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

[mysticbb]

I think you are asking the wrong questions, if we are trying to get at truer values of these teams and players. Acting as if we already know these things, and any variation from previous expectation is the result of something random or abnormal happening.

Let me rephrase that a bit. Do you think players have a limitation in their ability to help a team win? Each player with a different one? Do you think that players have an average performance level during a part of their career on which a coach can rely on? Do you think consistency is a factor?

As we speak the Bulls and Philly are creating an ever changing new reality on the court, giving us new information that we should be incorporating into a better understanding of the value of these teams and the players on them.

That is most certainly true, but how much do you think players can get away from their average performance level due to that? And how much of that is simple the normal variance in their performance level? Well, look at the free throw shooting, even the best free throw shooters in the game are missing free throws, but they are still showing variance in their conversation rate from game to game, from week to week, from month to month, from year to year. When you increase the respective sample, the variance becomes smaller. I start drawing conclusions after 7 to 10 games, and within that thread most people considered that already too small of a sample. And now we want to draw a conclusion based on 4 games and it is fine? No idea, but that really sounds a bit weird to me.

If the Bulls and Philly were to play another series immediately after this one, but with the same circumstances concerning who was available each game, I would look at this first series as the best evidence to understand the new series to come, rather than what happened during the season against a myriad of different opponents. But either way, you'd want to include this current series for determining value to a heavy degree. Especially for Philly as they were so bipolar during the season.

I agree that I would include this series as well, because it gives us more informations. In basketball more information can help us to reduce the noise (variance). Thus, ignoring it, would be likely not a good idea. BUT, for the most part it shows that the team performance in average against the myriad of opponents is the best indicator. Well, that gets beat by taking the minutes distribution and apply individual prior informed RAPM to it. Or as I showed using my SPM.


We talked about the crunch time thing, and I said already that if I weight those minutes more, Rose' value increased. But overall with more weight on those minutes the error in the out-of-sample test gets bigger. Which means more weight for crunch time is not an overall better predictor. Well, we can understand that by looking at the amount of minutes each team plays with "crunch on". It is really small, and all we do is making a smaller portion of the game more important in the calculation. But that smaller portion also means a smaller sample, and as I wanted to show previously the smaller sample makes variance a bigger issue (but I guess you know that).

Anyway, for the Bulls I agree, with Rose they played better during clutch minutes than without him for more than one season already. And watching the game is also giving an impression why that is the case. Rose can break down the defense with his dribble penetration. The Bulls just don't have a play they can use during those minutes, because that play is always working. Whether that is up to the coaching or a problem related to the player skills is not quite clear, but overall the better clutch minutes teams are usually relying just on things which are working in all other minutes too. The Bulls lack such thing besides Rose' dribble penetration. That makes Rose so important during clutch minutes. But as I pointed out, we talked about that already and we agreed upon that.

And sorry, saying that "everyone can make the call that the series would be close" is ignorant. We had the example of the prediction regarding the Knicks and in the end the best prediction that came out here from the "eye test" was basically saying that the Knicks will not win as many games as the win% would have suggested. When asked to put a number on it, it was basically a joke. The same would happen, if you asked someone to put a number on this series. What does close mean? What kind of scoring margin is expected? When you make that prediction without looking at numbers for a whole season, I bet you are at least as far off as the previous season scoring margin for each team is off as a predictor. Using the stats is giving you a much better predictor, much more accurate, even so accurate that it can beat Vegas on a regular basis. No individual can do that.

[Rerisen]

Do you think that players have an average performance level during a part of their career on which a coach can rely on? Do you think consistency is a factor?

Sure, I just don't think we always know what those levels are exactly (and specific situations can also make that level rise or fall). Especially where there isn't a real large amount of data. You can start drawing conclusions after 7-10 games, but not everyone is gong to agree with those conclusions, depending on the nature of those games 7-10 games, what happened in them, opponents, etc.

Also, even a coach often doesn't know what these consistency levels are. Tom Thibodeau has been obviously sort of wandering in the dark this series, with the sudden injuries, and his rotations throw into chaos. One game he gave Ronnie Brewer a DNP, only to find out that Philly's defense is able to swarm on and near to totally shut down all the off ball action of Kyle Korver and Rip Hamilton. So the next game he goes back to Brewer and plays him 29 minutes, including to close the game, and suddenly Brewer's defense on Philly's guards helps lock down their offense, keeping the game in the muck, which is very much advantageous to the Bulls staying competitive with their own struggling offense.

So already we have players being used in the series much differently than they were used during the season - and thus differently from the data set we would be using to try and predict the series.

And now we want to draw a conclusion based on 4 games and it is fine? No idea, but that really sounds a bit weird to me.

Not me. Though I see others might be trying in reviving this thread.

What I see, is this thread is too much about defending turf now. On one side we have those grasping the regular season Bulls without Rose as the ‘true Bulls’ and the other side, some saying these more struggling Bulls against Philly are the true Bulls. BOTH are the true Bulls. The difference in results is from circumstances that I don’t find hard to understand at all, nor which needs any invoking of luck based factors to make sense as the same team. Rather better understanding of team vs team matchups, and perhaps even a little bit of psychology, as frustrating as that realty is dealing with human players.

BUT, for the most part it shows that the team performance in average against the myriad of opponents is the best indicator. Well, that gets beat by taking the minutes distribution and apply individual prior informed RAPM to it. Or as I showed using my SPM

The overall team strength used with RAPM for individual players in or out is no doubt good predictors in general, and maybe the best numbers wise, and far better than going in blind in a situation you otherwise have little knowledge of. But I don’t think they are good at understanding the small nuance differences that exist between certain matchups, especially such as in a playoff series against the same team 5-7 times. Nuances that often end up leading to the slight discrepancies in the end real results vs predicted. But these nuances are by no means impossible or even that difficult to comprehend if you have good knowledge of a team and or specific matchups in a series.

It’s been said that the playoffs are all about matchups. Does that get oversold sometimes, perhaps, but matchups do matter. They are not nil. But using a sample of 27, 66, or even 82 games of one team playing against every team in the league to predict the outcome of two specific teams playing seven times…. You can probably get very close sure, but it may be incomplete since it does not account for matchups whatsoever. There are things that happen on a smaller scale than 5 vs 5 in a basketball game, sometimes concerning only two players, or four players, and their particular strengths and weaknesses. Exploit such a situation enough times toward the same advantage, it can begin to throw off more ham handed 5 v 5 or team v team based figures.

Anyway, for the Bulls I agree, with Rose they played better during clutch minutes than without him for more than one season already. And watching the game is also giving an impression why that is the case. Rose can break down the defense with his dribble penetration. The Bulls just don't have a play they can use during those minutes, because that play is always working. Whether that is up to the coaching or a problem related to the player skills is not quite clear, but overall the better clutch minutes teams are usually relying just on things which are working in all other minutes too. The Bulls lack such thing besides Rose' dribble penetration. That makes Rose so important during clutch minutes. But as I pointed out, we talked about that already and we agreed upon that.

Thing is we don't really have good numbers on how great this impact is (though I think this series will shed light on the answer). In the Bulls 18 wins without Rose this year, only two of those games were decided by 5 points or less! So the team actually didn't have to deal with that many crunch time situations.

But now in a series vs Philly, two grind it out defensive teams, and playoff basketball to boot, just about every game is going to be close and contain a much greater # of these crunch time, high pressure offensive possessions, and I think this is where the point differential might start stretching out from what previous predictions would expect.

Though in the end, it might only be 10 points difference either way by the end of the series, but still perhaps enough to swing it. And so that the expected differentials you have used might still look pretty darn good, but yet could ultimately be the difference in not getting the right winner.

So the point I'm making it that of course I think these number derived values have a lot of merit, but should never be the end of the story, and that more micro level analyzing, if done correctly, can bring even more clarity.

To get back to the question of the thread a little, Rose's value as a shot creator and fairly efficient (healthy) volume scorer, is a value to the team that can vacillate greatly, and is not consistently shown to max degree every game. Against poorer teams, or just any time the team is playing well, Rose can just 'game manage' take it easy, get his teammates involved, and they win by 20 still. Maybe Luol Deng helps you win a lot of such games more - since hard effort defense is more consistently game impacting. But when you get in a game where that shot creation becomes vital, and every possessions is an ugly struggle to even get a decent shot up, suddenly Rose's value starts becoming near irreplaceable. And this is just the traditional role of the offensive NBA superstar, to have that extra 'nitro' gear he can turn on by command, but by and large doesn't have to exercise every game, or play every game to his 30 Point capability max effort. In the playoffs, it seems this nitro gear has to be used far more often though, as role players diminish, bench players are used less, etc.

[Dr Positivity]

I don't think this series is enough evidence one way or the other... lots of Bulls injuries, a top 3-5 defensive team in the league in the Sixers, the mental impact of losing Rose

I feel pretty confident just saying though, based on how I judge teams, that the Bulls without Rose are inferior to Memphis and Indiana, two superstarless ensemble teams. A team like MEM is just a more talented version of this version of Chicago because Conley is a big step up on whatever Chicago gets from backcourt dribble drive and Zbo/Gasol instead vs like Boozer's offense is a pretty big blowout as well, while also being a team with tons of size/rebounding and perimeter athleticism. So I would say 43-46 Ws is a good measure for the Bulls. The 55-60 W pace they were playing at? No... just no. There is no way to argue *this* team is like 6-10 Ws better than the Memphis, Indiana types...

[te887848]

Come on guys, don't you get it? 27 is a bigger number than 4! It doesn't matter who the competition is during the 27 games - it's just a bigger number and that's all that matters.

[Doctor MJ]

See, for a minute there, I was about to become the voice of reason, then Asik missed two free throws, and this conclusively proved that the regular season was meaningless and RAPM was worthless, so I'm back to being a moron again. It's really a shame that I wasn't as able to factor in Asik's free throw shooting as well as other people did.

[ElGee]

See, for a minute there, I was about to become the voice of reason, then Asik missed two free throws, and this conclusively proved that the regular season was meaningless and RAPM was worthless, so I'm back to being a moron again. It's really a shame that I wasn't as able to factor in Asik's free throw shooting as well as other people did.

Excuses. Just. Excuses. You know very well 2 free throws by Asik are enough to completely invalidate a statistic and your entire on opinion on just about everything.

[RichardsRival3]

MVP just lead his team to a 4-2 loss against the 8th seed.

[Chris435]

MVP just lead his team to a 4-2 loss against the 8th seed.

Why don't we throw Rose out there without Deng & Noah and see how he does?

[RichardsRival3]

MVP just lead his team to a 4-2 loss against the 8th seed.

Why don't we throw Rose out there without Deng & Noah and see how he does?

You can bet your bottom dollar, the Bulls would win more than 2 games against the offensively challenged 76ers.

[Basileus777]

MVP just lead his team to a 4-2 loss against the 8th seed.

Why don't we throw Rose out there without Deng & Noah and see how he does?

Philly struggled offensively and Chicago crushed them on the boards, I'm not seeing how Noah's injury was a major factor when the Bulls lost because of their offensive ineptitude.

[Rerisen]

Rerisen- crunch time represents something like 5-10% of the possessions in a close game. G2 wasn't close, so it's strange to say it's made the difference or put so much emphasis on it. It's just another part of the game.

Crunch time is not really a hard and fast thing that starts at 5 minutes in a game within 5 points though. A lot of it to me is about pressure. I would argue an entire 4th quarter could be high pressure time in an important playoff game.

That's not my word, that is NBA players saying how much more pressure there is in the post-season, how the game are totally different.

This is what the Bulls scored in the 4th quarters of the 5 games without Rose: 23, 14, 19, 20, 15.

Embarrassing. Not sure why I shouldn't focus on that when that is when the team was collapsing totally out of character against their performance in all the other quarters during the series. (Bulls +/- through 3 quarter, Game 2-6 = +3. Bulls +/- in the 4th quarter, Game 2-6 = -24)

People are joking about the Bulls almost winning Game 6 and Asik's FTs. Well Asik shot 4/7 for the game, that's his usual. If the Bulls HAD won that game, it would not be because they solved any of the offensive struggles I talked about coming down the stretch of these games, but only because Philadelphia is somehow almost as inept as the Bulls offensively and especialy late. And they have had trouble closing tight games themselves for 2 years.

How many more possessions where the Bulls offense just panics, or a poor isolation player is put in position to individually create, in the playoffs vs regular season does it really take to throw off a prediction slightly or have one team win over another in a close series? No more than 2 or 3 more such plays a game can be all it takes.

I also don't get why you are throwing general level numbers at me, when we are dealing with very specific examples here. Specifically that the Bulls supposed 'true strength' without Rose was built on 27 games, where actually most were not that close at all, anything like this post-season series has been. As I pointed out, only 2 of their wins were within 5 points or less, and I think only 2 of their losses as well. So that even if only 3% of the Bulls possessions were 'clutch' ones in that stretch and 8% were clutch ones in this series, it's enough to make a difference, and that is valuable to know in determining player value.

Several of those games they just totally destroyed poor opponents in games where the pressure of not having Rose was simply irrelevant. To me those are not very valuable data games.

I also don't know why the suggestion should be that crunch time should be dismissed as any special period of the game, when practically every player in the NBA sees their %'s drop over this period. Players that can even keep up 40% at such a time are considered extremely clutch. And to me, its easy to understand why superstar talented scorers like a Derrick Rose can hold up their efficiency (while still falling themselves) but far better than guys not used to be in that position, or who have to throw up ad-hoc shots due to the frequency that system offense gets busted up in these times. Superstar scorers are relatively decent at improvising such situations, role players are usually terrible at it. Even in Game 5, in which the Bulls won, the Bulls were not scoring on system offense well. Luol Deng threw in a couple of near miracle off balance shots, that are far beyond his usual scope or skill to do.

Compared to most casual fans on this board, I am more statistically inclined than most. But what I'm seeing in this thread, is almost like a dogmatic entrenching on the stat side supporters, as if you guys feel the methodology itself is under attack as somehow greatly flawed. That's not it. It's just surprising to me that people are going so hard all out to defend what are really pretty small sample sets that we are suppose to believe are telling us ultimate truths about the Bulls, Derrick Rose, Luol Deng, etc.

As was pointed out on the other page, and as early as my first couple posts in this thread, we need only look at Philadelphia themselves to see how crazily different a team can play within the same season in a stretch as small as 27 games. But because the Bulls played at a level in the same span that somewhat mirrors what last year suggested, we are suppoesed to take that as gospel? And are somehow seen to be taking a big steaming dump on the whole model if we don't buy it? I just don't get that at all.

[mysticbb]

Come on guys, don't you get it? 27 is a bigger number than 4! It doesn't matter who the competition is during the 27 games - it's just a bigger number and that's all that matters.

Based on the 27 games from the regular season, the 76ers were supposed to win that game by 2pt. Based on RAPM/SPM, the 76ers were supposed to win by 1pt, based just on the 4 games without Rose in the playoffs, the 76ers were supposed to win by 8. Now, which of those tools is a better predictor?

Philly struggled offensively and Chicago crushed them on the boards, I'm not seeing how Noah's injury was a major factor when the Bulls lost because of their offensive ineptitude.

Noah didn't help the Bulls defensively. Maybe that is now more clear, because the Bulls defense was still there, they played better defense without Noah. BUT Noah is not a negative to the Bulls defense as well as he is a positive to the offense. Noah's offensive game (which is based on post passing and being able to convert in short range), is helping the Bulls. Missing Noah can explain part of the offensive struggle.

Rerisen, RAPM with the minute distribution the series had says the Bulls winning by +1.6, my SPM says they are winning by 1.1, SRS says they winning by 0.4. Well, it was -1.7 for the series. We are off from 2.1 to 3.3 per game over a 6 game sample. And people acting like the Bulls were completely destroyed.
And even though you are correct that Asik had his normal free throw rate, a bit luck for the Bulls here and we would talk about a 3-3 series right now. I don't blame Asik at all, he showed his normal performance level, I was disappointed about Boozer the most. But in the end sometimes it comes down to luck, the 76ers having more luck with their shooting in game 2, they having more luck with the free throws in game 3, having more luck with the ref decision in game 4 and the Bulls are not having that more luck with the free throw shooting by Asik in game 6.

Regarding the crunch time. Including a higher weight for this is making the tool worse as a predictor. That's just it. I always thought it is necessary to include such higher weight for crunch time, but in the end most of the things in crunch time are influenced by the variance. We only see a small amount of team having a small significant advantage or disadvantage in crunch time. And even on of the best crunch time teams during the last years, the Dallas Mavericks, were not able to show that kind of "skill" in this season. They lost 3 of their 4 games in crunch time against OKC.

[Rerisen]

RAPM with the minute distribution the series had says the Bulls winning by +1.6, my SPM says they are winning by 1.1, SRS says they winning by 0.4. Well, it was -1.7 for the series

This isn't including game 1 right? Rose played till the last minute in game 1. We should only look at what the model expects in the last 5 games without Rose vs reality.

Regarding the crunch time. Including a higher weight for this is making the tool worse as a predictor. That's just it.

I totally get where you are coming from. But the tool is being made to be most useful en masse for the whole league every team. Which it should be.

I'm just saying, specific to the Chicago Bulls, when you remove Derrick Rose, you are going to get worse results than for most teams removing most players in pressure situations where the team needs to score.

Last year a healthy Rose was the 2nd highest scorer per minute in clutch time in the league! And also averaged 9.8 Assists/48 to boot in those instances. The team is just uniquely weak to his loss in those situations. I think the 4th quarter differentials vs Philly are absolutely showing this. And actually to a degree that is pretty staggering, when we are considering numbers in which even slight differences normally represent massive gaps in team strength.

Of course even with this 'error' if we take it as one, its not going to throw such predictions off whack by huge degrees, as Elgee noted they make up a smaller % of the game (& Note the Bulls strength through 3 quarters in the series without him). But to the small degrees they are apparently off, I think it certainly has to do with things I talked about, rather than just, "Boozer played unusually bad" or the refs made bad calls, and that the Bulls are really better than they showed if you repeat the series same circumstances. I don't believe they are better.

At the same time I believe they could also replay those 27 games without Rose in the season and do just as well again. That is a incongruity in the numbers I can accept, because I believe I understand why it happened, and it makes sense to me. It has nothing to do with a kneejerk reaction such as, "I just refuse to believe Deng is better than Rose."

Anyhow, a lot of words have been spent on this, and now we have no more new games to process on it. However, next year it does indeed look like we will get more Rose-less and Deng-less game to investigate.

[EarlTheGoat]

Some folks in here continue to try and justify RAPM and other BS based predictions.

Do you guys refuse to accept reality?

Bulls without their best player, Derrick Rose, have lost against an 8th seeded team 4-2.

Maybe its time for some of you to question the prediction methods of your formulas instead of trying to pretend none of what happened actually counts, because that is pretty laughable.

[Rerisen]

Want to make one more general comment on this thread, as I think it is winding down.

I think it takes incredible hubris to believe that such a small subset of people in the world would truly know the secret that ‘Luol Deng is really more valuable (better?) than Derrick Rose’ - based on this one specific model of rating players. We have other models that can also be highly effective at predicting (if not quite as good) that suggest nothing of the sort - which should make us suspicious a little off the bat.

I mean this thread title isn't low level baseball Moneyball stuff like say, 'stolen bases are overvalued'. This is more like turn the basketball world on its head stuff. And I don't say that as an appeal to authority, but rather just the absolute incongruity of believing you can turn that basketball world upside down based on what is, in part, how the Bulls played for 27 games in one regular season.

To me it seems beyond obvious that the standard of proof should be so much more thoroughly tested and solidly founded than that, before so readily buying a conclusion. Not least of all by the people most knowledgeable and familiar with the model saying that. Such people should in fact be the *most* suspicious and absolute in their confidence before arguing such a position I would think. In order to not make a huge or premature blunder, of the type that could actually set back the evolution of adopting these valuation metrics, vs advancing them to the broader fan base.

Just such a response as I think you are seeing evidenced by the juvenile troll like bumping that went on in this thread when the Rose-less Bulls appeared not to live up to their suggested rating, however close they came.

[EarlTheGoat]

MVP just lead his team to a 4-2 loss against the 8th seed.

Why don't we throw Rose out there without Deng & Noah and see how he does?

Philly struggled offensively and Chicago crushed them on the boards, I'm not seeing how Noah's injury was a major factor when the Bulls lost because of their offensive ineptitude.

This also.

What happened in this series just comes down to this:

Bulls offensive ineptitude without their best player and MVP.

You are going nowhere if you have a team full of solid role-players or borderline all-stars that cant create their own shot. In the Playoffs, the most important thing you need is shot-creation.

[Rerisen]

The MVP award is very much about circumstance, this was lost on a lot of arguments against Rose for it. I do not think the voters were stupid to appreciate Rose's value to the Bulls last year based on circumstance.

Now the more interesting question, what if the Bulls had say a Ben Gordon, instead of Rip Hamilton. A much more efficient scorer, and guy capable of creating his own offense, even if not a better player necessarily than Rip. (or choose your own equivalent self-starter, I like Ben for how he filled this role previously on the Bulls).

Do the Bulls roll the Sixers? Not collapse so much in the 4th Qs? I think they very well might, then Rose's value perhaps is seen as less again.

In order for the Bulls to win a title, I think they very well might need to 'make Rose less valuable' in this way. But until that happens, he is pretty darn valuable / irreplaceable role wise.

[rrravenred]

You are going nowhere if you have a team full of solid role-players or borderline all-stars that cant create their own shot. In the Playoffs, the most important thing you need is shot-creation.

And of course excellent perimeter defense, crisp ball movement and floor spacing is absolutely irrelevant in the playoffs. Stands to reason. Obviously.

[lorak]

Based on the 27 games from the regular season, the 76ers were supposed to win that game by 2pt. Based on RAPM/SPM, the 76ers were supposed to win by 1pt,

How exactly this calculation should be made?

Code: Select all

minutes   RAPM   RAPM*MIN   player
39   -2,5   -97,5   watson
37   -0,6   -22,2   rip
39   0,7   27,3   asik
42   4,8   201,6   deng
27   -0,5   -13,5   boozer
29   4,0   116   gibson
11   0,9   9,9   brewer
9   1,4   12,6   lucas
5   2,0   10,0   korver
1   -0,2   -0,2   butler
      244   SUM
         
41   0,7   28,7   holiday
29   -1,7   -49,3   turner
31   -1,7   -52,7   hawes
43   2,6   111,8   ai
35   2,6   91,0   brand
30   3,3   99,0   young
26   0,9   23,4   williams
6   0,9   5,4   meeks
   257,3   SUM



What I'm missing here?

[ElGee]

Why isn't the end of the 3rd quarter high pressure?

I think people tend to miss the point by trying to obsess over semantics. The Chicago Bulls are not a good offensive team without Derrick Rose. This is true in the first quarter and it's true in the 4th quarter. And it's true in the last 3 minutes. (They are worse without Noah too.)

The obsession with crunch time needs to stop, whether you define it as final 3 and 3, final 5 and 3, final 6 and 5, final quarter within 12, or final quarter period (all of which I've done across my different clutch studies), the point is is that it is just another part of the game. One of the biggest things to understand, which I don't think anyone is understanding until very very recently, is that there are some strategy shifts that are occurring that are faulty (and some smart) at the very end of close games. So, we could say the "game changes" in a few ways as it winds down...but the game is always changing! So it seems like a strange a thing to harp on if people actually knew this.

The most interesting definition I've heard of this is ronnymac's (and others) about clutch being critical moments of the game or even the opening minutes of a key game (I think of Denver's +13 start tonight vs LA in G6). I find that interesting, not because of the nebulous "what's clutch and what isn't," but because it implicitly suggests EVERYTHING counts. The whole darn game counts -- you can make big plays at any time because they all count the same. If your team of professionals is (sadly) prone to huge psychological lapses, you can "save" them in the first Q or the fourth.

I digress...you cite the Bulls offense late and I say it's an issue throughout the game. I mean, the team got a layup on its biggest possession of the season. This is not to say they are a good offensive team, but they don't lose a game "because" of this sudden failure in crunchtime unless you are under the strange assumption that all other things in the game are equal, including the score heading into whatever you want to call pressure situations.

You cite Chicago's 4th Q points in the last 5 games...what about their 3rd Q in G2? -21 points. Yes, in G3 the 4th was a disaster. But in G4, they gave up 9 points in the 1st and never got it back. Tonight they lost the 1st by 2 points...and that was the "difference" in the game. And while I sympathize with your plight as a Bulls fan, it's also maddening to hear this dichotomous narratives when series are so close...the Celtics were about 3 or 4 plays away from being up 3-2 on Miami last year after 450 possessions. But that's not the story that gets told because it doesn't fit in a headline...as mystic said, it was a close series -- we shouldn't shape narratives because of 2 misses free throws, or the shocking inability of a team to get back in transition D on a FT miss with 6 seconds left.