The NBA Geek is a website devoted to smart, data-driven analysis of the teams and players of the NBA. It is very much a work-in-progress. Most of the statistical analysis used by the NBA Geek is inspired by the Wages of Wins blog of Professor David Berri as well as the two books The Wages of Wins by Berri, Schmidt, and Brook, and Stumbling on Wins by Berri and Schmidt. Both the books and the blogs emphasize quite a few points about performance in basketball:
- Most pundits think that points per game is the most important metric, and the NBA labor market seems to reflect this, as points per game is highly correlated with salaries. However, points per games is not significantly correlated with wins. Despite this, even when scoring totals come at the expense of scoring efficiency, players who shoot poorly-but-frequently are often paid high salaries, and considered by both casual and hardcore fans as well as NBA analysts to be "productive" players.
- The NBA labor market and most NBA analysts undervalue rebounds. Although rebounds do correlate with salary in the NBA, the effect is not strong. (But don't get me started on the argument that we overvalue rebounds. Rebounds have a significant effect on Wins Produced numbers, but it turns out shooting efficiency is even more important.
- The NBA labor market appears to be indifferent to turnovers, field goal efficiency, and steals, even though both of these have strong correlations with #winning (I couldn't quite resist hashtagging that).
- In other words, as just about every coach from Junior Varsity up to the NBA level will tell you, winning basketball games tends to be heavily influenced by such factors as taking high-percentage shots and getting control of (and keeping control of) the ball. Yet most analysts, and indeed most NBA General Managers, tend to ignore most of these factors when evaluating talent, and concentrate solely on scoring totals.
There's a lot more information in the FAQ, but I strongly encourage you to read the books.