Hmm, I don't think my post about why raw +/- is bad is getting to people. +/- is a holistic stat, meaning it looks at the entire system at a given time (5 players on team A, 5 players on team B, refs, coaching decisions). It's not really possible to get individual performance statistics from +/- and thus not good for drawing conclusions about individuals. The state of the system changes dozens of times over the course of a game and coaching tactics changes from game to game depending on which opponent you're facing. The variance in this stat is so high that there will never be a large enough sample size to overcome it. Over the course of an NBA season, there's probably something like 80,000 5v5 permutations across the league over 1300 games. Getting large sample sizes on all these permutations is not possible because there aren't enough games/minutes over a single season. And you can't rely on +/- stats from year to year because players change team which creates more 5v5 permutations. Can raw +/- be useful for any type of conclusions or summaries? The short answer is no. Though lots of people, here and in the media, think otherwise. Again, Jose Calderon had a great game (triple double) against the Roxs recently but has a +/- of zero (http://espn.go.com/nba/boxscore?id=400278066). Looking at the raw +/- numbers, all the starters were positive, he was the worst. John Lucas, the backup PG was +11. From +/- alone, you would think Jose Calderon had a horrible game and the backup played great. Then you look at the pts, assists, rebounds, etc and you see the Calderon was the player of the game. So +/- fails as a summary stat as well. Sure most of the time +/- will correlate to what actually happened in the game. The problem is that it doesn't correlate too often. You always have to have qualifiers when mentioning +/-, so it gets to the point where it's like, why bother quoting it.
Roxs, your argument seems to be that one can't draw statistically significant conclusions from simple +/- numbers. I agree. But I would also point out that stats serve different roles. They may be the result of an analysis, from which one may draw statistically significant conclusions. Or they may be merely descriptive of what happened, while providing little insight into why it happened. A stat that falls in the latter category is not, in my view, useless. You need a way to describe what happened before you can begin an analysis and draw real conclusions. Let's say I'm committing max dollars to a player. I learn that we're getting outscored by 5 points for every 48 minutes he's on the floor, while we outscore the other team by 5 when for every 48 minutes he's not on the floor. Of course that on its own doesn't tell me that this player is overpaid or that he's a bum. Nevertheless, its useful to know, because now I can start to examine why we're seemingly not taking advantage of his abilities. Can you provide an example in this thread of a conclusion/summary that was arrived at, and then we can discuss whether said conclusion/summary is valid or invalid.
the adjusted +/- might be more useful: http://stats-for-the-nba.appspot.com/teams/HOU.html "In "Regularized Adjusted Plus-Minus" (RAPM), the goal is to provide more accurate results by employing a special technique called "ridge regression" (a.k.a. regularization). It significantly reduces standard errors in Adjusted Plus-Minus (APM)." With sample-size and injury/flu caveats. Should be more accurate towards the end of the season. More explanation on RAPM: http://godismyjudgeok.com/DStats/2011/nba-stats/a-review-of-adjusted-plusminus-and-stabilization/ "It has long been recognized that raw Plus/Minus is not particularly indicative of each player’s actual performance or value. The largest issue is that the stat is totally context-dependent: if I was on the court with Lebron James, Dwyane Wade, and Chris Bosh, I’d probably have a pretty good Plus/Minus despite the fact I shouldn’t be playing at all. And even if I was a really good player, I still wouldn’t have good raw Plus/Minus numbers if I were playing with Jamario Moon, Anthony Parker, Ramon Sessions, and J.J. Hickson. In addition, the opponents matter as well. Some players just play in garbage time against the other team’s scrubs. Their Plus/Minus numbers may be pretty good, but only because they were playing against Brian Scalabrine! This big issue, the dependence of Plus/Minus on the other 9 players on the court, gave rise to a new statistic aimed at adjusting for this problem: Adjusted Plus/Minus."
Adjusted +/- has its uses. But its the result of analysis, rather than a direct description of what happened. For small sample sizes where adjusted +/- results are too noisy to be helpful, I think I would rather simply know what happened directly without necessarily drawing a firm conclusion from it.
