Talk Me Down: How is Deng not MVP of the Bulls?

[Doctor MJ]

Re: Deng plays more time with the bench. Again, the quality of players he shares the court with are accounted for. There's also the matter that using the fact that the starter who plays the most minutes by definition must spend a decent amount of time with bench guys against him just seems odd.

From what I've seen in obviously being an avid Bulls watcher is the 'accounting' of other players is not something that in reality has constant value.

For instance, when the Bulls play scrub teams without Derrick they stomped many of them. And I actually don't think they would have won much bigger even with Rose. But when they play much better teams that can pressure the Bulls limited ball handlers, the Bulls ability to produce good shots without Rose falls to nearly nil.

Therefore, the other players RAPM is attempting to account for with adjustments do not have static value across all situations. Which would normally wash out in the numbers, but relevant to this year, is skewed because of the vast majority of the team's minutes without Rose were played against very poor competition.

Some good thought here.

In general, you are right, the examination of Deng's various +/- numbers being out of whack with his production stats is not really about him vs Rose. Though you did label the post that way. But rather if we are to believe these numbers, Deng is not just more valuable to the Bulls than perceived, but actually more valuable than a bunch of superstars are to their teams, who are generally considered superior. Is Deng really as valuable as Kevin Durant?

An interesting thing to bandy about. I made the thread I did because I thought just focusing on Deng on the Bulls was the most fundamental aspect of this.

I'll admit that depending on what I experience in this thread, I may start mentioning Deng on longer MVP lists.

[drza]

DocMJ, you don't happen to read my Rotowire stuff do you? Because I swear that the lead for my Hoops Lab article this week was about how Deng was very arguably the MVP of the Bulls, for many of the same reasons that you name. Thus, I'm already in the mindset to fully support the argument that you were putting forth.

I also liked David Stern's summary: Deng's defense is ridiculously important to the Bulls' success, and his offense is great as well. Rose has very good impact on offense, but next-to-no impact on defense. In summary, that's why Deng measures out as the more impactful player.

[Rerisen]

I think though that maybe the most useful way to turn this conversation is to say, okay, let's go with the idea that Deng's impact is what it is because the Bulls have no one else like him, how amazing is it that they so desperately need a player like Deng who puts up such average numbers and whose role isn't even very easy to describe?

To get it off Rose a bit, which is volatile if we are going to claim Deng is more valuable, Deng has been a player that has graded out very high in +/- for years and years, I would say well beyond his PER.

So it's not that he is uniquely valuable to this team, even if he has now finally been put into the perfect 3rd option role to make him ever more so. But he's just a lot better than given credit for. Is the truth somewhere in the middle, between his pedestrian PER, efficiency, and production, and yet short of his superstar like RAPM ratings, I think probably.

[Doctor MJ]

basketballvalue has a 2-year APM number going, it also ranks Deng ahead of Rose.

Yet last year Deng's APM was just +1.04 to Rose's +11.66. This year year Deng is at +17.50, with all kinds of strange Bulls lineups due to the injuries, Rose and Deng both missing a noticeable amount of games. To me this suggests unreliability in the number, and I truly believe has a large degree to do with the Bulls being a machine in the regular season, especially against bad teams.

But the construction of the team is rather unique with so much reliance on one guy offensively that this domination cannot hold up without Rose against stiffer competition. There is just no way for numbers to understands how players value can increase or decrease in a fashion that is not linear as competition scales up or down. But which yet could indeed be happening, without being captured unless the sample size without Rose grew much larger, and to include many more higher quality opponents.

The effects of poor sample size are major. Like I say, I'm not trying to make this all about +/-, but it seems relevant to the discussion.

As you talk about the team not being able to hold up against good teams without Rose, I want to remind you that you were just talking about the weak backups at the 3 forcing Deng to play bigger minutes, which is another way of saying that the team can't hold up without Deng.

I suppose what one could conceivably argue is that Deng can have more impact than even Rose against middling teams, but that against tougher teams that would disappear without Rose. I don't know though. That would make some sense if Deng was offense-oriented, but Deng's impact primarily comes on defense.

[Doctor MJ]

DocMJ, you don't happen to read my Rotowire stuff do you? Because I swear that the lead for my Hoops Lab article this week was about how Deng was very arguably the MVP of the Bulls, for many of the same reasons that you name. Thus, I'm already in the mindset to fully support the argument that you were putting forth.

I also liked David Stern's summary: Deng's defense is ridiculously important to the Bulls' success, and his offense is great as well. Rose has very good impact on offense, but next-to-no impact on defense. In summary, that's why Deng measures out as the more impactful player.

Ha. Well, no I haven't been drza, but I would be if I had more time this season. Great minds.

[Doctor MJ]

basketballvalue has a 2-year APM number going, it also ranks Deng ahead of Rose.

Yet last year Deng's APM was just +1.04 to Rose's +11.66. This year year Deng is at +17.50,

Why look at APM when RAPM is available?

2011: Deng +4.8, Rose +2.2
2012 (non prior informed): Deng +3.2, Rose +1.6
2012 (prior informed): Deng +5.8, Rose +3.0

Where do you get the prior vs non-prior breakdown?

[Rerisen]

basketballvalue has a 2-year APM number going, it also ranks Deng ahead of Rose.

Yet last year Deng's APM was just +1.04 to Rose's +11.66. This year year Deng is at +17.50,

Why look at APM when RAPM is available?

2011: Deng +4.8, Rose +2.2
2012 (non prior informed): Deng +3.2, Rose +1.6
2012 (prior informed): Deng +5.8, Rose +3.0

If I'm not mistaken, RAPM requires even more lengthy minute accumulations to become real reliable. More than one year, several years even. Which is fine for Deng's value, yet Rose was not nearly as good a player even 2 years ago.

So for Rose at his current level of play, the only true sample we have is last year and 35 games from this season. What is our standard error in such a time frame.

Generally speaking, I have long felt Deng was undervalued vs his box score points. with on/off stats being a big part of that. But to the point of being more valuable than Rose... cannot swallow that one.

[drza]

Another thing to keep in mind about Deng, this +/- impact isn't a new thing. I see Rerisen alluded to it above, but Deng has measured out as extremely valuable in the +/- stats for a long time now. In Ilardi's '04 - '09 study, Deng ranked #47 overall. In Englemann's '08 - '11 study, Deng ranked #27 overall. Deng ranks above Rose in each of the 4-year Englemann study, the 2-year basketballvalue APM study, and the RAPM for both 2011 and 2012 (prior year informed or not, the links for both are on Englemann's site).

Thus, I really don't buy the argument that Rose is somehow uniquely allowing Deng to have his impact. He has his impact whether Rose is around or not, and has been for a lot of years. I think this is really a case where by trend we're calling Deng a "role player" or a "third option", when in terms of impact he really may be the MVP of the team. Just in an unorthodox way that doesn't show up in the box scores.

[Rerisen]

I suppose what one could conceivably argue is that Deng can have more impact than even Rose against middling teams, but that against tougher teams that would disappear without Rose. I don't know though. That would make some sense if Deng was offense-oriented, but Deng's impact primarily comes on defense.

I agree, Deng's primary impact is defense. Kind of has to be with him scoring 15.9 PPG at .504 TS% currently, and not being a great dribbler or playmaker either.

Yet if we are to believe just how much impact Deng is making on defense, this would have to be rewriting traditional limits as expected from the Small Forward position, right. As in up near dominant center level impact.

[therealbig3]

@drza, or anyone really:

Well, much in the same way that Ginobili seemed to be the MVP of the Spurs, right? I'm not saying that as a "gotcha!" btw, just pointing out another instance where a perceived lesser player seems to have more impact than the star of the team.

I'm legitimately curious, what to make of that? If Deng is to be considered the MVP of the Bulls, are we going to say Ginobili was the MVP of the Spurs? Or that Harden is the MVP of the Thunder?

Again, not asking in a condescending way, just interested in how far we should take that train of thought.

[Doctor MJ]

A lot of good points here. You saw similar above and beyond +/- figures (RAPM as well) for say Lamar Odom, on several recent years of the Lakers, who is interestingly, a similar player to Deng in ways.

But was Lamar really more valuable than Kobe, as in, the Lakers could do better missing Kobe than Lamar in those years.... probably not.

What you are probably getting is a unique combination of two players playing extremely well together, one a superstar and one a role player that plays off said superstar. With the role player by whatever twist of chance or perhaps anomaly, playing well enough when they apart, in far less minutes, that the stat is spitting them out actually ahead of the superstar. But if we grew the gap in the minutes they were separate to equal the minutes they played together, I really question if such trends would hold.

Excellent, was hoping the thread would go here.

Kobe & Odom is a great thing to bring up. Do you realize that the Lakers are actually do better with Kobe on the floor this year than they were last year? Kobe's APM's gone way up, and the Lakers are playing like crap. What's happening?

Well obviously, everyone around Kobe is playing like crap. Why is that? Well, I think the answer has a lot to do with Phil Jackson forcing a healthy system to be in place, while Brown's never really done anything like that. The result is an offense far more focused on Kobe's abilities and predilections, which then can't cope without him on the floor.

Good chance we have something similar with Rose & Thibs. In the hands of a lesser coach, Rose still does his thing, but Deng's impact becomes far less. One might even be tempted to call Deng a "system player", though I still don't think that's accurate enough for my taste.

What we're really talking about here is an epiphyte. An emergent entity in a complex environment that can only thrive with the right structure around him, but can do great things with that structure.

[lorak]

2011: Deng +4.8, Rose +2.2
2012 (non prior informed): Deng +3.2, Rose +1.6
2012 (prior informed): Deng +5.8, Rose +3.0

Where do you get the prior vs non-prior breakdown?

prior: http://stats-for-the-nba.appspot.com/ranking_rec
non prior (it's on main page too, only a bit lower): http://stats-for-the-nba.appspot.com/ranking12_no_prior

[Doctor MJ]

If I'm not mistaken, RAPM requires even more lengthy minute accumulations to become real reliable. More than one year, several years even.

RAPM vs APM, the entire benefit of RAPM is to increase reliability. Taken over a long enough time span, it becomes less accurate than classic APM.

That said, more sample size always helps both of these stats, just like it does any other stat, and yes I'm reluctant to conclude too much off of 30-ish games.

[Doctor MJ]

2011: Deng +4.8, Rose +2.2
2012 (non prior informed): Deng +3.2, Rose +1.6
2012 (prior informed): Deng +5.8, Rose +3.0

Where do you get the prior vs non-prior breakdown?

prior: http://stats-for-the-nba.appspot.com/ranking_rec
non prior (it's on main page too, only a bit lower): http://stats-for-the-nba.appspot.com/ranking12_no_prior

Thanks.

[Doctor MJ]

I agree, Deng's primary impact is defense. Kind of has to be with him scoring 15.9 PPG at .504 TS% currently, and not being a great dribbler or playmaker either.

Yet if we are to believe just how much impact Deng is making on defense, this would have to be rewriting traditional limits as expected from the Small Forward position, right. As in up near dominant center level impact.

Well reasoned, and yes. What we're seeing in longer term APM studies is that while the traditional hierarchy of positional importance is roughly accurate, and while guards really seem not to ever dominate, we are getting some small forwards ranking right up there with the big men. It is definitely a surprise to me.

Honestly, I feel like this is not so much a re-evaluation of how defense is played in the NBA so much as I think it's a reflection of the modern game where 3-point shooting is huge, players are used to all shots being contested, and team coordination for defense is no longer a foul waiting to be called. I also think no coach has his finger on this new defensive worlds more than Thibs.

[Doctor MJ]

@drza, or anyone really:

Well, much in the same way that Ginobili seemed to be the MVP of the Spurs, right? I'm not saying that as a "gotcha!" btw, just pointing out another instance where a perceived lesser player seems to have more impact than the star of the team.

I'm legitimately curious, what to make of that? If Deng is to be considered the MVP of the Bulls, are we going to say Ginobili was the MVP of the Spurs? Or that Harden is the MVP of the Thunder?

Again, not asking in a condescending way, just interested in how far we should take that train of thought.

Well to me this goes back to the minutes played. With Ginobili you can find quite a few metrics that make him look like a megastar, and those are not to be tossed aside, but there's no denying that playing less minutes diminishing actual impact significantly. So yes there are times when Ginobili outperformed Duncan even when the volume stats implied otherwise, but on the whole the Spurs stood on Duncan's strong shoulders.

At this point, I see Harden in a similar boat. Until the Thunder are relying on Harden like they rely on Durant, I'm not going to really have any doubts about who is more valuable over the course of a season. Yes that assumes that the coach knows what he's doing when he limits Harden, but that seems a more reasonable assumption than the alternatives.

[Rerisen]

^Well, much in the same way that Ginobili seemed to be the MVP of the Spurs, right? I'm not saying that as a "gotcha!" btw, just pointing out another instance where a perceived lesser player seems to have more impact than the star of the team.

I'm legitimately curious, what to make of that? If Deng is to be considered the MVP of the Bulls, are we going to say Ginobili was the MVP of the Spurs? Or that Harden is the MVP of the Thunder?

Again, not asking in a condescending way, just interested in how far we should take that train of thought.

Well the crux of the question is maybe what does it mean to be 'most valuable'. You see this same debate in the regular NBA MVP discussion.

And is it possible one player can be more valuable than another within a total system, but less valuable if that system collapsed? This is what Doctor MJ was getting it I believe, talking about a complex system.

For instance, the Bulls win 62 games with Rose AND Deng, and certain numbers are telling us Deng impacted more of that (note others like Win Shares, Ewins, greatly disagree).

But in another hypothetical, we might reason that the Bulls win far less games without Rose for a full season than without Deng. As how would Deng respond to getting gameplanned for a whole season like Rose does now (or at least more anyway), with the team having no legitimate #1 or even #2 really, option.

Without Deng or say Boozer, Rose being a shot creator obviously would have capacity to up his scoring quite a bit to fill gaps, even if it meant his efficiency sliding down some, perhaps to Iverson levels. But Deng's efficiency is already at that level, if he has to score 20-25 nightly for the Bulls, what happens then? What happens at the end of games, when Rose routinely took over and scored more points in clutch time last year than any other player short of Kobe? Historically, Deng is a player whose offense has dwindled in 4th quarters, as his off ball game and lack of creating seems to hamper him against higher intensity defense.

This isn't something that has been tested just by the Bulls running roughshod over a few bad teams without Derrick to where those late games issues never came up. It's quite a bit different for a team to sustain a short such run of a few games without a top guy, than to do it over a full year, when specific weakness will get scouted and exploited more and more.

Then you also get into leadership areas, confidence of having that bail out star around. To me by the time the Bulls finally ran out of gas against NJ their last game without Rose, and scored only 85 points losing at home, I know most fans were thinking we can't keep this up much longer without Derrick. Running out meaningful minutes to John Lucas and Mike James and hoping for Rose impersonations. I think there was probably something true in that feeling. That the longer they went without Rose, the more those games would begin to happen, as the team's intensity began to fall off the championship push it stays on when Rose is there.

[Rerisen]

Well reasoned, and yes. What we're seeing in longer term APM studies is that while the traditional hierarchy of positional importance is roughly accurate, and while guards really seem not to ever dominate, we are getting some small forwards ranking right up there with the big men. It is definitely a surprise to me.

Honestly, I feel like this is not so much a re-evaluation of how defense is played in the NBA so much as I think it's a reflection of the modern game where 3-point shooting is huge, players are used to all shots being contested, and team coordination for defense is no longer a foul waiting to be called. I also think no coach has his finger on this new defensive worlds more than Thibs.

Makes sense, as so much more of a team's high efficiency offense comes from threes. And it used to come all from the paint and FTs, where big men ruled. If you forced teams to shoot outside all game, you had a very good chance to win, as even if people got hot, most of the buckets were worth just 2.

As far as perimeter defense impact, is there a better example of this than the Miami Heat right now, with so much of their defensive strategy dependent on the speed and versatility of LeBron and Wade? A SF and SG.

It harkens back to the 90s Bulls really who controlled offenses in similar ways with Jordan and Pippen flying all over, then just a 'body' in the paint as Phil would often talk about Luc Longely as.

Though sometimes with Miami, I also feel Joel Anthony is a critical cog in that defense that gets underrated.

[drza]

@drza, or anyone really:

Well, much in the same way that Ginobili seemed to be the MVP of the Spurs, right? I'm not saying that as a "gotcha!" btw, just pointing out another instance where a perceived lesser player seems to have more impact than the star of the team.

I'm legitimately curious, what to make of that? If Deng is to be considered the MVP of the Bulls, are we going to say Ginobili was the MVP of the Spurs? Or that Harden is the MVP of the Thunder?

Again, not asking in a condescending way, just interested in how far we should take that train of thought.

The thing is, I don't think that the APM numbers argued that Ginobili was the MVP across the whole of the years that the Spurs were winning. From '04 ' -09 Ilardi's study had Duncan and Ginobili 3rd and 4th with less than 0.2 separating their scores...on-court. But Duncan was playing a much larger volume of minutes, which would then argue that Ginobili was having a superstar impact on the game but since he played less Duncan was having the larger impact overall. Which I'd absolutely endorse. (By the way, the same pattern showed up in the yearly RAPM studies).

And for your other example, both the 2-year APM from basketballvalue and the 2011 full-season RAPM have Durant ranked higher than Harden.

So no, I don't think it's the same thing. In both examples you gave, the APM studies agree with the consensus that the superstar was the more impactful player overall...though they DO highlight that, like Deng, Ginobili and Harden are better than their box scores indicate. But his situation is different...here the "lesser" player has both factors going: Deng ranks out higher in the +/- studies consistently over multiple years, AND Deng also plays more minutes. That makes it a harder sell to me that Deng's impact is some kind of by-product.

[therealbig3]

^Hmm, fair enough, for some reason, I thought Ginobili was always ranked a little bit ahead of Duncan on the APM lists, which made him a very interesting player to me, and it actually almost turned me off to APM. But I guess it's true, they do agree that Duncan during his prime was having similar impact on the court while playing more minutes, and was thus the better player.

The 4-year RAPM from 08-11, Durant's entire career, and the last 2 years for Harden, has Harden at +4.0, and Durant at +1.1, which is why I included Harden in my post.