a special burden on swingmen to justify their value, since their two positions, SG and SF, have the lowest replacement costs. Accordingly, I would only sink maximum money into elite swingmen, not merely above average swingmen like Michael Redd. Those contracts tie down too many of a team's valuable dollars, and lead a team down the road to mediocrity.

For example, since signing his maximum dollar contract several seasons ago, Michael Redd's "TWCar" (Total Win Contribution above Replacement) has never exceeded +0.286 and has been as low as +0.126. Thus, his "Championship Percentage" has never exceeded 10.4% and has been as low as 9.8%.

By contrast, swingman Kobe Bryant last season posted a TWCar of +0.752, which represents a Championship Percentage of 27.3%. So he's clearly worth the money. So is small forward LeBron James, who last season posted a TWCar of +0.789 for an even higher Championship Percentage of 28.6%.

Besides telling a team how to invest their money, "Championship Percentage" can also tell them how to construct their roster, and how much of a "supporting cast" their best player requires to get them to a championship. For instance, I estimate that in 1971 Kareem Abdul Jabbar had a TWCar of roughly +1.671 for a Championship Percentage of 60.7%, so he didn't need much. Oscar and Bobby D were plenty. And in many of Wilt Chamberlain's best seasons, he had Championship Percentages of nearly the same (he just had a bunch of mince meat surrounding him) so it was unfair to compare he and Russell solely on the basis of championships.

I'll flesh this statistic out a bit more in future posts. With the Bucks off tomorrow night, I intend to do a complete statistical comparison of Michael Redd and Ray Allen since the infamous Allen-for-Payton trade for Wednesday. It should be interesting.

0 Comments:

Post a Comment

<< Home