538 Article: "The Hidden Value of the NBA Steal"

[Doctor MJ]

It's great to see 538 relaunched. I've always been a big fan.

Gotta say though, this article on steals leaves me frustrated, and wondering what you think:

http://fivethirtyeight.com/features/the ... nba-steal/

The thesis of the article is that because of steals having high regression value, a player's SPG is an incredibly important stat. To my mind though there can be no responsible analysis of this without also dealing with the fact that actual regression of overall player impact has shown again and again that ballthieves just aren't very important compared to other defenders and hence the type of indvidual stat regression being done here ends up producing PER-like stats that clearly overrate ballthieves relative to what overall regression analysis (RAPM) does.

I really don't understand why this seems to be so hard for someone like the author (the Skeptical Sports guy) to spot. And once spotted, it's not even like it's a mystery as to what the issue is: There's high variance in a players steals each game, and when the steals aren't paying off they represented costly gambles.

Anyone disagree?

[bondom34]

It's great to see 538 relaunched. I've always been a big fan.

Gotta say though, this article on steals leaves me frustrated, and wondering what you think:

http://fivethirtyeight.com/features/the ... nba-steal/

The thesis of the article is that because of steals having high regression value, a player's SPG is an incredibly important stat. To my mind though there can be no responsible analysis of this without also dealing with the fact that actual regression of overall player impact has shown again and again that ballthieves just aren't very important compared to other defenders and hence the type of indvidual stat regression being done here ends up producing PER-like stats that clearly overrate ballthieves relative to what overall regression analysis (RAPM) does.

I really don't understand why this seems to be so hard for someone like the author (the Skeptical Sports guy) to spot. And once spotted, it's not even like it's a mystery as to what the issue is: There's high variance in a players steals each game, and when the steals aren't paying off they represented costly gambles.

Anyone disagree?

TBH Doc, I saw the title, thought it was BS, and didn't even read it earlier today. Its a stat that doesn't have a ton of value without knowing the amount of times a player gambles and misses steals. A guy like Corey Brewer on Minnesota gambles a ton, and it doesn't pay off, but he still gets his steals, and it's not correlating with wins there obviously. As you said, the variance in steals in a game is greatly going to affect winning, I honestly don't think I'll be reading much 538 as none of their articles to date have been very good in my opinion.

[bondom34]

So, ended up reading it out of boredom, and it was pretty much what has been said earlier. It makes no real sense. To go back to the previous example, not to harp on the guy, but Brewer is 2nd on Minnesota in steals, and slightly ahead in blocks (though way behind in rebounding). Using the replaceability, the assists should be replaceable, but blocks/steals not so much, and though Brewer is an okay player, I doubt that his value would be as high as this would make it seem. Just pulling up some per game stats this year for guys, this would massively overrate Thad Young (2.2 steals, 0.4 blocks, 18 points), Monta Ellis (1.8/0.3/19), and Jimmy Butler (2.0/0.6/13). Oh, and Tony Parker apparently sucks.

[Narf]

It's great to see 538 relaunched. I've always been a big fan.

Gotta say though, this article on steals leaves me frustrated, and wondering what you think:

http://fivethirtyeight.com/features/the ... nba-steal/

The thesis of the article is that because of steals having high regression value, a player's SPG is an incredibly important stat. To my mind though there can be no responsible analysis of this without also dealing with the fact that actual regression of overall player impact has shown again and again that ballthieves just aren't very important compared to other defenders and hence the type of indvidual stat regression being done here ends up producing PER-like stats that clearly overrate ballthieves relative to what overall regression analysis (RAPM) does.

I really don't understand why this seems to be so hard for someone like the author (the Skeptical Sports guy) to spot. And once spotted, it's not even like it's a mystery as to what the issue is: There's high variance in a players steals each game, and when the steals aren't paying off they represented costly gambles.

Anyone disagree?

TBH Doc, I saw the title, thought it was BS, and didn't even read it earlier today. Its a stat that doesn't have a ton of value without knowing the amount of times a player gambles and misses steals. A guy like Corey Brewer on Minnesota gambles a ton, and it doesn't pay off, but he still gets his steals, and it's not correlating with wins there obviously. As you said, the variance in steals in a game is greatly going to affect winning, I honestly don't think I'll be reading much 538 as none of their articles to date have been very good in my opinion.

RAPM is one of the worst stats I've seen. Or to put it another way, there are a lot of guys RAPM says are great defenders who, when put on the basketball court, makes the team defense significantly worse.
It calls certain players elite who clearly aren't elite defenders (Lopez for instance).
There are plenty of players they get right, but that doesn't make it accurate or even close to it.

As for steals, I tend to think they are very underrated. You suggested "Minnesota wasn't winning" with their steals. That's not true. Minnesota is giving up the highest FG% in the NBA and a .500 team with above average defensive efficiency.
http://espn.go.com/nba/hollinger/teamst ... rder/false
That was 12th earlier, but Minnesota's players have clearly given up now and dropped to the 14th ranked defense.
Anyway, I just thought I'd put that out there. DVORP > DRAPM and both are so flawed I don't think either is worth much without context.

Just like steals without context are usually right, but not worth much without context.

All that said, I agree his final analysis was crap. But possessions/FTAs/FG% are all quite important and people tend to focus only on FG%. I might suggest of the 3 that possessions are the most important (steals/turnovers/rebounds).

[bondom34]

It's great to see 538 relaunched. I've always been a big fan.

Gotta say though, this article on steals leaves me frustrated, and wondering what you think:

http://fivethirtyeight.com/features/the ... nba-steal/

The thesis of the article is that because of steals having high regression value, a player's SPG is an incredibly important stat. To my mind though there can be no responsible analysis of this without also dealing with the fact that actual regression of overall player impact has shown again and again that ballthieves just aren't very important compared to other defenders and hence the type of indvidual stat regression being done here ends up producing PER-like stats that clearly overrate ballthieves relative to what overall regression analysis (RAPM) does.

I really don't understand why this seems to be so hard for someone like the author (the Skeptical Sports guy) to spot. And once spotted, it's not even like it's a mystery as to what the issue is: There's high variance in a players steals each game, and when the steals aren't paying off they represented costly gambles.

Anyone disagree?

TBH Doc, I saw the title, thought it was BS, and didn't even read it earlier today. Its a stat that doesn't have a ton of value without knowing the amount of times a player gambles and misses steals. A guy like Corey Brewer on Minnesota gambles a ton, and it doesn't pay off, but he still gets his steals, and it's not correlating with wins there obviously. As you said, the variance in steals in a game is greatly going to affect winning, I honestly don't think I'll be reading much 538 as none of their articles to date have been very good in my opinion.

RAPM is one of the worst stats I've seen. Or to put it another way, there are a lot of guys RAPM says are great defenders who, when put on the basketball court, makes the team defense significantly worse.
It calls certain players elite who clearly aren't elite defenders (Lopez for instance).
There are plenty of players they get right, but that doesn't make it accurate or even close to it.

As for steals, I tend to think they are very underrated. You suggested "Minnesota wasn't winning" with their steals. That's not true. Minnesota is giving up the highest FG% in the NBA and a .500 team with above average defensive efficiency.
http://espn.go.com/nba/hollinger/teamst ... rder/false
That was 12th earlier, but Minnesota's players have clearly given up now and dropped to the 14th ranked defense.
Anyway, I just thought I'd put that out there. DVORP > DRAPM and both are so flawed I don't think either is worth much without context.

Just like steals without context are usually right, but not worth much without context.

All that said, I agree his final analysis was crap. But possessions/FTAs/FG% are all quite important and people tend to focus only on FG%. I might suggest of the 3 that possessions are the most important (steals/turnovers/rebounds).

I'm not saying steals are a bad thing by any means. What I was trying to get at was the conclusion made no sense. Steals are good, but they don't tell me much about a defender, and they sure aren't worth 9 points a pop as claimed. As well, as Doc said guys w a bunch of steals aren't generally important to great defenses, as a big man is (now I realize Min isn't a great defense, but that part stands true). The article claims to show something that just isn't there.

[5DOM]

I'd like to see the regression model because from what I understand he's just used difference in SRS's and the 6 box score stats for his linear regression, and I can't imagine it's a very good model.

This whole article just seems like a lot of jumping to (possibly wrong) conclusions to me because it's so lacking in details and considerations, which is funny because I thought Silver's very against that. I suppose it could very well be true that steals are hard to replace, but I am still unsure of its value in evaluating players because of what you guys have already mentioned.

538 has so far been a disappointment for me as well. It's a good looking website though.

[bondom34]

Also something I read from someone on APBR, but if steals are this important, why wouldn't they correlate much higher on a team level with wins?

[bondom34]

Figured I'd post this here, was browsing a Rockets Truehoop blog post, and this article came up as the topic. Author did more regression analysis, and the steals thing showed to basically be worth the most points. However, steals are so much more rare than other occurrences, he re-ran it using standard deviations as a unit instead of the straight numbers, and in the end concluded rebounding technically is more "worthwhile" statistically.
http://www.red94.net/treatise-data-stea ... more-14228

[Doctor MJ]

RAPM is one of the worst stats I've seen.

Sorry to be inflammatory, but you should know whenever I read "X is one of the worst stats", that to me makes me prone to interpret all that follows as "I don't understand how to use stats but I'm really sure there's something very wrong with this".

Or to put it another way, there are a lot of guys RAPM says are great defenders who, when put on the basketball court, makes the team defense significantly worse.

The noise in +/- data is a real issue. This is not sufficient to call it a bad stat, but it is an issue.

It calls certain players elite who clearly aren't elite defenders (Lopez for instance).
There are plenty of players they get right, but that doesn't make it accurate or even close to it.

Lopez? I'm not sure where you got the idea that RAPM typically loves either Lopez brother on defense. Maybe give some more details and some links. I'm not going to say it's accurate enough you should simply accept the values there as stand in for an overall defensive ranking, but I would certainly question what fundamentally "better" stats are out there.

As for steals, I tend to think they are very underrated.
...
Just like steals without context are usually right, but not worth much without context.

They are like steals to you, yet you think they are a horrible stat while steals are "underrated and "usually right". Like most who dismiss an "advanced" stat, it seems you don't apply the same standards to more established stats. It would never occur to you to ignore steals simply because you need context too, so neither should you behave that way toward RAPM.

DVORP > DRAPM and both are so flawed I don't think either is worth much without context.

Are you referring to this:
http://www.buildingbetterball.com/2013/ ... 2-defense/

Which rates Drummond from last year as the best defender in the league.

If so I'd say that DVORP is both more flawed in the ways you complain about DRAPM being flawed, and far less worthwhile fundamentally because of how it's created. Thing to understand about RAPM is that it's are best use of +/- stats, and that +/- stats provide an orthogonal tool to use to augment box score stats. There was a gap in analysis before +/- stats that simply has to be filled in some way by any analyst worth his salt, and RAPM is what we use to fill that right now. It is not the end all be all, but it's part of something we need to use.

While one could use DVORP as one part of his entire analysis as well, it's the comparison between the two more than anything else that sends the red flags. There's simply no replacement for +/- stats. It's a window you have to use. If you think you have a box score oriented metric that does it "better", then you don't understand how to build an all around approach to basketball analysis.

Beyond all that though there's the matter that you're responding here based on your dislike of RAPM without showing any indication that you understanding of why RAPM damns the 538 analysis - and it does damn that analysis, so you not understanding that means there's just so much you don't understand.

Bottom line is that RAPM is simply "overall regression analysis". The 538 approach is "individual factor regression analysis", and hence it's based on the assumption that regression analysis is a good approach to things. If you reject RAPM based on this, then you need to reject the 538 approach as well, and since the point of this topic is not to worship RAPM but to reject the 538 approach, that means you've gone into a topic that you should essentially agree with and picked a fight over some other issue for reasons that don't factor into the topic at all.

And to be clear, here's the issue with 538: Doing individual factor regression as part of overall player analysis makes the assumption that you could essentially some up the individual factors and get an overall picture. But RAPM is the overall picture, and it violently disagrees with individual factor analysis when it comes to steals. The lack of correlation in and of itself destroys the argument, but we don't have to rely on statistical theory to simply state this if we simply thinks things through a little bit:

Regression analysis on steals as a factor implies an independence of steals from other things that paint the broader picture that aren't covered among the rest of the box score factors. But that obviously isn't the case because we don't have an "attempted steals" factor (among other things) at this time. Recognizing that, the question isn't whether individual factor regression analysis is problematic but simply if that problem is a big one. At that point we'd turn back to overall regression analysis to see if they seem to actually "add up" similarly. And of course, they don't.

[NinjaSheppard]

I found Harlabos Voulgaris' response interesting

Haralabos Voulgaris ‏@haralabob Mar 25

@MikePradaSBN @MattMcLDC @ZachLowe_NBA Could be biased, but this might be the worst article I have every read.

Haralabos Voulgaris ‏@haralabob Mar 25

Well gotta give @fivethirtyeight credit - They’ve ran an article that claims a steal is “worth” as much as 9ppg http://fivethirtyeight.com/features/the ... nba-steal/ …

Haralabos Voulgaris ‏@haralabob Mar 25

Up next an article claiming a 3 point make is actually worth 15 points per game. Those who can’t… write?

Haralabos Voulgaris ‏@haralabob Mar 25

My point is, its silly to create a “data driven” website and then write articles using regressions on box score stats (which suck). The End

Haralabos Voulgaris ‏@haralabob Mar 25

@GuyITC I’m saying 1. they aren’t worth anywhere near 9 points of “player value” 2. Not all steals are equal, most steals are useless.

Haralabos Voulgaris ‏@haralabob Mar 25

@GuyITC @professorsilly Funny you mention that bc rarely are those same guys valued highly by our defensive metrics. (This was in response to ball thieves)

[Narf]

DRAPM is one of the worst stats I've seen.

Sorry to be inflammatory, but you should know whenever I read "X is one of the worst stats", that to me makes me prone to interpret all that follows as "I don't understand how to use stats but I'm really sure there's something very wrong with this".

And when I see that, I think you're arrogant, self righteous, and trying to hide the fact that you can't make a statistically relevant argument to back up anything you say. That's fallacy of course, but I don't feel like arguing stats with someone who disregards anything about me without a civil discussion or anything resembling a benefit of the doubt to my intelligence. So I have no interest in responding after this.

And if you didn't catch it, I clearly said the 538 analysis was "crap". I just disagree with the statements people made about steals not having much value in determining a players defense. I think "gambling for steals" is thrown out there way too much, and the correlation between steals and good defense is stronger than people think.

Anecdotally, I can show a group of players that are greatly mischaracterized by DRAPM by how they play with 4 other players. When a far lesser DRAPM player is replaced with them the defense gets better, meaning the only thing that's changed is a "better" defender replaced a "worse" defender and DRAPM had it backwards. You could do the same for DVORP, but I personally have found DVORP to have a stronger correlation between ACTUAL defense and it's rating. It's not about how the number is created, it's about whether or not that number does what it's supposed to do.

There's also very large fluctuations between players leaving one team and going to another with new players around them. Kevin Garnett was a better defender in Minnesota, but DRAPM says he's been a better defender since then. This is someone I watched his whole career, I don't need to prove that statistically to you. Not that KG wasn't a great defender in Boston, just that he was even better in Minnesota. But his teammates weren't. So because of his teammates, his DRAPM went up.

In essense, this is no different than +/-. But we don't pretend that +/- is without context. You know you have to look at who is replacing them to get a feel for what that defensive +/- is. Actual defense is measured by how many points the team gives up with that player on the floor. Individual contributions are a little harder to verify. But if the defense is terrible with a player on, and average with his replacement on, he's a bad defender any way you cut it. He's also a bad defender if his replacement is a bad defender and they have the same defensive points per possession allowed.

And clearly +/- is not accurate unless put into context. You have to have some understanding of the overall defense, some understanding of the value of his replacement, and some understanding of the players they play most of their minutes with (for instance, Brewer plays most of his minutes with Kevin Martin, which significantly lowers his +/- and DRAPM....making both of them fairly inaccurate). But that's easier to look up with +/-, with DRAPM people just assume it's right. Defensive matrixes don't just have outliers. They are terrible, and simply get it wrong far too often to be used as if it is an authoritative number. It's like counting someone's championships to claim they were a great player. As if Karl Malone wasn't greater then most of the PFs that received rings. Or if KG had spent his entire career in Minnesota and never won one. Great defensive teams make great DRAPM, whether you're a great individual defender or not. The individual player is not greater on that team then he was on the last one, he just had a better team that year.

82games.com keeps track of 5 man units and actual defensive points per 100 possessions while those units are in. Knock yourself out comparing drapm with players and their replacements in 5 man rotations if you'd like. Then (gasp) look at those same numbers the last year and the year before. They won't correlate well.

And no, I didn't get my opinion from some article you found. I had it long before I read anyone else who agreed with it. It wasn't hard to look at DRAPM's ratings and see clear as day they were terrible based on the players I already knew. I think both DRAPM and DVORP are so flawed they are nearly worthless. But of the 2, DRAPM is more flawed. Feel free to disagree.

[Narf]

I wish I hadn't brought up Brewer, as he's a very special case.

So, Brewer is a unique case. His big asset for playing along side of Martin is a flaw when playing along side good defenders. He's a ball hawk, and sometimes loses his man. He gambles, but also disrupts and that has a blanket effect on the other team's offense. When playing along side one of the worst defenders in the league, this becomes an asset. Martin is so bad, that having Brewer half leave his man open to come over and help actually makes the defense better. As long as he gets enough steals and force enough bad plays to make up for the open shots his man gets here and there.

Overall Brewer is at least above average on defense IMHO. He does give up easy baskets, but that isn't so different from giving up not-easy baskets. He also makes amazing defensive plays sometimes. He also wins gambles when going for steals, or forces bad passes/mistakes. He also recovers. And he will always be a polarizing person because he's so good and so bad on the same play that it's hard to put a value on him.

But, the point was right. Because when someone is paired with a terrible defender to help "make up" for their bad defense, it alters their defensive stats significantly.

[Doctor MJ]

I found Harlabos Voulgaris' response interesting

Haralabos Voulgaris ‏@haralabob Mar 25

@MikePradaSBN @MattMcLDC @ZachLowe_NBA Could be biased, but this might be the worst article I have every read.

Haralabos Voulgaris ‏@haralabob Mar 25

Well gotta give @fivethirtyeight credit - They’ve ran an article that claims a steal is “worth” as much as 9ppg http://fivethirtyeight.com/features/the ... nba-steal/ …

Haralabos Voulgaris ‏@haralabob Mar 25

Up next an article claiming a 3 point make is actually worth 15 points per game. Those who can’t… write?

Haralabos Voulgaris ‏@haralabob Mar 25

My point is, its silly to create a “data driven” website and then write articles using regressions on box score stats (which suck). The End

Haralabos Voulgaris ‏@haralabob Mar 25

@GuyITC I’m saying 1. they aren’t worth anywhere near 9 points of “player value” 2. Not all steals are equal, most steals are useless.

Haralabos Voulgaris ‏@haralabob Mar 25

@GuyITC @professorsilly Funny you mention that bc rarely are those same guys valued highly by our defensive metrics. (This was in response to ball thieves)

Wow. Harsh.

It's interesting about the author in question, who is the Skeptical Sports guy who did such great work with Rodman. He's very clearly very smart in math and capable of insightful work. He's got a lot more common sense than a lot of the APBRmetrics crew, and obviously more than Berri.

But once again we see a basketball n00b come in and get championed by other smart data guys for work that we can tear apart like tin foil. What's bizarre to me is that the issues I see don't seem like they require basketball knowledge to see. OBVIOUSLY you sometimes can't use individual factor regression as a more detail oriented approach to overall regression, so if you have access to both (which you typically do) you should crosscheck before you bloviate.

So why don't they?

I feel like the answer is just overconfidence. They come in thinking not that their analysis is perfect but that it surely must be an improvement over what is already in play in the intellectual void of jock-world. That shouldn't matter - they should dot their i's and cross their t's always - but they think they can get away with it. And it seems that the Gladwells and Silvers of the world aren't up to be able to tell that they publishing guys who are becoming laughing stocks among those who are the true contextual experts.

[Doctor MJ]

DRAPM is one of the worst stats I've seen.

Sorry to be inflammatory, but you should know whenever I read "X is one of the worst stats", that to me makes me prone to interpret all that follows as "I don't understand how to use stats but I'm really sure there's something very wrong with this".

And when I see that, I think you're arrogant, self righteous, and trying to hide the fact that you can't make a statistically relevant argument to back up anything you say. That's fallacy of course, but I don't feel like arguing stats with someone who disregards anything about me without a civil discussion or anything resembling a benefit of the doubt to my intelligence. So I have no interest in responding after this.

Yeah I should have tried to find a better way to say it. I'm sorry.

The point is really just that it doesn't make sense to talk about ranking how "good" various stats are when they don't tell us the same story as each other. If you're inclined to throw out +/- stats because you see some data in there you think is unreasonable then you need to change your perspective on how to incorporate various stats into your analyses. Something can only be justified as useless to you if you've can a superior version of the same type of tool handy, which in the case of RAPM you basically can't. You can be cautious about using it, you can criticize others for putting too much faith in it, but if you don't use it then you're ignoring a set of data that there's no rational basis for ignoring.

And if you didn't catch it, I clearly said the 538 analysis was "crap". I just disagree with the statements people made about steals not having much value in determining a players defense. I think "gambling for steals" is thrown out there way too much, and the correlation between steals and good defense is stronger than people think.

Anecdotally, I can show a group of players that are greatly mischaracterized by DRAPM by how they play with 4 other players. When a far lesser DRAPM player is replaced with them the defense gets better, meaning the only thing that's changed is a "better" defender replaced a "worse" defender and DRAPM had it backwards. You could do the same for DVORP, but I personally have found DVORP to have a stronger correlation between ACTUAL defense and it's rating. It's not about how the number is created, it's about whether or not that number does what it's supposed to do.

There's also very large fluctuations between players leaving one team and going to another with new players around them. Kevin Garnett was a better defender in Minnesota, but DRAPM says he's been a better defender since then. This is someone I watched his whole career, I don't need to prove that statistically to you. Not that KG wasn't a great defender in Boston, just that he was even better in Minnesota. But his teammates weren't. So because of his teammates, his DRAPM went up.

Well fundamentally with RAPM we're still just talking about a variation on player "lift" (meaning how is the player in question lifting his team compared to what they'd do without them), and most typically the potential bias here runs in the other direction. If you're on a team with bad players, you can make a night & day difference but joining players already capable of functioning well without you means that there's going to be redundancy and diminishing returns.

What the data tells us about Garnett's defense in the two cities is that Garnett actually had more lift when surrounded by better talent, and that's so rare as to be extremely noteworthy. If you think you see this a lot in RAPM, you're confused.

Speaking more on what happened with Garnett. To the extent you're saying he wasn't actually a more capable defensive player in Boston than he was in Minny, I quite agree. What we're talking about is him having greater defensive efficacy in Boston than in Minny. And that was based on him being allowed to focus more on defense in Boston than he did as the two-way star of Minnesota, as well as getting to play for Thibodeau who is probably the best defensive mind of the basketball world. It's good to note that context, but it's not a flaw in RAPM that it tells us "Wow, Garnett's really having huge impact on defense in Boston!".

In essense, this is no different than +/-. But we don't pretend that +/- is without context. You know you have to look at who is replacing them to get a feel for what that defensive +/- is. Actual defense is measured by how many points the team gives up with that player on the floor. Individual contributions are a little harder to verify. But if the defense is terrible with a player on, and average with his replacement on, he's a bad defender any way you cut it. He's also a bad defender if his replacement is a bad defender and they have the same defensive points per possession allowed.

And clearly +/- is not accurate unless put into context.

Just jumping in here because it begs me to say how much of my ability to use +/- in all forms (raw, on/off, APM, RAPM, xRAPM, etc) is based on my ability to read the context. You're not wrong at all on this, it's just you went down a wrong turn when you chose to look at a particular stat in terms of "good" & "bad" instead of just using the noise to develop your intuition for the context further.

And I'm sorry if I continue to sound condescending, it's just that I've got quite a lot of experience in this and I've seen so many people talk like you're currently doing, and sometimes I don't give enough cushion in my words to keep them from feeling their pride being violently attacked. There's no way around it though, I think I do know better than you on this, and since you didn't come in here making your own content but rather criticizing something I'm using, I get impatient sometimes.

You have to have some understanding of the overall defense, some understanding of the value of his replacement, and some understanding of the players they play most of their minutes with (for instance, Brewer plays most of his minutes with Kevin Martin, which significantly lowers his +/- and DRAPM

Jumping in again just to say: There's absolutely no reason to think that being on the floor with a bad defender makes your RAPM more likely to be worse. The entire reason why we have "adjusted" +/- is combat that issue.

....making both of them fairly inaccurate). But that's easier to look up with +/-, with DRAPM people just assume it's right. Defensive matrixes don't just have outliers. They are terrible, and simply get it wrong far too often to be used as if it is an authoritative number. It's like counting someone's championships to claim they were a great player. As if Karl Malone wasn't greater then most of the PFs that received rings. Or if KG had spent his entire career in Minnesota and never won one.

I agree with you. I figured this is the source of your frustration, but I have to step in when you go from there and literally call a stat bad because others use it improperly, especially when you say that in response to a conversation started by me and I didn't use it like that. When you reach the point that you assume someone is using a stat poorly without any evidence they are, you've got an issue. When you don't simply assume but respond by starting with "RAPM is one of the worst stats I've seen." you're broadcasting to the world your issue.

And so when I respond to you by letting you know how that looks it's tough love man.

Great defensive teams make great DRAPM, whether you're a great individual defender or not. The individual player is not greater on that team then he was on the last one, he just had a better team that year.

82games.com keeps track of 5 man units and actual defensive points per 100 possessions while those units are in. Knock yourself out comparing drapm with players and their replacements in 5 man rotations if you'd like. Then (gasp) look at those same numbers the last year and the year before. They won't correlate well.

And no, I didn't get my opinion from some article you found. I had it long before I read anyone else who agreed with it. It wasn't hard to look at DRAPM's ratings and see clear as day they were terrible based on the players I already knew. I think both DRAPM and DVORP are so flawed they are nearly worthless. But of the 2, DRAPM is more flawed. Feel free to disagree.

I would say the fundamental thing you need to consider here is that even though we talk about pluses and minuses, your basic mathematical terms, we know that all of this is much more complicated than that. It's still over simplistic, but start thinking about the synergy & redundancy of sets of players in terms of multipication, division, exponents, etc.

When Garnett went to Boston, you didn't simply add a defender of X1 quality to a bunch of guys of X2, X3, X4, etc quality, and the clearest sign of that is that no one expected the team to dominate the league with their defense. If you had told people that Garnett would get the box score stats he did in Boston in the summer of 2007, no one would have been impressed, and no one would have talked about the tremendous effect that would mean he was having on the Boston defense.

Hence, it was an absolute shock to the basketball world, and a clear sign that we need stats that help us embrace such nonlinear relationships in the sport. The +/- family of stats helps us do that, and the proof of that is in the fact that it is in those stats that you see some manner of individual explanation for the shocking emergent property that was Boston's defensive dominance.

Now, clearly you'd see one rebuttal to this being that +/- is simply echoing the team success. Okay, and what's the problem there? When you talk about year to year variance in +/- data that goes along with year to year variance in overall team defensive efficacy, obviously something important is changing, and if you've been around basketball enough you know that something may have nothing to do with an actual roster change. That +/- stats can actually see something and say something when problems happen simply makes it useful.

Another rebuttal: Well if it's just an effect of team success, then it doesn't say anything about actual causality. Of course it doesn't, and that's why you need context. +/- data is either a way to confirm hunches or a way to tell you need to look closer at something. It is not the entire analysis. I agree with you that sometimes people use it as such and that's a problem.

But what you also need to understand, particularly in a context like this, is that people have been at this quite a while, and when that's the case they use shorthand. A player's RAPM isn't ever alone enough for me to say something big about him, but 'round these parts analysis of, say, Garnett's defensive strengths and role as a defender in Minny & Boston has been talked about so much that it hardly seems necessarily to re-has all that every time I mention a stat of his where he looks good.

Also ftr, I was asking what stat you were using when you refer to DVORP. VORP is a baseball term that's been claimed by multiple basketball statisticians and none of them have really gotten that much traction because people aren't really sold that they do what they say (calculate "value over replacement player").

[Doctor MJ]

I wish I hadn't brought up Brewer, as he's a very special case.

So, Brewer is a unique case. His big asset for playing along side of Martin is a flaw when playing along side good defenders. He's a ball hawk, and sometimes loses his man. He gambles, but also disrupts and that has a blanket effect on the other team's offense. When playing along side one of the worst defenders in the league, this becomes an asset. Martin is so bad, that having Brewer half leave his man open to come over and help actually makes the defense better. As long as he gets enough steals and force enough bad plays to make up for the open shots his man gets here and there.

Overall Brewer is at least above average on defense IMHO. He does give up easy baskets, but that isn't so different from giving up not-easy baskets. He also makes amazing defensive plays sometimes. He also wins gambles when going for steals, or forces bad passes/mistakes. He also recovers. And he will always be a polarizing person because he's so good and so bad on the same play that it's hard to put a value on him.

But, the point was right. Because when someone is paired with a terrible defender to help "make up" for their bad defense, it alters their defensive stats significantly.

As it should. Again, RAPM is a stat about efficacy. About lift. About value. It is not a statement about absolute goodness, and you know what? The dirty little secret about stats is that almost none of them can be taken as a statement about absolute goodness. I mean your FT% is your FT%, but basically everything else depends on context.

I'll also note that the dilemmas you speak of with Brewer and the variance you describe in his efficacy aren't noise but are a fundamental part of what he is as a basketball player and it is something valuable to capture in data form. When I see oscillation in the +/- data of a role player I don't see that as an error, I see that as an inverse measure of his portability as a player.

[Witzig-Okashi]

I had a discussion with a user on this subject around two months ago (or at least six weeks, cannot remember for certain).

I had wondered if there was a stat that could measure the value of steals in a game, since other than offensive fouls like charges, they are the only ones that definite possessions guaranteed for a team that forces turnovers (blocks do not come away with that guarantee, of course).

The only problem is that there would have to be some metric to negate the risks involved with gambling for steals; how would players like Iverson or Francis who would gamble for steals on a regular basis look on the metric with players like Rubio or Paul that seem to get more of their steals on-the-ball/guarding a man. Can this even be determined with a metric?

Sorry if I had stated something others may have alluded to prior to this post, but the article is interesting, to put it vague terminology (intentionally)....

Nice post, Doc....

[RSCD3_]

I had a discussion with a user on this subject around two months ago (or at least six weeks, cannot remember for certain).

I had wondered if there was a stat that could measure the value of steals in a game, since other than offensive fouls like charges, they are the only ones that definite possessions guaranteed for a team that forces turnovers (blocks do not come away with that guarantee, of course).

The only problem is that there would have to be some metric to negate the risks involved with gambling for steals how would players like Iverson or Francis who would gamble for steals on a regular basis look on the metric with players like Rubio or Paul that seem to get more of their steals on-the-ball/guarding a man. Can this even be determined with a metric?

Sorry if I had stated something others may have alluded to prior to this post, but the article is interesting, to put it vague terminology (intentionally)....

Nice post, Doc....

I am that user haha

It was interesting especially now because of the steals-nba success correlation in ncaa competition that I have read about recently.

If the average +/- value of a gamble/steal attempt could be determined by setting up a definiton of a steal attempt ( would probably need sports vu cameras for this) it could possibly quantify the difference between which players help or hurt their team by trying to force turnovers.

l

Sent from my SCH-I800 using RealGM Forums mobile app

[Dr Positivity]

538 is a mess so far IMO. It seems like a site targeted at a mainstream audience casually interested in data/stats instead of the hardcores, so the articles are shorter and simplistic. The problem is this just makes the feel incomplete and half baked and brutally victim to small sample size issues, which contradicts Silver's fox mentality.

Not sure I know how a site could make stats articles for a mainstream audience, but so far 538 does not seem the way.

[5DOM]

538 is a mess so far IMO. It seems like a site targeted at a mainstream audience casually interested in data/stats instead of the hardcores, so the articles are shorter and simplistic. The problem is this just makes the feel incomplete and half baked and brutally victim to small sample size issues, which contradicts Silver's fox mentality.

Not sure I know how a site could make stats articles for a mainstream audience, but so far 538 does not seem the way.

http://fivethirtyeight.com/features/pre ... -of-chalk/

[thizznation]

That article didn't make sense...