During the NBA Finals, I made an effort to estimate player contributions to the final score, using Model-Estimated value, and a metric which I unimaginatively called “Credit.” MEV is a linear-weighting player productivity measure (read about it here), and Credit (which I’ve modified somewhat since the Finals coverage), attempts to divide credit (or blame) for a team’s success among individual players:
Player Credit = Player MEV / Team Total MEV * (Team Points / (Team Points + Opponent Points))
This way, total Credit for all players on both teams sums to one in every game, and players on teams that win by a lot are allocated more Credit to divide amongst themselves, whereas in tight games, each team has closer to 0.50 Credits to attribute. In the spirit of the upcoming Olympic Games, I hope to return to semi-regular coverage, though not necessarily in the form of any actual posting. Rather, I will endeavor to update, after each game leading up to and played in Beijing, MEV and Credit statistics for each member of Team USA. Each game’s statistics, as well as cumulative stats can be found here:
I hope you find it useful and insightful over the next few weeks.
The Arbitrarian 
Those credit percentages seem about right to me from my observation of the game. Looking forward to seeing these after the upcoming warm-up in China.
Very nice.
I love looking at these each time a game is done. Keep up the good work – one quick typo correction, you have Tayshaun Prince’s name misspelled as “Teyshaun”.
FYI only – The USA basketball site has slightly different stats (http://www.usabasketball.com/seniormen/2008/stats/USA.HTM) for Deron Williams and Kobe Bryant on FG and FGA, not sure why that is. Do they go back and revise them?
I_hate_flowers: Thanks for pointing that out. I’ve updated the stats to reflect the official USA Basketball numbers.
The reason for the discrepancy regarding Kobe and DWill’s stats is that in the Turkey game Kobe scored nine points–which you can verify by consulting the play by play sheet–but was only officially credited with seven points in the boxscore. DWill scored four points but officially he was credited with six points. The official Team USA stats reflect the incorrect info, so Kobe is listed as scoring 67 points instead of 69, while Dwill is credited with 32 points instead of 30.
I realize that this slight glitch will not alter your ratings that much but this is an example–along with the charting I did of Chris Paul’s scorekeeper-inflated assist totals in last year’s playoffs–of how numbers only give an incomplete picture of what happens on the court.
The human eye may miss things but numbers do not capture everything either and sometimes the numbers don’t even capture such basic truths as who actually scored and who delivered the pass that led directly to a basket.