Wins Produced is a metric developed by David Berri, Martin Schmidt, and Stacey Brook to measure how any given player's individual contributions relate to winning basketball games. It is documented quite thoroughly in the book The Wages of Wins and further on The Wages of Wins blog, The Wages of Wins Journal, (as well as the many blogs in the Wages of Wins Network such as and Arturo's Silly Stats and others). You can see the formula for calculating wins produced here. Any attempt to summarize the book in just a paragraph won't do it justice (seriously, read the book!) but the primary factors that influence wins produced are, in order of importance: Shooting Efficiency (both from the field and from the free throw line), factors that involve gaining and keeping possession of the ball (rebounds, steals, and turnovers), and team factors (assists, blocks, and personal fouls).
Unlike other all-in-one metrics such as John Hollinger's PER and the NBA's Player Efficiency, which reward players for all shot attempts, Wins Produced treats missed field goal attempts as a negative factor. Here's an explanatory excerpt of this latter effect from the Wages of Wins FAQ:
NBA Efficiency = PTS + ORB + DRB + STL + BLK + AST – TO – All Missed Shots
Game Score = PTS + 0.4 * FGM – 0.7 * FGA – 0.4*(FTA – FTM) + 0.7 * ORB + 0.3 * DRB + STL + 0.7 * AST + 0.7 * BLK – 0.4 * PF – TO
These measures all align because each tells a similar story about player scoring. For example, imagine a player who takes twelve shots from two-point range. If he makes four shots, his NBA Efficiency will rise by eight. The eight misses, though, will cause his value to decline by eight. So a player breaks-even with respect to NBA Efficiency by converting on 33% of his shots from two-point range. From three-point range, a player only needs to makes 25% of his shots to break-even.
Most NBA players can exceed these thresholds. Therefore, the more shots most NBA players take the higher will be his NBA Efficiency total. As a consequence, players who take a large number of shots tend to dominate the player rankings produced by this measure.
The above is one of the reasons that we are so adamant that one must look beyond scoring totals to measure a player's offensive efficiency.
WP48 stands for Wins Produced Per 48 minutes, and is a positionally-adjusted measure of a player's per-48-minutes box-score statistics. It takes into account the team's per-48-minutes box score statistics and the average performance of other players at the same position (some guesswork is always involved because power forwards sometimes play center, shooting guards sometimes play small forward, etc). A full explanation of all the math can be found here. There is also a lot of helpful stuff at "The Basics" section of Arturo's Silly Stats.
The average WP48 value is 0.100, since a team of 5 such players would produce .5 wins per 48 minutes, which would equate to winning 41 of 82 games. In most sites that use the WP48 metric, we divide players into these categories:
It's important to note that yes, players can have a negative WP48, meaning that such a player is an albatross that costs his team(s) wins. No, this does not mean that a team would be better off playing 4-on-5, because the effect of not fielding a player on a team's WP48 would quite obviously be negative.
RAW PRODUCTION is a player's Production, adjusted for defensive rebounds, assists, and team defense, but before adjusting for position. To illustrate why this is important, consider that someone who grabs 8 rebounds per 48 minutes is an exceptional rebounder if he plays the point guard or shooting guard position, but a relatively poor rebounder if he is the team's center. Similarly, a point guard will have higher turnovers per 48 minutes than a power forward because the power forward is not asked to bring the ball up the floor while some other guy is constantly poking at the ball trying to take it away. I'm sure you can think of other reasons that adjusting for position is important.
Sigh. This argument comes up a lot. So here's the thing: most people undervalue (or just completely ignore) rebounds. our metrics do not. Rebounds matter, so players that rebound better get higher scores than players who don't (all other things being equal). This leads those of us who subscribe to this metric to make wild, outlandish claims like, "Marcus Camby is really good at basketball." Which in turn leads to the ever-popular straw man argument, "All you care about is rebounds!" In a word: No. As it turns out, shooting efficiency has a greater effect than rebounding on player performance. A 1% change in points per field goal attempt (or adjusted field goal percentage times two) leads to a 5.2% change in wins produced, for instance. A 1% change in rebounding leads to a 3% change in wins produced (this data was stolen from the Wages of Wins FAQ).
It should also be noted that David Berri recently updated his Wins Produced formula to account for the fact that some defensive rebounds are team rebounds (i.e. we calculate how many rebounds a player "stole" from teammates, and how many were "stolen" from him). This is done in Step Two when calculating production. Turns out, Kevin Love "steals" a lot of defensive rebounds (and Darko Milicic does not, and gets lots "stolen" from him). It also turns out that even after you adjust for this, Kevin Love is still "superstar" good (and, yes, Darko Milicic is still terrible).
Lots has been written about rebounding, why it is important, and why WP48 does not overvalue rebounds. Here are a few points from the Wages of Wins FAQ:
You'll find further reading at Arturo's Silly Stats. Be warned. If you comment about how rebounding is a team stat and that guys like Kevin Love only get lots of them because his teammates don't rebound, I'll probably make fun of you, especially if you haven't read any of the above.
Basically, so that we are comparing apples to apples. Yes, I realize comparing a player with 100 total minutes played to one with 1000 minutes played is silly because the first player's statistics are heavily subject to variance due to small sample size. However, given a large number of minutes, clearly the per-48-minutes stat says more about a player's production than a per-game stat.