So a two point field goal will yield you one unit of win score, and a three point field goal will yield you 2. Free throws work the same way basically as two point field goals, you are credited for the points you score but deducted a sum equal to half your total ft attempts, the logic being that two free throws have the same cost as one shot.

Lets take for example Kobe’s game against the Rockets on Mar. 30. He scored 52 points. Amazing right? But let’s look at his winscore.

Scoring – He scored 53 points on 44 shots and 14 fts. So divide the ft’s by 2, and you come up with a deduction of 51 possessions employed. So all that scoring netted him just 2 units towards his winscore.

Gaining and maintaining possession – Kobe was exactly even on this metric, which isnt very good either. He had two rebounds, and two assists, which are half as valuable as a point or possession, for a total of three. However, he is deducted three for the cost of his three turnovers.

Crimes and Rejections – Kobe gets deducted 1 for committing two personal fouls, which like assists and blocks are worth half as much as a point, rebound, steal or turnover.

So for the entire game Kobe accumulated just one unit of winscore, and despite scoring 53 points, he was actually well below the average winscore production for the shooting guard position.

It sounds a bit heretical doesnt it, but I believe this logic is quite sound. You CAN score 53 points in a game and yet still hurt your team’s chances of winning the game.

Of course quite often when you when you have a fifty point game you DO help your team. On April 15, Kobe posted a winscore of 28, on 18-25 shooting, 11-13 ft, and a net of 10.5 towards winscore from gaining and maintaining possession. That was his best game of the season and one of the top 15 in the league this season. I am not saying scoring is a bad thing, just inefficient scoring.

Kobe is a great player btw, but strangely most of his value comes not from his scoring, but from his above average ability for his position to gain and maintain possession.

If you want to check out a great website, which has win score stats for every game and every player this year, go to:

http://www.jasonchandler.com/basketball/

Hope that helps…

You do a good job of supporting WOW, so much so that I think I’ll check it out.

One question, I sort of breezed through everyting you wrote so maybe I missed the logic behind it, but:
Why do made FGAs count against you? If everything goes according to plan is seems one player should take a shot for your team everytime down the court. If you happen to be that player, and you make the shot, why are you penalized (I guess not really penalized, but why does it count against you)?

I was an all-league HS basketball player once. I scored 14 points per game on 55% shooting, I never made a shot outside the paint. And I was second in my league in rebounding to a player who is now in the NBA. This bsckground I think predisposed me to the WOW, and also perhaps explains my vast affection for David Lee.

These stats guys arent going anywhere. I think most people will cotton to it eventually. Many baseball fans have started to understand that batting average and RBI’s are less important than OBP and slugging. And I think basketball fans will slowly internalize the truth that high scoring, electric 42% shooters dont really help their teams win games and understand that Curry Fingers is not a great player even if he can fill it up. If they do, we might eventually see a better quality of play out on the court.

That transition – converting team stats to individual ratings – is where I see most of the potential weak points/blind spots, although my free time and math powers are not sufficient to sort them all out. I’ll play with a few numbers and try to write a more substantial post.

I actually find much to admire in Berri’s work. Your previous post reminded me that many writers, and posters, are not even talking the same language. If the Crawford-Collins backcourt breaks through and leads us to a title next year, I’ll have to join them.

http://sonicscentral.com/apbrmetrics/viewtopic.php?t=1232&postdays=0&postorder=asc&highlight=wages+wins&start=60

Second, I clearly didnt understand this as well as I thought I did, He does in fact proceed from team statistics. He finds a point and a possession are worth the same in terms of team wins using regression analysis. Increase the number of points a team scores or allows by 100, holding everything else constant, and the team will win 3.3 more games. This gives a value of .033 wins for a point scored and -.033 for a point allowed. The same logic works for possessions. A possession employed is worth -.034, and a possession acquired is worth +.0344. From there, they either use regression to determine the value of each individual action available on the court, or they simply take the values found for points and possessions and apply them. Either way, a point, a shot attempt, steal, turnover, and rebound are all found to have equal value in terms of team wins.

He then takes that logic and it applies it to individual statistics.

Alright, I think I have it right and now feel absolved from trying to explain this again. It was an interesting experiment, clearly I need to delve back into my old statistics textbook.

On page 109 of the book “Summarizing all our Steps” might help. The entire Shaq or Kobe chapter is really what you need, if you still care. :-)

http://www.amazon.com/gp/reader/0804752877/ref=sib_dp_pt/103-9006677-3336653#

Owen – can you link me to this?

I’m being lazy – I’ll have to re-read the book – but I thought I recalled his starting point as being team statistics. Did he really type in every individual’s box score line for an entire season?

p.s. “This is an approach Oliver worked out, not Berri, and is a cornerstone of basketball statistics.” (Bill James and Pete Palmer did the same thing in the ’80s, in baseball)

p.p.s. I am still suspicious of position adjustments, since there’s such variety in the roles various teams give each position. e.g. NYK vs. Golden State.

It’s not a counting stat, like turnovers per game. Mutombo commits very few turnovers per game because he almost never touches the balls on the offensive end.

When he does actually touch the ball, he has about a one in seven chance of turning it over. But he only touches the ball seven times a game, so it shows that he averages about one turnover per game.

If you look at a player like Kobe or Nash or Wade who touch the ball for about forty plays per game, they would average between five and six turnovers at Mutombo’s rate.