Why The Mavs Are The League's Deepest Team - And Why That May Be A Problem

[dude1394]

I'm also curious where the on/off court numbers you are talking about are?
82games.com has josh as on +7.9, off +4.2.
http://www.82games.com/0506/0506DAL.HTM

[dude1394]

Originally Posted by Five-ofan

Tim Thomas.

So diow was covering chris kaman?

[Five-ofan]

I got them from the same place. Click on Joshs name from that link. The numbers you have are his overall +- which makes sense since he is the mavs 3rd best offensive player. Josh howard contributes more offensively than defensively. That actually doesnt surprise me at all.

[Five-ofan]

Originally Posted by dude1394

So diow was covering chris kaman?

Largely yes. He also covered maggette some. I kind of adopted the clippers since I like Maggette, livingston , brand and kaman so i watched all of their games.

[jthig32]

Originally Posted by Five-ofan

I got them from the same place. Click on Joshs name from that link. The numbers you have are his overall +- which makes sense since he is the mavs 3rd best offensive player. Josh howard contributes more offensively than defensively. That actually doesnt surprise me at all.

It doesn't surprise me either, I think JOsh os overrated defensively. But like Dude, I find it hard to believe that the Mavs are better defensively when Josh is off the court.

[chumdawg]

One thing here, dude and Matt. Be careful about characterizing the Suns defensively. About halfway through the year last season, they were in the top five in the league defensively.

Basically, athletes will be athletes. Boris Diaw is NOT an awful defender. He's WAY too athletic for that. Athleticism will generally win out in the end.

Now, I know that the Suns faded from the otherwordly defensive numbers they put up early last year. But you dismiss them at your own peril. Athletes WILL win out in the end. Boris Diaw ain't goin' away, that I can guarantee you. Guys like that will crash the boards and defend some shots.

[dirno2000]

Nice observation Chum. The Suns were surprisingly good defensively before Kurt Thomas went down. After that they’re defensive efficiency slipped. In part because they had a tough time rebounding their own glass.

The addition of Marcus Banks, one of the better defensive PG's in the league will only help them defend the perimeter.

Were I a Suns fan, my concern would be Amare. Post microfracture Chris Webber is a turnstile. Amare's younger and his body may bounce back but after seeing Jason Maxiel light him up in Las Vegas, I'd say the jury’s still out.

[nashtymavsfan13]

If Amare comes back 100% next season, watch out. If he doesn't, I don't really fear any team.

[Pirate]

This thread has been an example of an incredible misuse of plus/minus stats, where it is used to say something that it really doesnt mean.

The stats cited:
JHo goes off the floor and the Mavs are better defensively.
Diaw goes off the floor and the Suns are worse defensively.

The (flawed) conclusion made:
This indicates Diaw is better than Josh on defense.

Reality:
It says no such thing.

What plus/minus does accurately (and the only thing it does) is to compare a player to the player who replaces him. That's all it can do. It is such an overused and mangled metric now.

What the above examples really tell us is the following ...
When JHo went out of the game, he typically was replaced by a BETTER defensive player. Griffin. Check.
When Diaw went out of the game, he typically was replaced by a WORSE defensive player. (Whoever.)

The only way you could accurately use these plus/minus stats to compare Diaw and JHo would be if Griffin was the replacement for both players, and of course that isnt the case.

[Five-ofan]

Originally Posted by Pirate

This thread has been an example of an incredible misuse of plus/minus stats, where it is used to say something that it really doesnt mean.

The stats cited:
JHo goes off the floor and the Mavs are better defensively.
Diaw goes off the floor and the Suns are worse defensively.

The (flawed) conclusion made:
This indicates Diaw is better than Josh on defense.

Reality:
It says no such thing.

What plus/minus does accurately (and the only thing it does) is to compare a player to the player who replaces him. That's all it can do. It is such an overused and mangled metric now.

What the above examples really tell us is the following ...
When JHo went out of the game, he typically was replaced by a BETTER defensive player. Griffin. Check.
When Diaw went out of the game, he typically was replaced by a WORSE defensive player. (Whoever.)

The only way you could accurately use these plus/minus stats to compare Diaw and JHo would be if Griffin was the replacement for both players, and of course that isnt the case.

I sort of mentioned that when i said if the suns went from 110 to 115 with and without him it would be a little different but overall i think we would all agree that the rest of the mavs are better than the rest of the suns defensively right? So my point that the suns with diaw are better than the mavs with howard defensively probably does mean that he is a better defender.

[grndmstr_c]

Josh was replaced by Stackhouse as often as he was replaced by Griffin.
And the other two main frontcourt players for Phoenix were Marion and KT, who aren't defensive slouches.

I tend to side with 5-0 on this bit. A responsible reading of the available +/- stats is, IMO, consistent with Diaw having been a stronger defensive player than Josh last year during the regular season (the playoffs, when both had a strong net but only Josh had a strong on-court, are a different matter - defensively capable backup at SF anyone?).

[V2M]

Josh has long, quick hands but I don't believe he has quick enough feet to keep his guy in front of him.

I think Devin Harris was our best perimeter defender last year. His problem may be that he's not strong enough to hold up in the post... but given that not too many guards are posting any more, I think Devin may actually be our most important perimeter player, this year.

[Pirate]

The problem with all these answers to my plus/minus objection is that the rebuttal flies real quickly into the subjective area - ie he looks better defensively, he has long arms, he this or that, the other players this or that ....

That's a far cry from asserting that the plus/minus is some sort of OBJECTIVE proof that Diaw has some sort of defensive advantage over JHo. It proves nothing in comparison between players on different teams.

Now maybe you still THINK that Diaw is a better defender. Or maybe you think JHo is a better defender. But it's just your opinion. Plus/minus is not any valid statistical comparison.

[grndmstr_c]

It's valid. There's error from uncontrolled variables just as there is with any inferrential quantitative tool, but that's no reason to throw it out completely.

[EricaLubarsky]

plus minus measures impact rather than ability, but it is a good tool

[Pirate]

Plus/minus measures a player against his replacement. Period. That's all it can do. Since the replacement for a player on one team is so different from the replacement for a player on a different team, that "uncontrolled variable" as you've termed it is a variation in the crucial statistical BASE you are working off of. It renders inter-team comparisons useless.

Plus/minus works well in telling a team which of its lineups are most effective. For comparing players on different teams? Not reliable at all. In fact, it is even incredibly flawed in trying to compare players on the same team who play different positions.

[grndmstr_c]

Error's are going to negatively impact the precision of inferrential statistics regardless of where they occur. It's a boon to those of us with minds open enough to appreciate +/- analyses that one of the primary sources of error in those statistics is so incredibly transparent. You want to continue to ignore them, fine - it takes a fair bit of effort to bend the data at 82games to questions of the sort that got this debate going, and some people may want to spend their time doing other things - but your complete disdain for the potential of these tools for player evaluation is misguided at best.

[Pirate]

You can continue to don your pompous "professional stat-meister" air and stick your nose in the clouds at us mere mortals, but repeated misuse of a particular stat metric doesnt make that misuse any more reliable or authoritative.

Comparing plus/minus between players on different teams isnt valid as an "objective stat" at all and shouldnt be construed as such. The people who create that stat will tell you that very thing, but people want to overuse the stat to say something thats nice and tidy but unfortunately not possible to say from such stats.

Use it for what it can tell you objectively and its valid. Use it to say added things and you get an excuse for nothing more than opinions if the misused stats are all there is to base them on.

Nothing wrong with opinions. But thats all these we are getting here so far.

[dude1394]

I can't see how a plus-minus stat between two players especially at different positions and vastly different styles of play are that useful. It can be used to bolster the case of one player being a good defender and ergo consideration for said player to have consideration against another...but the +/-'s between them don't seem that relevant.

[grndmstr_c]

If I'm sticking my nose in the air it's only to more deeply partake of the sweet fragrance of basketball enlightenment. I'm going to go back and quote myself, because you're obviously losing sight of what my position is and exactly how far I'm willing to go in my interpretation of these stats in this case:

"I tend to side with 5-0 on this bit. A responsible reading of the available +/- stats is, IMO, consistent with Diaw having been a stronger defensive player than Josh last year during the regular season"

If you really think I'm overstepping my bounds with that then I seriously doubt there's any common ground to be found here. And to be clear, I've never claimed that the interpretive use of +/- stats was 100% objective. Surely someone who so steadfastly refuses to incorporate statistics into their repertoire can appreciate the creative side of statistical inference without feeling the need decry a perfectly reasonable application of the statistics themselves. The use of artistic license with stats, far from invalidating the opinions they inform, is a neccessity, and all that's required to make the exercise a productive one is a commitment to holding what license is taken accountable to a conservative formulation of the meaning of those stats, which is a far cry from being closed off to the possiblity of taking any license at all as you seem to be.

[V2M]

It may not make much sense to compare +/- numbers of two players playing different positions on different teams... BUT here the comparison is for the differentials in those +/- stats. Considering how wide the margins are, we can't entirely discard them. Given the extraneous variables, we may not be able to conclusively prove anything... likewise, we cannot totally ignore those as baseless, especially when the sample sizes are sufficiently large (in general, the regular season samples are large enough... playoffs of any one single season aren't, in my opinion!).

[Pirate]

"interpretive use"
"Creative side of statistical inference"
"use of artistic license with stats"
"license is taken"

Those are all fancy words for "if you can find a stat that looks like it's relevant, its valid to claim any conclusion you can get away with, even if it isnt really being said by the stats."

Got it!

[grndmstr_c]

Originally Posted by Pirate

"interpretive use"
"Creative side of statistical inference"
"use of artistic license with stats"
"license is taken"

Those are all fancy words for "if you can find a stat that looks like it's relevant, its valid to claim any conclusion you can get away with, even if it isnt really being said by the stats."

Got it!

No, they're delicate ways of saying, "facts are made to be interpreted and have their meaning debated, and if you disagree with my interpretation then why don't you engage me in a real debate about it instead of hiding from opinions that differ from your own by labeling them invalid". Got that?

Hell, you could just say: "even taking into consideration the arguments forwarded in favor of Diaw that centered on the two players' respective teammates and their net and off-court defensive numbers, I don't consider the margin of the on-court difference large enough to be particularly informative and so am inclined to trust in the opinion I've formed from watching them both play," and that would be fine with me.

[Pirate]

V2M ... The sample size doesnt shed any added light, when the foundation of the comparison is skewed in the same manner every time. You are - with greater number of games for each player - merely repeating the same "experiment" (the one that didnt tell you anything in the first place).

Here is an example of how this comparison is faulty.

Let's put the ability to play defense on a scale from 1-10, with 1 being horrendous and 10 being all-world. From plus/minus, we know Griffin is a bit better than Jho, Diaw is a lot better than House. Here are some possibilities of objective 1-10 scales of raw defensive performance that fit the plus/minus data.

Griffin 10 Jho 9 Diaw 3 House 1
Griffin 8 Jho 7 Diaw 5 House 2
Griffin 5 Jho 4 Diaw 3 House 1
Griffin 2 Diaw 2 Jho 1 House 1
Diaw 10 House 7 Griffin 2 Jho 1
Diaw 8 Griffin 6 Jho 5 House 4

All we know for certain is
... Jho and House are not the best of the 4 (but each of them could be 2nd best).
....Griffin and Diaw are not the worst of the 4 (but each of them could be 2nd worst)
....Griffin is better than Jho
....Diaw is better than House

Now lets use a specific choice that mirrors the plus/minus and jump to the same sort of "conclusions" being asserted here: Griffin 9 Jho 8 Diaw 5 House 1

- You compare Griff and Jho, and you say "Jho is a negative " (-1). Valid.
- You compare Diaw and House, and you say "Diaw is a HUGE plus" (+4). Valid.
- You go on to say, "Diaw is such a big positive at +4, that he has to be better than Jho who is a pathetic -1." Now you are dead wrong. By looking at the OBJECTIVE scale, we see that JHo is actually significantly better (+3) than Diaw.

If you play a ton of games using our example, Diaw will keep showing up better by a lot than House. Jho will keep showing up not quite as good as Griffin. But Diaw at 5 still wont be better than Jho at 8, even though Diaw has a much higher plus/minus.

[grndmstr_c]

You act like there's no way to get a handle on those things, though, when in fact on-court and off-court +/- numbers,specific lineup +/- numbers, and player pairs data are all useful in snuffing out the kinds of confounds you're characterizing as debilitating.

[Pirate]

GC ... I am not saying you shouldnt interpret facts. I am saying that misapplication of data leads to worthless conclusions, regardless of how diligently you try to make the misapplication seem logical.

Plus/minus compares Jho to Griffin and does it quite well. Take one out, put the other in, and then what happens? It gives results.
Then what about when you plug in Diaw for Griffin? Dont know. He has never been plugged in for Griffin.
If you compare Diaw to HOUSE, he looks incredible. Is he incredible when compared to Griffin? Dont know. Compared to Jho? dont know. All we know is how he compares to House.

Same logic the plus/minus misusers use would say - I am better looking than my ugly friend. You are uglier than your wife. So I must be far better looking than you!!!

[Pirate]

GC ... They are commonly misused. You are correct. You may enjoy misusing them, and many others do likewise, but the creators of the plus/minus themselves tell you that using it as a comparative metric between teams and even between players at differing positions on the same team is a flawed use of the data.

We all want to be able to answer "who is best" with a nice neat simple number. Many like to use it that way. But it really cant define that, even though we wish it did. The highest plus/minus always will go to the star with the worst backup, and the star with the best backup will be comparatively mediocre on plus/minus even though he may be a far better player.

[grndmstr_c]

You're still operating on the assumption that the only numbers that are being looked at are net +/-, which are not at all the only numbers 5-0 and I are looking at. Until you take off those blinders you're going to continue to miss out on what these stats have to offer. I suppose I should at least be thankful that you recognize that your own limited perspective on the numbers would be ill-suited to these kinds of player comparisons.

[nashtymavsfan13]

You make some nice points there Pirate.

[Pirate]

GC ...You think I am missing or ignoring something so my point MUST BE invalid. Neither is remotely true. What I am doing is focusing on the most important core to the comparison, rather than be sidetracked by attempted barbs and sloppy stats. I understand the per100 possessions calculations, but it was the plus/minus component of that which is being cited here as the comparative proof.

The assertion was that Diaw is a +5 p/100/p, therefore he MUST BE amazing defensively. My answer is, he is +5, but that only tells us how he is compared to his replacement.

The assertion was that Josh is a -0.5 p/100/p, therefore he MUST BE mediocre defensively. My answer is, he is -0.5, but that only tells us how he is compared to his replacement.

It's the same core stats. And it still has the same flaw that prevents objective inter-team comparison.

My conclusion would be far different than the one thats trying to be forced here. I think that instead of seeing Diaw as amazing defensively, we are seeing that Diaw's replacement couldnt play a lick of defense. Jthig said this about Diaw leaving the game, which makes more sense to me than rushing out to nominate Diaw to the All-time NBA All-defense team: "For Phoenix, last year, it mostly likely meant they were ultra small, with someone like Jones or Eddie House coming in his place. In House's case especially, that's going to be a defensive drop off no matter who he's replacing."

[V2M]

That's a valid point, Pirate... I have to admit. Just looking at the +/- we can't say conclusively whether Josh is better or not. But earlier in the thread someone (maybe 5-0?) brought up a point saying that presuming the rest of the Mavs (minus Josh, of course!) are better defensively than the rest of the Suns (minus Diaw)... then the fact that with Diaw, the Suns are better than the Mavs with Josh, proves that Diaw has to be better than Josh.

To rephrase in simple equations:

(M - J) > (S - D)
S > M
==> D > J

Note: M being the Mavs, of course; S = Suns; J = Josh; D = Diaw

Of course, the only debatable argument here could be whether the rest of the Mavs are indeed better than the rest of the Suns?!

[nashtymavsfan13]

Originally Posted by V2M

That's a valid point, Pirate... I have to admit. Just looking at the +/- we can't say conclusively whether Josh is better or not. But earlier in the thread someone (maybe 5-0?) brought up a point saying that presuming the rest of the Mavs (minus Josh, of course!) are better defensively than the rest of the Suns (minus Diaw)... then the fact that with Diaw, the Suns are better than the Mavs with Josh, proves that Diaw has to be better than Josh.

To rephrase in simple equations:

(M - J) > (S - D)
S > M
==> D > J

Note: M being the Mavs, of course; S = Suns; J = Josh; D = Diaw

Of course, the only debatable argument here could be whether the rest of the Mavs are indeed better than the rest of the Suns?!

I don't think that equation proves Diaw is better than Josh, I think it shows that Diaw is more vital to the Suns defense than Josh is to the Mavs defense. Because the Suns D isn't as good as the Mavs. Both players are good defensively, but that equation doesn't show who's better.

[Pirate]

V2M ...That equation is insupportable for many many reasons.

To make it easy, lets start with the concept that they are EQUAL with JHo and Diaw in the lineup. Then you take JHo and Diaw out. But in that situation, that's not the ONLY variable. You also bring someone else onto the floor. The Mav replacing JHo is better than Jho defensively, Mavs get better on D. The Sun replacing Diaw is worse than Diaw defensively, Suns get worse on D.

So far we've learned nothing comparatively. All we know is that the Mavs have a defensive upgrade who backs up Jho, and the Suns have a defensive nightmare backing up Diaw.

But doesnt it tell us something that a backup on the Mavs is better than JHo? Not really - we all know (or at least have been told) that Avery prefers D, so it stands to reason that his stand-ins will supply that. And likewise on the Suns, since D'Antoni cares only about O, it isnt surprising that he brings a sub into the game that cant play it at all.

Is it possible, from the numbers being strewn here, that Diaw is better than Jho defensively? Yes, the numbers dont disprove that - but they also dont disprove that Jho is better than Diaw. In other words, they really dont tell us anything conclusive at all, other than JHo's sub is better than he is on D, and Diaw's is worse than he is on D.

We want these numbers to prove something. We want to be able to compare players using "stats." We want to be able to definitely point to a number and say "You are better, you are worse." But you can twist and tweak these numbers all day, and they still cant do it.

[grndmstr_c]

Okay, I'm just going to spell out some of the main points in favor of Diaw. Pirate, your ad nauseum objection to the overinterpretation of #1 taken in isolation is not, nor has it ever been in contention, and it should be more than clear from what I've outlined below that there's considerably more to the argument than what you have thus far appeared capable of recognizing. So if you're going to continue to belabor the thread then at least try to make it substantive and debate the entire argument instead of just one tiny piece of it.

1) Diaw had a superior defensive net +/- compared to Josh last year
2) I expect most would agree that outside of those two players Dallas featured a stronger defensive cast than did Phoenix
3) the defensive off-court pp100 for Diaw and Josh are consistent with #2, above, in that Dallas gave up significantly fewer pp100 with Josh on the bench than did Phoenix when Diaw was on the bench (see V2M's post for the implications of this consideration)
4) while the SF position for Dallas was sometimes manned by Griffin, a player with a strong defensive reputation, when Josh wasn't on the court, that position was manned for an equal number of minutes by Jerry Stackhouse, a player with a poor defensive reputation who is undersized for the position, so it would seem unlikely that the quality of Josh's backup was skewing his defensive net +/- towards the negative side
5) Josh's defensive on-court numbers are far more comparable to Stackhouse's than they are to Griffin's
6) The three high-minute front-court rotation players for Phoenix were Diaw, Marion, and Kurt Thomas. Both Marion and Thomas have defensive reputations that are on-par with Griffin's and considerably better than Stackhouse's, suggesting that it is unlikely that Diaw's strong defensive net +/- was due to his backups being weak defenders.
7) Diaw has had a positive defensive net +/- each of the last three years, whereas Josh has had a positive defensive net +/- only once during that span of time

[V2M]

Pirate and Nashty... apart from what GMC mentioned above, here's a couple notes from Five-0 earlier in this thread:

"They give up 106.1 PP100 with him on the court and 105.6 without him. The suns on the other hand? They give up 105.2 with diaw on the court and 110.3 with him off the court. A huge difference. In essence with Diaw on the court the mavs are better than the mavs defensively and without him they are the suns we all know."

"... despite the mavs being better defensively overall, The suns with Boris Diaw on the court are BETTER defensively than the mavs with Josh Howard on the court.... "

Without getting into all the statistics yet again... just presuming those numbers above are right... I have no problem inferring that, all things being equal, Diaw has a bigger defensive influence than Josh. Yes, we can always argue that there's a zillion other subjective variables that aren't taken into consideration... but if that's so, one can never conclusively argue that even Bowen or Artest is any better than Josh, either. We can, as a matter of fact, reduce all player comparisons (offensive or defensive) to subjective arguments.

[Five-ofan]

They are right, [link] http://www.82games.com/0506/05DAL10D.HTM [/link] for Josh's stats and [link] http://www.82games.com/0506/05PHO11D.HTM[/link] for Diaw's stats.

[V2M]

Originally Posted by Five-ofan

They are right, [link] http://www.82games.com/0506/05DAL10D.HTM [/link] for Josh's stats and [link] http://www.82games.com/0506/05PHO11D.HTM[/link] for Diaw's stats.

Thanks Five-0.

It's interesting to see other defense-related stats:

eFG% Allowed: Phx is 48.7% with Diaw and 49.9% without him (-1.2% diff); Dallas is 48.0% with Josh and 47.2% without Josh (-0.8% diff).

Rebounding: Total rebounding improves 1.7% with Diaw on court for Suns and 0.8% with Josh on court for Mavs. Defensive rebounding, as a % of Total Rebounding, improves from 67% without Diaw to 71.2% with Diaw (+4.2% diff); It goes up from 69.8% to 70.1% for Josh (+0.3% diff).

Fouls Committed: Dallas is 22 with Josh and 21 without him; Phx is 19 with Diaw and 22 without him.

FTs attempted by opponents: Dallas is 25 with Josh and 26 without him; Phx is 22 with Diaw and 26 without him.

Of course, once again, I'm not insinuating that any one of these stats by themselves tell the whole story... but collectively, I think they're painting a clear picture that Josh is not as good a defender as Diaw is. Also, somehow, there's been a perception created around here that Josh is our best defender... and at the very least, these numbers prove that he's certainly not our best defender, especially if we're better off defensively when he's off the court!

[Five-ofan]

None of those numbers in themselves prove anything but ive watched quite a few suns games and around 80 mavs games a year for the last few years plus the playoffs and they just confirm what my eyes tell me. Diaw is better defensively. I said that he was better defensively before i looked at the numbers. They just confirmed what i had already thought. As for josh, he is a talented but stupid defender. He has the skill set to be good defensively but his defensive iq is horrendous. As a matter of fact everything that people say about daniels defensively i think goes double for josh.

[Pirate]

Guys, you are just restating the very same stats that I have gone to great lengths to show the fallacy in. The plus/minus stats - even if they are per 100 - are still plus/minus, with a scale BASED ONLY ON YOUR REPLACEMENT's ABILITY rather than on any standard objective measure.

Those are subsidized with raw opinions, and nothing more.

When we gert to the end, we still have opinion on top of opinion, and nothing conrete - which was my point all the way through.

GC - You take great pains to toss barbs my way. But even though you CLAIM to have added objective data, you dont. It's the same stuff, but you fail to realize it apparently.
item 1 - plus/minus
item 2 - opinion ("defensive reputation") that may or may not be accurate and is subjective at best
item 3 - plus/minus proves itself and proves the opinion, totally circular
item 4 - opinion ("defensive reputation") that may or may not be accurate and is subjective at best
item 5 - plus/minus
item 6 - opinion ("defensive reputation") that may or may not be accurate and is subjective at best
item 7 - plus/minus

V2M ...Both of your items are merely the same plus/minus (with its flaws) that I have been adddressing the whole time. It tells us nothing that is objective. It is a comparison of player-vs-replacement, and the replacement for Diaw is vastly different than the one for JHo.

I am not particularly trying to assert JHo is a better defender than Diaw. But the strong DEFINITIVE assertions being made here, based only on plus/minus, were way beyond the scope of what plus/minus can DEFINITIVELY tell us.

Adding opinions doesnt alter the fact that we are still offering murky opinions.

And thats fine. If you think Diaw is a better defender, you can have that opinion.

But what has been offered here offers no proof whatsoever, by misusing stats outside their valid purpose.

[Five-ofan]

No they dont but the fact that the suns with diaw on the court are better than the mavs with howard on the court does go a long way and that isnt plus minus.