*New Adj. +/- Rankings. Best Off/def and overall players

[jinxed]

Azereus,

And of course APM drops when you're playing a better team that's beating you and goes up when you beat a bad team: it's you losing or beating a team that makes you better or worse, right? The NBA isn't like a tournament everyone plays against one set team and then derives their APM based off how they did against that one team. It's adjusted, yes, but ultimately if you play only bad teams and crush them, your APM will be higher than if you played amazing teams and got repeatedly destroyed.

Nope, that's not how it works at all. That's pure +-. Adjusted plus-minus takes into all the other players on the court with you..on your team..and against you and then evens it out. It's basically a stat that says if Mike Bibby was on the court with 4 average players and the opposing team consisted of five average players, how much would Mike Bibby's team win or lose by over 100 possessions? (sort of, don't take that too literally)

**Edit,

Actually a better way to say it might be , 'how much does having Mike Bibby vs. The average NBA player X effect the Hawks performance over 100 possessions That is what adj. plus minus is a measure of.

the quality of your opponents is inversely proportional to your own quality as a player/team. The Nets would have insane APMs if they were a NCAA team, even though they're the same players.

well yes, but only because their numbers would be adjusted to the average NCAA player versus the average NBA player. Any player on the Nets would be a great NCAA player, so they would deservedly have high APM if they were in the NCAA's. But that's not what we are talking about. We are talking about how good the Nets players are in the NBA, and how much they contribute to the final score of an NBA game.

You're missing out on this whole ADJUSTED part of of ADJUSTED plus/minus.

[jinxed]

As for Mike Bibby

Better stats in bold and larger (just bold is kinda hard to see)

07-08: 14.9 PER, .526 TS, .498 eFG, 0.5 ORB, 32.9 AST%, 15.5 TOV%, 2.6 AST/TO, 21.5 USG%, 14.1 PPG, 1.737 OAPM

08-09: 16.3 PER, .544 TS, .518 eFG, 1.6 ORB, 24.2 AST%, 10.6 TOV%, 3.125 AST/TO, 20.6 USG%, 14.9 PPG, 1.079 OAPM

How much of the variance in offensive performance do you think all those stats combined explain? I once used points, offensive, assists, and FG%...a little different then those slightly better stats but in total it only explained about 50% of the variance in offensive performance, most of it due to FG% (this was at the team level by the way)

Anyway, what that means is that about half of the different variables that go into offensive performance are not accounted for by those stats. Such as setting screens, running plays, floor spacing, breaking the press or double teams, the hockey assist, working the 24 second clock, feeding the post player, dribbling the ball to the proper part of the floor that allows teammates to get into the right position, etc etc..all of those other variables that ARE accounted for by plus/minus , but NOT accounted for by PER. It's within those variables that bibby's performance could have dropped from one year to the next.

[jinxed]

To further illustrate and better explain my point. When we take this definition of Adj. Plus Minus

"how much does having Mike Bibby vs. The average NBA player X effect the Hawks performance over 100 possessions"

We can see how it doesn't matter if the Hawks lose by 20 or win by 20. If you replaced Mike Bibby with average NBA player X, what effect would that have had one the outcome of the game? Would the Hawks have won by 18 or 23? If the answer is 18..then Mike Bibby's APM is -2. If they would have won by 23 then Mike Bibby's APM is +3.

The same goes if they would have lost by 23 with player X instead of Bibby then Bibby's APM would be +3, even though his team lost by 20.

Did I explain that well?

[azuresou1]

I'm under the impression that APM only adjusts for quality of opposition and teammates and not minutes? Could you point out where it says otherwise?

Iliadi is using per 40 minutes numbers. Just read his article about that on 82games.com.

Read it, and the interesting thing is that even he addresses several weaknesses that you and jinxed have either ignored or dismissed, such as the player interaction effect, system (which he lists as coaching effects), etc.

Also, this doesn't address my question about how Bibby posts better stats almost universally other than AST% which he counters by lowering his TOV% even more (thus posting a better AST/TO ratio), the Hawks win more games, and the Hawks post more impressive margins of victory, and yet his Offensive APM drops 0.7 points.

And of course APM drops when you're playing a better team that's beating you and goes up when you beat a bad team: it's you losing or beating a team that makes you better or worse, right? The NBA isn't like a tournament everyone plays against one set team and then derives their APM based off how they did against that one team. It's adjusted, yes, but ultimately if you play only bad teams and crush them, your APM will be higher than if you played amazing teams and got repeatedly destroyed.

Well, that explains why Wade lead the league in adjusted +/- last season ... wait ..., anyway, I doubt that you have any idea what you are talking about. Please, read those links before you answer again.

So if the Nets went to the NCAA and the proceeded the demolish every team by 30-40 points... their adjusted +/- wouldn't rise?

+/-, adjusted or not, is a zero sum game.

[azuresou1]

To further illustrate and better explain my point. When we take this definition of Adj. Plus Minus

"how much does having Mike Bibby vs. The average NBA player X effect the Hawks performance over 100 possessions"

We can see how it doesn't matter if the Hawks lose by 20 or win by 20. If you replaced Mike Bibby with average NBA player X, what effect would that have had one the outcome of the game? Would the Hawks have won by 18 or 23? If the answer is 18..then Mike Bibby's APM is -2. If they would have won by 23 then Mike Bibby's APM is +3.

The same goes if they would have lost by 23 with player X instead of Bibby then Bibby's APM would be +3, even though his team lost by 20.

Did I explain that well?

I'm going to just quote this post instead of all the others, since it's less cumbersome.

So basically you are saying that the average NBA player improved more than Mike Bibby did that year, resulting in his drop in OAPM, correct? Since the Hawks won more games and had a higher margin of victory last year? What counts as the average NBA player? Is it the average of all players in the league, all starters, all guards, all point guards, all starting point guards? What's the criteria used to determine it? Because if it's the comparison is against the universal pool of NBA players, that completely disregards team composition and role, and assumes that simply by slotting in someone with a better OAPM, that the team will score more points.

If the Magic replace Dwight Howard with Chuck Hayes, do they become a better defensive team? Possibly, but I'd bet heavily against it, because the Magic's defensive philosophy is to funnel to Dwight, and 6'6 Chuck Hayes may play amazing post defense, but that's not going to help him when Josh Smith can probably jump right over him after driving past Shard (exaggeration, obviously).

Basketball is a team game; there is too much clutter and hidden variables to run a regression and call it a day.

[mysticbb]

Read it, and the interesting thing is that even he addresses several weaknesses that you and jinxed have either ignored or dismissed, such as the player interaction effect, system (which he lists as coaching effects), etc.

I didn't dismiss them per se. You have to take those weaknesses or flaws in your mind once you start to interpret those stats. Well, let me explain it with your words:

Basketball is a team game; there is too much clutter and hidden variables to run a regression and call it a day.

The list can't be interpreted as showing the "best players" in the game. You can name the list most impact on their teams (or for other stats most productive or something like this), but you can't even say the best player on a team is per se the guy with the highest adj. +/-. "Best" refers to something which you can't sum up in one single number, but the single number can be helpful to determine which player is the best.

[azuresou1]

Well then jinxes' claim on the first page is false then.

I don't believe in judging any player with a single number - to me it's seeing the forest for a grove inside.

[jinxed]

So basically you are saying that the average NBA player improved more than Mike Bibby did that year, resulting in his drop in OAPM, correct? Since the Hawks won more games and had a higher margin of victory last year? What counts as the average NBA player? Is it the average of all players in the league, all starters, all guards, all point guards, all starting point guards? What's the criteria used to determine it? Because if it's the comparison is against the universal pool of NBA players, that completely disregards team composition and role, and assumes that simply by slotting in someone with a better OAPM, that the team will score more points.

If the Magic replace Dwight Howard with Chuck Hayes, do they become a better defensive team? Possibly, but I'd bet heavily against it, because the Magic's defensive philosophy is to funnel to Dwight, and 6'6 Chuck Hayes may play amazing post defense, but that's not going to help him when Josh Smith can probably jump right over him after driving past Shard (exaggeration, obviously).

Basketball is a team game; there is too much clutter and hidden variables to run a regression and call it a day.

APM is a measure of PERFORMANCE, not who is the 'best' player..whatever that means? Yes it is context dependent, no one is saying that APM predicts what Chuck Hayes would do if was on the Magic. You are right about that sir. What APM does say is this is how well D.Howard performed, within his teams system, in the role that he was given.

I'm not saying the average NBA player improved more than Bibby, or if Bibby got worse, APM can't say that alone.

[rsavaj]

Am I reading the data correctly? Amare's defensive stats>his offensive stats?!

[Jimmy76]

Am I reading the data correctly? Amare's defensive stats>his offensive stats?!

that puzzled me as well

I can see why he could maybe hurt the offense if you gave him too many iso's and just ended up slowing down and stagnating the game

its hard to believe hes a good defensive player however

[slick_watts]

This may not be the thread for this, but has any proponents of adjusted +/- seen Kevin Durant's numbers this year? Wayne Winston (the guy who said he'd trade KD based on his career adjusted +/-) has him ranked 2nd in the league right now.

What does this say about adjusted +/- as a tool? Is Kevin Durant himself solely responsible for the better ranking or can most of it be attributed to 'noise', the play of his teammates, etc.?

[Jimmy76]

This may not be the thread for this, but has any proponents of adjusted +/- seen Kevin Durant's numbers this year? Wayne Winston (the guy who said he'd trade KD based on his career adjusted +/-) has him ranked 2nd in the league right now.

What does this say about adjusted +/- as a tool? Is Kevin Durant himself solely responsible for the better ranking or can most of it be attributed to 'noise', the play of his teammates, etc.?

The reason for Durants low +/- were some very coachable quickly changed problems on defense (mainly) and Durant being the coachable guy he is fixed it. It was only a matter time before he flipped to exact opposite side of the +/- scale.

[slick_watts]

This may not be the thread for this, but has any proponents of adjusted +/- seen Kevin Durant's numbers this year? Wayne Winston (the guy who said he'd trade KD based on his career adjusted +/-) has him ranked 2nd in the league right now.

What does this say about adjusted +/- as a tool? Is Kevin Durant himself solely responsible for the better ranking or can most of it be attributed to 'noise', the play of his teammates, etc.?

The reason for Durants low +/- were some very coachable quickly changed problems on defense (mainly) and Durant being the coachable guy he is fixed it. It was only a matter time before he flipped to exact opposite side of the +/- scale.

Is there any data available that suggests this or is it just your own observation? Kevin Durant's Net defense points per 100 possessions last season was +8.2. This season it's -10.5. That's a swing of 18.7 points. Kevin Durant is definitely improved defensively, but is he affecting the game that much? It seems to me the +/- numbers are largely affected by the different approach the team has had to defense, not just Kevin Durant specifically.

[Jimmy76]

This may not be the thread for this, but has any proponents of adjusted +/- seen Kevin Durant's numbers this year? Wayne Winston (the guy who said he'd trade KD based on his career adjusted +/-) has him ranked 2nd in the league right now.

What does this say about adjusted +/- as a tool? Is Kevin Durant himself solely responsible for the better ranking or can most of it be attributed to 'noise', the play of his teammates, etc.?

The reason for Durants low +/- were some very coachable quickly changed problems on defense (mainly) and Durant being the coachable guy he is fixed it. It was only a matter time before he flipped to exact opposite side of the +/- scale.

Is there any data available that suggests this or is it just your own observation? Kevin Durant's Net defense points per 100 possessions last season was +8.2. This season it's -10.5. That's a swing of 18.7 points. Kevin Durant is definitely improved defensively, but is he affecting the game that much? It seems to me the +/- numbers are largely affected by the different approach the team has had to defense, not just Kevin Durant specifically.

Personal observation but any interpretation of +/- is going to be. Im guessing its a mixture of noise and improvement but Durant has definitely passed the eye test of improving his defense and the stats reflecting that isnt surprising. Im pretty sure Garnett went from a huge negative to the best +/- player ever like overnight (I remember reading that anyways).

[slick_watts]

If that's the case, how reliable can adjusted +/- be if there are night and day swings? I thought the whole point of 'adjusting' was to isolate one player's impact on the game but it doesn't seem to do that effectively at all on anything but an enormous sample size.

[Jimmy76]

If that's the case, how reliable can adjusted +/- be if there are night and day swings? I thought the whole point of 'adjusting' was to isolate one player's impact on the game but it doesn't seem to do that effectively at all on anything but an enormous sample size.

Players can make oen or two improvements that will radically help their game. Players do sometimes improve that much overnight. Just look at Josh Smith all he said was "ill stop taking 3's and work harder on defense" and became a much much better player. Night and day swings are because of night and day swings in players.

[raleigh]

If you must know, here are the stats I primarily look at when judging a player: TS%, PPG, TRB%, AST%, STL/BLK%, TOV%, Opponent Counterpart 48-Minute PER, and MPG.

Very similar to the concoction I use. I've essentially given up on statistical measures for overall defensive effect, although I've found DRtg and ORtg (Oliver?) to be a bit more promising...I'm currently in love with the "Advanced" stats on basketball-reference.com.

Adjusted +/- seems more useful in evaluating a GM or Coach, and would be difficult to translate across teams.