He was discussing it before it became a "very popular stat". How many people knew what adjusted +/- was 3 years ago, when he was referencing it? Hell, how many people know what it is now? This is from 2006: http://www.chron.com/CDA/archives/archive.mpl?id=2006_4099359 [rquoter] Morey began his NBA career analyzing customer relations but gradually moved into basketball operations, serving as an adviser to general manager Danny Ainge. What are some of the statistics he studies? "Efficient use of possessions is an undervalued, under-appreciated thing relative to just a guy who scores," he said. "Is he using those possessions efficiently? That's a key thing that's undervalued." And? "The unit that is what I'd call `ground truth' in the NBA is measured another way. There's a player on the floor with four other players, and he's facing an opposing group of five. While those 10 guys are on the floor, they're playing a mini-game for the time they're on the floor. Who won? "What created them winning and losing? Maybe they created extra possessions through turnovers or rebounding." Maybe it's one or two players being part of the varied lineups throughout an entire game that is more responsible for success than a box score would show. Maybe a player some people see as valuable really isn't. [/rquoter] Again what he is describing there (what he calls, almost reverentially, "ground truth") is what adjusted +/- is trying to capture about players. The public domain numbers at basketballvalue are the results of one model for outputting a top-level summary of how each player may impact their teams during those "mini-games" Morey talks about. Morey has his approach which may not be radically different, as do the Mavs, and the Cavs, and maybe some other teams. The Rockets do have a lot of data at their fingertips, which allows them to track other things other than simply how the score changes with players on the floor. They can try to analyze more deeply why the score changed as it did.
I wasn't referring to rebounding% (i.e. the percentage of rebounds a player grabs while on the floor). The Rockets track something else to understand how players impact their team's rebounding%. A player like Shane Battier may not grab a lot of rebounds, but by analyzing how certain combinations of players rebounding with him on the floor and off the floor over time, the Rockets can get a sense of whether he is hurting or helping in that department. That the Rockets look at things this way is discussed in Chris Ballard's recent book, and Morey and I believe Sam Hinkie have referred to it before in interviews for various articles. This is very much a related topic to adjusted +/-.
I stand corrected. Still, my point wasn't that Morey doesn't find adjusted +/- useful at all but that he probably doesn't place as much value in it as some others. That "why the score changed" is more important, sort of like the key to the safe that is +/-. Morey has those answers, we don't. Without those "keys" we are left with a stat that is similar in value to a locked safe.
Ah, different reb%...but yes, it still quantifies a small piece of information which I feel is something you can draw more valuable conclusions from. I don't have any gripe with any specific type of stat. I dislike the sweeping range types of production that adjusted +/- attempts to describe. Oh, and I usually woundn't say this but I tried to rep you. Even when I disagree, I can learn from your posts. You've made some good points.
I think I understand. Adjusted +/- approach can be broken down into offensive/defensive components. The main thing is that an analyst's job shouldn't be done just with the +/-. It is helpful to have top-level summaries of how a player (or lineup, or player combination) might be performing, but then you dig in deeper to understand why. Its like if you were analyzing the performance of a complicated, instrumented application on a new computer system using some sophisticated performance tool with access to low-level hardware counters. Do you start off by examining a those counters (e.g. L2-cache misses) at every region of the application to see where the bottleneck is? Probably not. It would make more sense to first identify problem regions based on execution time, and then once those areas are isolated you dig in deeper and see why the performance is poor. I look at the problem of analyzing lineup performance in basketball in a similar way. The +/- stuff, which represents point differential, is analogous to execution time measurements for an application. Yes, it may not tell you why the performance is off, but its useful to know because it gives you an idea where to focus your attention. Thanks.
I think chuck has been great this year and its not only by the +/- stuff. The year the rockets lost to the jazz the first time around, the rockets starter had i believe the best +/- of any starting five and lost to the jazz because one game they had 4 guys score. Basketball unlike many sports to me rely on impact players making impact plays. Thats not a knock against chuck, i just think ifg hayes is starting and playing 20 mins on your team, your team isnt going to be very good. He's a situational player and through injury has been forced into this role he's in now. I appreciate a guy that can go beyond what is expected of him.
Define "impact player making impact plays" Is it only a person who makes a shot with 5 seconds left on the clock to win a game... or can it be a player who takes a charge with 5 seconds left on the clock to win a game? The definition of "impact" is based on opinion.
There really aren't a whole lot of players that are consistently impactful in the playoffs. That's what stars are for. You need role players to be solid, and be able to step on various occasion to make an impact. If you have a number of such players, then hopefully you'll get 1 or 2 each game making an impact. As for the Utah series, I'd say Chuck was impactful in 2 of their 3 wins. First game of that series, Chuck had a huge impact in the first half (close to double figure rebounds by half time including a bunch of offensive rebounds, if I recall) which helped the Rockets stay in the game when their shooting being off. 5th game of the series, Chuck made a game winning defensive play on the final possession. I'm not saying Chuck had a good playoffs overall, but he can make impact plays for you. Just not as a scorer.
Not exactly sure I'm following the point/math here. If the Rockets starters had one of the better +/-'s, and better than the Jazz, for starters in the playoffs, then they lost the series because their bench sucked, or the Jazz bench was awesome, or both... This may be what you are saying, but the way I read it, it sounded as if you were discounting the +/- stat because despite the Rockets starters being great by that measure, they still lost the playoff series. If so, it isn't the stat that is misleading, just the analysis of it. I know DM has better stats than adjusted +/-, more detailed and more customized. But as a fan, without access to insider statistics, it has to be one of the more useful stats available. Certainly, sample size is an issue to be aware of, but over 82 games, or 2 or 3 seasons, it would seem to me that an individual plays with enough combinations of players to be able to gauge how that players, as an individual, impacts whether or not his team outscores the other team when he is playing, and certainly to gauge that individuals effectiveness when part of a larger 3,4 or 5 man grouping. As a fan, this is all I care about. Is my team winning? Who, or what combination of players, are on the court that is helping my team win? I'd agree that the Chuck love can be a little much at times. As can the Battier love. While its definitely early, Battier's performance this year has to be a concern.
Nice find. Wayne Winston has his own, proprietary methods of calculating Adjusted +/- for players (he originated the idea, "WINVAL", maybe ten years ago) and its what the Mavs have used. And his numbers are basically in accordance with what we have publicly available, with some differences here and there. This is stuff the Mavs paid hundreds of thousands of dollars for, and Mark Cuban isn't a dummy. It's useful.
