Xmas in Charity Tournament

Bob

Apparently he never missed a shot

beat

BobH,

Well, I looked at the formulas and rationale involved in the stats. Frankly, there seem to be important gaps in the explanation of the methodology, and the best I could do was to guess at how they handled certain things that were not spelled out.

As I under stand the procedure, it's not W/L record but an estimate of the proportion of a team's "wins produced" by a given player that seems to be the goal of the math. In the cited example involving the Bulls, Rose's figure was dwarfed by those of Noah and Ron Brewer (and it wasn't because Rose played fewer games; the stats were compiled on a "per minute played" basis).

On one hand, credit must be given to the statisticians for trying to incorporate virtually every basketball stat known to mankind in calculating these estimates. Even rebounds "lost" to one's own teammates are included. And there are deductions for points lost by a given player's turnovers. Adjustments are even made by the position of a given player.

The statisticians seem to correlate each statistic with the effect it has on a win, which provides a weighting to give that stat in the overall formula. This is also an advance over just throwing a bunch of stats together and tacitly assuming they all should be weighted equally.

The sheer number of stats that are incorporated lends some credence to the ability of the formulas to rank various players in terms of "share of winning." As even more statistics become available as part of the plethora of stats currently being introduced, the formula will very likely gain more credence.

What bothers me is that, when they correlate each stat with its influence on wins, they do it only one stat at a time. If multiple stats were considered simultaneously, the relationship of some stats to other stats would undoubtedly yield different weights for given stats than are derived by considering each stat individually. But, more than that, there are still more offensively oriented than defensively oriented stats, and I have to believe there's a skewness toward offense in the overall ratings. Moreover, I didn't notice any factors in the formula pertaining to the quality of combinations with whom a given player operated.

My overall assessment would be that this kind of formula, calculated as (partially) explained, would be more useful in evaluating the relative values of each of several players (because any lacks in the methodology would apply to all of them) than in evaluating the true "worth" of any single player.[i] So, although I would not necessarily endorse the details of the formula, the fact that so many stats were considered make me happy that Christmas ranked highly.

As more refined statistics become available and are incorporated into the formula, the results will continue to inch upward in their value. But, if I thought basketball would ever reach the point where statistics could outweigh informed judgment in player evaluation, my interest in the game would wane considerably.

Sam

Bob,

You precisely nailed it—both the good and the bad. During some seasons, I've been accused of being overly optimistic about the Celtics and their chances. This year, I just seem to be doing what comes naturally.

Sam