The Problem that is Nene

[tontoz]

I taped the Pelican-Kings games and watched it over the weekend. Davis is already starting to fill out.

One time Rudy Gay drove and was winding up for a big windmill slam. Gay is long with big hops but Davis met him at the top and it was returned to sender. After the block Davis didn't do any showboating just ran down the floor like any other play. Hard not to like this guy.

Different game but his mindset reminds me of Duncan.

[OnAirDonnie]

If everyone is healthy, the ideal rotation would be

PG Wall - Miller
SG Beal - Webster - Miller
SF Temple - Porter - Webster
PF Pierce - Hump - Blair
C Gortat - Nene - Seraphin

I don't think Pierce is going to survive all season guarding 3s. He has noticeably slowed down and last year played his best ball as an undersized 4.

Nene lacks the mobility to defend well at the 4, but would look great with min behind Gortat

Hump has played better than most expected - thinking he was done - and should be rewarded with a more prominent role.

[FAH1223]

My problem with Nene is two fold.

His free throw shooting sucks.

And his efficiency from the field sucks. His shot selection is poor with that dumb jumper. He should be driving to the basket and putting in work there.

Then again, I understand the spacing issues with playing him and Gortat at the same time.

In Denver he was just so damn efficient.

[Dat2U]

The two biggest predictors of wins are a) TS% and b) ball possession stats (Offensive boards, steals & turnovers). For NBA teams, if you meld those two stats and list teams from best to worst, the list matches a list by total wins at a level of @95%. IOW, there's no way to argue w/ the significance of those numbers for a team.

I've tried to argue with you that stats = wins isn't a zero sum equation but apparently that was met with deaf ears. Anyone with time on their hands and a love of math can backward map their way into a formula that equals up to a lot wins.

Actually, dat, no -- that can't be done "easily" (i.e. it has to be predictive, not a "backward map"). And an unsupported ex cathedra claim that it can be done doesn't change anything. So if you want to tell me I have "deaf ears" -- back at you, amigo: you've got blinders on.

As to "stats = wins isn't a zero sum equation", I fear that's patently false. At the end of the game, they don't give the W to the team whose PF is tallest or the team that made the most incredible passes or the team w/ the largest number of thunderous dunk or the team whose players can 'get their own shot' or the team that... you get my point. They give the W to the team w/ the most points.

"Points" is a stat. And, it's a stat that is the direct result, and can only be the direct result, of the other stats of the game.* The effort of the analysts, mostly academics (economists), is to figure out how to weigh each of those stats so that rolling them up predicts win-loss record w/ maximum accuracy.

* Please think that through before automatically denying it. What else could produce that final stat of points? Now, you may want to argue that to get those stats you need certain kinds of players, but then you'll have to do some research to show that correlation (e.g. do taller power forwards get more rebounds?).

Correlation does not imply causation. A stat like Wins Produced actually does a horrible job of predicting future wins, you're better off using a basic stat like points scored.

[payitforward]

I've tried to argue with you that stats = wins isn't a zero sum equation but apparently that was met with deaf ears. Anyone with time on their hands and a love of math can backward map their way into a formula that equals up to a lot wins.

Actually, dat, no -- that can't be done "easily" (i.e. it has to be predictive, not a "backward map"). And an unsupported ex cathedra claim that it can be done doesn't change anything. So if you want to tell me I have "deaf ears" -- back at you, amigo: you've got blinders on.

As to "stats = wins isn't a zero sum equation", I fear that's patently false. At the end of the game, they don't give the W to the team whose PF is tallest or the team that made the most incredible passes or the team w/ the largest number of thunderous dunk or the team whose players can 'get their own shot' or the team that... you get my point. They give the W to the team w/ the most points.

"Points" is a stat. And, it's a stat that is the direct result, and can only be the direct result, of the other stats of the game.* The effort of the analysts, mostly academics (economists), is to figure out how to weigh each of those stats so that rolling them up predicts win-loss record w/ maximum accuracy.

* Please think that through before automatically denying it. What else could produce that final stat of points? Now, you may want to argue that to get those stats you need certain kinds of players, but then you'll have to do some research to show that correlation (e.g. do taller power forwards get more rebounds?).

Correlation does not imply causation. A stat like Wins Produced actually does a horrible job of predicting future wins, you're better off using a basic stat like points scored.

Tell me Dat, if a team improves stats that have 90%+ "Correlation (that) does not imply causation" with more wins, does it or does it not get more wins? Please answer: more wins or not?

If it does, well I guess that's what "correlation" means, huh?

If it doesn't, please enlighten me as to what "correlation" means. Are you sure you want to make that claim. That it doesn't? Wouldn't you rather obfuscate? Accuse me of something? Or would you rather answer: yes or no.

Btw, what's the subject of this thread, dat?

[nate33]

If everyone is healthy, the ideal rotation would be

PG Wall - Miller
SG Beal - Webster - Miller
SF Temple - Porter - Webster
PF Pierce - Hump - Blair
C Gortat - Nene - Seraphin

I don't think Pierce is going to survive all season guarding 3s. He has noticeably slowed down and last year played his best ball as an undersized 4.

Nene lacks the mobility to defend well at the 4, but would look great with min behind Gortat

Hump has played better than most expected - thinking he was done - and should be rewarded with a more prominent role.

You plan to start Garrett freaking Temple at SF on a team that has Pierce, Porter and Webster?

[fugop]

Tell me Dat, if a team improves stats that have 90%+ "Correlation (that) does not imply causation" with more wins, does it or does it not get more wins? Please answer: more wins or not?

If it does, well I guess that's what "correlation" means, huh?

It obviously does not get more wins unless you're making a causal claim. For example, a team could have all five players crash the offensive glass, significantly improve its offensive rebounding percentages, but reduce its chances of winning.

You're actually just being slightly unclear in your phrasing; raw TS% isn't important, but TS% relative to your opponent. Of course both having more possessions than your opponent and scoring more points per possession than your opponent guarantees a win. Its a tautology, and offers limited insight.

[Zonkerbl]

Actually, dat, no -- that can't be done "easily" (i.e. it has to be predictive, not a "backward map"). And an unsupported ex cathedra claim that it can be done doesn't change anything. So if you want to tell me I have "deaf ears" -- back at you, amigo: you've got blinders on.

As to "stats = wins isn't a zero sum equation", I fear that's patently false. At the end of the game, they don't give the W to the team whose PF is tallest or the team that made the most incredible passes or the team w/ the largest number of thunderous dunk or the team whose players can 'get their own shot' or the team that... you get my point. They give the W to the team w/ the most points.

"Points" is a stat. And, it's a stat that is the direct result, and can only be the direct result, of the other stats of the game.* The effort of the analysts, mostly academics (economists), is to figure out how to weigh each of those stats so that rolling them up predicts win-loss record w/ maximum accuracy.

* Please think that through before automatically denying it. What else could produce that final stat of points? Now, you may want to argue that to get those stats you need certain kinds of players, but then you'll have to do some research to show that correlation (e.g. do taller power forwards get more rebounds?).

Correlation does not imply causation. A stat like Wins Produced actually does a horrible job of predicting future wins, you're better off using a basic stat like points scored.

Tell me Dat, if a team improves stats that have 90%+ "Correlation (that) does not imply causation" with more wins, does it or does it not get more wins? Please answer: more wins or not?

If it does, well I guess that's what "correlation" means, huh?

If it doesn't, please enlighten me as to what "correlation" means. Are you sure you want to make that claim. That it doesn't? Wouldn't you rather obfuscate? Accuse me of something? Or would you rather answer: yes or no.

Btw, what's the subject of this thread, dat?

It could mean that the statistics you can easily measure are actually correlated with some other much more meaningful variable that is not easily measured.

Consider steals. Steals can be generated by playing good positional defense, and are therefore highly correlated with the underlying variable you are interested in, namely good positional defense.

However, steals can also be generated by gambling for steals by jumping into passing lanes, leaving your offensive counterpart completely wide open the 90% of the time that it fails. A player like Gilbert Arenas may have a high steals number and yet play some of the worst positional defense in the league.

[Zonkerbl]

Similarly for blocks. Generally a sound defensive big will get his share of blocks, and blocks are therefore correlated, generally, with good positional defense. However, good positional defense is much more important than blocks - a player like Javale McGee can generate a lot of blocks and be an absolutely putrid defender. Nene is a very solid positional defender and generates relatively few blocks.

[Dat2U]

Actually, dat, no -- that can't be done "easily" (i.e. it has to be predictive, not a "backward map"). And an unsupported ex cathedra claim that it can be done doesn't change anything. So if you want to tell me I have "deaf ears" -- back at you, amigo: you've got blinders on.

As to "stats = wins isn't a zero sum equation", I fear that's patently false. At the end of the game, they don't give the W to the team whose PF is tallest or the team that made the most incredible passes or the team w/ the largest number of thunderous dunk or the team whose players can 'get their own shot' or the team that... you get my point. They give the W to the team w/ the most points.

"Points" is a stat. And, it's a stat that is the direct result, and can only be the direct result, of the other stats of the game.* The effort of the analysts, mostly academics (economists), is to figure out how to weigh each of those stats so that rolling them up predicts win-loss record w/ maximum accuracy.

* Please think that through before automatically denying it. What else could produce that final stat of points? Now, you may want to argue that to get those stats you need certain kinds of players, but then you'll have to do some research to show that correlation (e.g. do taller power forwards get more rebounds?).

Correlation does not imply causation. A stat like Wins Produced actually does a horrible job of predicting future wins, you're better off using a basic stat like points scored.

Tell me Dat, if a team improves stats that have 90%+ "Correlation (that) does not imply causation" with more wins, does it or does it not get more wins? Please answer: more wins or not?

If it does, well I guess that's what "correlation" means, huh?

If it doesn't, please enlighten me as to what "correlation" means. Are you sure you want to make that claim. That it doesn't? Wouldn't you rather obfuscate? Accuse me of something? Or would you rather answer: yes or no.

Btw, what's the subject of this thread, dat?

Why obfuscate when I can simply point out that you or any mathematician can regress team stats to assign a value for points, rebounds, etc to get your desired result (aka mapping for wins) but a question that's been asked many times is how in the world do you translate that information into a player level analysis without losing it's predictive power? If your stat can't predict squat, what is it absolutely good for? Nothing, now lets say it again.

[dobrojim]

If everyone is healthy, the ideal rotation would bePG Wall - Miller
SG Beal - Webster - Miller
SF Temple - Porter - Webster
PF Pierce - Hump - Blair
C Gortat - Nene - Seraphin

I don't think Pierce is going to survive all season guarding 3s. He has noticeably slowed down and last year played his best ball as an undersized 4.

Nene lacks the mobility to defend well at the 4, but would look great with min behind Gortat

Hump has played better than most expected - thinking he was done - and should be rewarded with a more prominent role.

Was that a typo or brain fart? Ideal and Temple starting cannot go together. They just can't.

And Nene defends 4s pretty well.

[closg00]

Updated after Ruz's info.

[Ruzious]

Out probably 2 to 3 games.

Jorge Castillo: Nene saw a doctor today for right foot. Wittman said he doesn’t think he’ll be out a long period of time. Mentioned “a game or two” #Wizards Twitter @jorgeccastillo
Rumors tagsWashington Wizards, Injuries, Nene Share on FacebookShare on TwitterSubscribe to HoopsHype rumors
November 22, 2014 Updates
- See more at: http://hoopshype.com/rumors/tag/washing ... nTv3T.dpuf

[closg00]

Out probably 2 to 3 games.

Jorge Castillo: Nene saw a doctor today for right foot. Wittman said he doesn’t think he’ll be out a long period of time. Mentioned “a game or two” #Wizards Twitter @jorgeccastillo
Rumors tagsWashington Wizards, Injuries, Nene Share on FacebookShare on TwitterSubscribe to HoopsHype rumors
November 22, 2014 Updates
- See more at: http://hoopshype.com/rumors/tag/washing ... nTv3T.dpuf

Thanks, we shall see. Sometimes Nene can "play" at hobbled level Nene.

[tontoz]

We did fairly well without Nene late last season. He has been pretty bad on offense so far. His TS is at 48%, horrid for a big.

[hands11]

http://www.monumentalnetwork.com/videos ... n-11-24-14

Sounds like Randy doesn't think Nene wont be out long 1-2 games

[Zonkerbl]

Actually, dat, no -- that can't be done "easily" (i.e. it has to be predictive, not a "backward map"). And an unsupported ex cathedra claim that it can be done doesn't change anything. So if you want to tell me I have "deaf ears" -- back at you, amigo: you've got blinders on.

As to "stats = wins isn't a zero sum equation", I fear that's patently false. At the end of the game, they don't give the W to the team whose PF is tallest or the team that made the most incredible passes or the team w/ the largest number of thunderous dunk or the team whose players can 'get their own shot' or the team that... you get my point. They give the W to the team w/ the most points.

"Points" is a stat. And, it's a stat that is the direct result, and can only be the direct result, of the other stats of the game.* The effort of the analysts, mostly academics (economists), is to figure out how to weigh each of those stats so that rolling them up predicts win-loss record w/ maximum accuracy.

* Please think that through before automatically denying it. What else could produce that final stat of points? Now, you may want to argue that to get those stats you need certain kinds of players, but then you'll have to do some research to show that correlation (e.g. do taller power forwards get more rebounds?).

Correlation does not imply causation. A stat like Wins Produced actually does a horrible job of predicting future wins, you're better off using a basic stat like points scored.

Tell me Dat, if a team improves stats that have 90%+ "Correlation (that) does not imply causation" with more wins, does it or does it not get more wins? Please answer: more wins or not?

If it does, well I guess that's what "correlation" means, huh?

If it doesn't, please enlighten me as to what "correlation" means. Are you sure you want to make that claim. That it doesn't? Wouldn't you rather obfuscate? Accuse me of something? Or would you rather answer: yes or no.

Btw, what's the subject of this thread, dat?

Having said all that, let's give the example that proves the rule. Let's suppose you have a head coach (let's arbitrarily give him initials EJ) who has absolutely no idea how defense is played at the NBA level. EJ constructs a defense that generates steals by encouraging wing players to gamble in the passing lanes, leaving their offensive counterparts wide open. He also ignores eFG% and focuses on raw shooting percentage, deliberately choosing to leave three point shooters wide open in order to "pack the paint." This team will improve some statistics that are roughly correlated with wins and yet is more likely to lose.

Now let's assume this idiot coach gets fired, and all other idiot coaches who have no idea what they are doing. Then you can assume an improvement in all the various correlated variables should lead to an improvement in winning.

So I think you are right, PIF. But there are some rare exceptions. Because correlation does not necessarily imply causality.

[payitforward]

Tell me Dat, if a team improves stats that have 90%+ "Correlation (that) does not imply causation" with more wins, does it or does it not get more wins? Please answer: more wins or not?

If it does, well I guess that's what "correlation" means, huh?

It obviously does not get more wins unless you're making a causal claim. For example, a team could have all five players crash the offensive glass, significantly improve its offensive rebounding percentages, but reduce its chances of winning.

You're actually just being slightly unclear in your phrasing; raw TS% isn't important, but TS% relative to your opponent. Of course both having more possessions than your opponent and scoring more points per possession than your opponent guarantees a win. Its a tautology, and offers limited insight.

Tautology, no, that's a term in logic. But you mean it's true by necessity. Yes, that was kind of my point.

And the correlation I pointed out was, exactly "over a season." Hence, yes, "raw" (i.e. overall team) TS% is important. It's what we're talking about. I.e. I was talking about Nene's *overall* TS%.

Your point about having all 5 players crashing the offensive glass is non-relevant just on its own terms -- you'd have to tell me what that would accomplish. I.e. if a team got *all* its potential offensive boards, i.e. got the ball back after every missed shot... well, I think the point is obvious. But of course that doesn't happan.

But I didn't say it would happen! Or suggest coaches have everyone crash the glass, or even change strategies in any way. What I said is that the net of high offensive boards / low turnovers / high steals help teams win in a way that is very very significantly correlated with actual results over time. Are you sure you want to disagree with that? Because, after all, it's an empirical fact, it can be demonstrated. As can the TS% claim.

Moreover, the distinction between correlation and causation is irrelevant in this case. Statistical facts don't work that way. We're only looking for correlation! Statistics can only give you that.

E.g. if I said that, statistically, college rebounding skills tend to translate to the NBA better than college shooting skills, you could check whether that was true and, if so, you might conclude that it makes sense to keep that fact in mind as you target players to draft. But the statistical fact wouldn't suffice to predict that *a particular player* would become a terrific rebounder.

Statistical facts correlate with statistical outcomes. If a team improves a stat that has been shown to correlate w/ improved results (over significant time and sample sizes), then it's really kinda likely that the team will experience improved results over time.

Frankly, if you deny that -- lets say, because you just don't like the fact that *I* said it because you don't like the way I post or whatever -- how can you expect to be taken seriously? I.e. in that case, what *should* a team try to improve? Stats that *don't* correlate w/ improved results? How far down the drain do you want to wash your own mind, dat?

[payitforward]

Let me say that I've had enough of this unpleasantness. Why Ruzious and Dat2 decided they wanted to fight with me about something as obvious as all this stuff, I don't know.

I can't remember a single issue about which I've differed w/ Ruzious, and the only issue I've debated w/ Dat in the past (as far as I can recall) was whether Trevor Booker was a good player for the Wizards -- obviously a dead issue. But the both of them displayed some very tangible mean-spiritedness in our exchanges. Or, maybe I'm just being overly sensitive -- that's certainly possible, and I'd be unlikely to be a good judge of whether it's true or not. If so, my bad.

In either case, screw it I'm doing a reboot. Go Wizards!

[padovani31]

Season comes and season goes and the same topic discussed. I really wish Nenê traded for the ones who blames him see the real impact of his game. For now, 2 games without him, two losses.

Ask people in Denver about the trade.