http://dberri.wordpress.com/2007/11/23/the-houston-rockets-fail-to-launch/ Berri, author of Wages of Wins, has his own methods for statistically evaluating players/teams based on his "Wins Produced" metric. He uses some these numbers to analyze the Rockets start.
Is basketball analyst a new class in World of Warcraft? Standard deviation spell! From a numbers basis, it's obvious. All the role players are falling short of last year's performances in the box score. They've missed their share of wide open shots. But they're also looking a bit lost w/ the new coach and system. Hopefully they all pick it up and start feeling more comfortable in the flow of the game and the shots will fall.
I'm highly skeptical of Berri's methods. I'm a big Chuck Hayes fan, but he determines that Chuck's per minute box score stats surpass anyone else on the team (based on "Wins Produced per 48 minutes"). I think the type of stats Chuck puts up are generally underrated, but that's going way overboard. I think he presumes that stats which are good indicators of success at the team level (FG%, rebounding, limiting turnovers, etc.) are equally important at the individual level. That simply isn't the case.
Then if you put a bunch of all stars on the court they should always win? We know that's not the case. The solution IS in 'TEAM'.
Chuck's numbers are good because he readily creates synergies with others, esp the Stars. No offense to Chuck but on a team w/o dominate players like a TMac or Yao, Chuck's numbers could be terrible. That is an element of 'team' that this analysis fails to assign appropriately. After enough games, lenovos are probably more useful.
Yeah I hate the '48 minute' stats. Totally worthless. There's a reason why guys don't play a full 48, and it's not just because of endurance.
Actually, he played roughly 500 minutes his rookie year and rarely with the two stars. Granted, he mostly played against reserves, but his box score stats were actually much higher than they have been the last couple years. Incidentally, his +/- numbers were very high then as well.
I don't have any problem with per minute stats for role players. They are a better indicator of on-court effectiveness than per-game stats. I just think Berri's "Wins Produced" metric is flawed. I think the argument that a player wouldn't be able to play a full 48 is beside the point. Rates are given "per 48 minutes" just because it's easier for our brains to judge their on-court effectiveness that way (we're used to looking at the stats for star players who play big minutes). They just as easily could be "per 40 minutes" or "per 24 minutes" or "per 10 minutes" or "per 1 minute".
