All-Defensive Teams announced.

[Lamrock]

But the one minute they did play, they were damn efficient!

[Andrew]

I'd go with the players who are getting the minutes. Presumably, the individual awards are intended to recognise outstanding individuals and someone who is playing less than 15 minutes doesn't seem to fit that description. Even if those players can be shown as being highly efficient or excellent per-48 or per-36 stats, that doesn't necessarily make them better than players who are slightly less efficient or have slightly lower numbers and did play somewhere between 32 and 48 minutes per game.

While fewer minutes means less of an opportunity, it also means less time to make mistakes. The longer a player is on the court, the greater the chances he'll pick up a turnover or foul here and there and the more challenging it becomes to maintain the same efficiency. Furthermore, a player who's played 36 mpg has demonstrated what he can do when actually given that amount of court time. A player 15 mpg player whose numbers are projected over 36 mpg hasn't proven he could maintain the same effort and produce those numbers if given that opportunity.

I think you've also got to look at noteworthy marks and milestones to a certain extent. To use an extreme example, if a player set some ridiculous record with 800 blocks it's an effort that deserves recognition, even though blocked shots are not the be all, end all of defense or defensive statistics.

[benji]

I hate having to go over this over and over and over again.

A player 15 mpg player whose numbers are projected over 36 mpg

They aren't projections, they are standardizations.

The longer a player is on the court ... the more challenging it becomes to maintain the same efficiency.

Again, is this true? I have yet to see any significant evidence that this is true. Indeed, everything I've ever seen says that players maintain their efficiency and even play better when they get more minutes.

I'd go with the players who are getting the minutes.

So the five players who played the most minutes in any season should get the First Team awards, and the one player who played the most should be MVP and DPOY?

Even after three "paragraphs" I'm still completely unclear on how you would choose all-defense players outside of "players who get minutes" or "set noteworthy marks".

[Andrew]

They aren't projections, they are standardizations.

Alright, I used the wrong term. My point is that there's no proof a player would actually average those numbers if they played 36 minutes rather than 8, 12, 15, 13, whatever. There's no proof that he wouldn't, but that doesn't automatically prove that he would.

Again, is this true? I have yet to see any significant evidence that this is true. Indeed, everything I've ever seen says that players maintain their efficiency and even play better when they get more minutes.

I wouldn't say it's always the case but I think it's fair to say that with increased time, these is just as much chance of a player making mistakes and affecting their statistical production as they do of maintaining or improving upon their previous numbers.

So the five players who played the most minutes in any season should get the First Team awards, and the one player who played the most should be MVP and DPOY?

No, and I did not say that. I would suggest that the players playing the most minutes are the more oustanding players in the league compared to someone who's barely playing. Maybe that player who's hardly seeing any court time is the more talented player or could one day be the better player but they haven't proven that yet.

I would not simply choose players with the five best minutes per game averages. I would choose the players with the best defensive statistics who played a significant amount of minutes (let's say more than half a regulation game, for example). I would choose those players over someone who's warming the bench for over 30 minutes of the game. I would also at least give consideration to players who set individual records. Flawed or not, that's how I'd make my picks.

[benji]

My point is that there's no proof a player would actually average those numbers if they played 36 minutes rather than 8, 12, 15, 13, whatever. There's no proof that he wouldn't, but that doesn't automatically prove that he would.

But that is not what the stats are telling us. They are telling us, per 36 minutes the player averages that.

I think it's fair to say that with increased time, these is just as much chance of a player making mistakes and affecting their statistical production as they do of maintaining or improving upon their previous numbers.

It is perfectly fine to have a hypothesis. That's the first step. The second step, and the question is: Where's the evidence to support or deny the hypothesis?

I would suggest that the players playing the most minutes are the more oustanding players in the league compared to someone who's barely playing. Maybe that player who's hardly seeing any court time is the more talented player or could one day be the better player but they haven't proven that yet.

I would suggest that NBA executives are not perfect and therefore do not always play the best players the most minutes...especially with their fetish for known crap over unknowns.

I don't understand this obsession people have with the "they haven't proven it" meme. If I say Carl Landry averaged 17.3 pts and 10.5 rebs per 36 minutes this season, he's proven it. He did it.

I would choose the players with the best defensive statistics who played a significant amount of minutes (let's say more than half a regulation game, for example

Should've said that in the first place...it's barely different from my general criteria. (Although I'd expand it to a third of minutes for a third team.)

[Lamrock]

I hate having to go over this over and over and over again.

No you don't.

I look at it a different way and say that it goes to the guy getting 35 MPG because his defensive output overall is greater than the 15 MPG guy, even if his per-minute output is less. Sure, if the 15 MPG guy played 35 MPG, theoretically he would have a greater output, but playing just 15 minutes he does not, thus he helps his team more than Mr. 15 MPG.

[benji]

I'm not arguing that a player who has played less should be considered equally with one who has played the entire game, infact, I haven't even argued anything yet.

The problem is, if you start following that theory, then we're just going to have to basically pick the top five minutes players for the season. I only have data of Atlanta through Milwaukee and:
[table][mrow]Player[mcol]MP
[row]iverson,allen[col]3424
[row]johnson,joe[col]3343
[row]davis,baron[col]3196
[row]bryant,kobe[col]3192
[row]richardson,jason[col]3149[/table]
[table][mrow]Player[mcol]Stops
[row]iverson,allen[col]914
[row]davis,baron[col]862
[row]bryant,kobe[col]806
[row]smith,josh[col]800
[row]richardson,jason[col]764[/table]
As a counterpoint:
[table][mrow]Player[mcol]PT%[mcol]Stops36
[row]smith,josh[col]72%[col]10.2
[row]farmar,jordan[col]43%[col]9.8
[row]rondo,rajon[col]58%[col]9.6
[row]lowry,kyle[col]53%[col]9.6
[row]wade,dwyane[col]49%[col]9.5
[row]allen,tony[col]35%[col]9.5
[row]garnett,kevin[col]59%[col]9.4
[row]hayes,chuck[col]40%[col]9.4
[row]carter,anthony[col]49%[col]9.4
[row]wallace,rasheed[col]59%[col]9.4[/table]
Or:
[table][mrow]Player[mcol]PT%[mcol]Stop%*
[row]farmar,jordan[col]43%[col]13.8%
[row]smith,josh[col]72%[col]13.4%
[row]rondo,rajon[col]58%[col]13.2%
[row]wade,dwyane[col]49%[col]13.1%
[row]lowry,kyle[col]53%[col]13.0%
[row]garnett,kevin[col]59%[col]12.9%
[row]carter,anthony[col]49%[col]12.7%
[row]iverson,allen[col]86%[col]12.6%[/table]
*(Rebounded Counterpart Missed Shots + Counterpart Turnovers + Rebounded Blocks + Steals - Value of Shooting Fouls)/Opposing Team Possessions

Or...same thing over Counterpart Possessions instead:
[table][mrow]Player[mcol]PT%[mcol]Stp%
[row]camby,marcus[col]69%[col]59.0%
[row]garnett,kevin[col]59%[col]58.9%
[row]smith,josh[col]72%[col]57.4%
[row]wallace,ben[col]41%[col]56.6%
[row]billups,chauncey[col]64%[col]56.1%
[row]rondo,rajon[col]58%[col]55.9%[/table]

[Matt]

OK, turnovers forced is a good start. Now missed shots, is it a stop if the defender doesn't put a hand up or jump but just stands there offering no resistence? Ideally any measure would consider the shooting abilities of the offensive player too.

[benji]

Yes, again, ideally. But I already explained that these are estimations from available compiled data as I do not have the capability to log all 2400 games, nor desire to further parse the PbP.

It is still far more information than simply "WOW, I SAW THAT GUY LIKE TOTALLY HUSTLE, MAN THE PACERS HAVE THE BEST DEFENDERS!"

[Matthew]

It's good information, but it doesn't take into account the ability of the offensive players being guarded nor does it take into account help defense. Because there are still holes in the data, it shouldn't just be used in determining who is a good defender.

[benji]

So, then, you're arguing there is no possible way to evaluate defensive production or performance?

[Matthew]

No I'm saying those stats can't be the be all and end all of who is the best defenders because they don't encompass all areas of defense.

[benji]

Who said they were?

[Matthew]

I said they weren't.

[benji]

But why would you bother just stating the obvious? We already have people to do that.

[Matthew]

Nobody said what I said (Matt/ Shep kind of touched on it), and is it the obvious? People see numbers and they go o0o Jordan Farmer is a good defender!! But they don't stop to think hey what help defense does he provide? Who is he guarding? etc etc. That's the point I was making.

[benji]

Yes, but it's an obvious point.

The discussion was about the All-Defensive Teams and methods to improve analysis to choose the players. Bringing up that all methods have flaws, so we can't truly evaluate the players is fine and good, but boring and obvious.

The numbers are information, you can reject information if you want, but some people like to gather as much as possible.

[Matthew]

Where did I reject it? I simply said it shouldn't be the only way of ranking defenders.

[benji]

Yes, but that is obvious and nobody even claimed that. So you're just arguing against strawmen.

[Matthew]

What other alternatives have you or anybody else posted? I explained why I disputed the defensive stats earlier.

Maybe I should just agree with you. LULZ Benji you are so right! You criticise everyones ideas of how defensive players should be ranked, and you offer stats as an answer. But when someone questions stats, they are pointing out the obvious or arguing with strawmen! What idiots!

[benji]

Well that one was certainly a strawman.

You can dispute what you think the stats are supposed to be saying, you cannot dispute what the stats are actually saying.

I posited a few stats as an alternative to simply total stops/minutes. Never were the stats claimed to be any kind of answer or anything "end all" but instead a way forward to improved analysis methods. You argued that the stats I offered do not measure everything in the world, something that is obvious, as well as something that no one ever claimed.

Maybe I should just agree with you. LULZ Benji you are so right!

This would require you to actually understand what my position is.

[Matthew]

When did I say what the stats are saying is wrong? You're arguing with strawmen!

And you've never made it clear what your position is on this topic, so don't blame me for you being vague.

[benji]

You must have assumed the stats were claiming to be the "end all" and were attemping to claim to "encompass all areas", why else would you bring it up to say they weren't?

[Matthew]

No I must've had. You posted all those stats and never acknowledged the flaws in them. I just pointed them out.

[benji]

I did not state the stats were perfect, and no one rational would believe they were, as it should always be assumed that nothing is perfect.

Whenever you post anything do you point out every single flaw in the footnotes of your post? Sadly not.

I offered a series of methods to further analysis of players, you pointed out the obvious fact that they contain limitations and are not perfect. That's fine, but I'm wondering where the people are who believed otherwise.

Indeed, now I regret calling the statistical methods flawed, as they certainly are not. They return the same values given the same data. The flaw was on your part, by assuming they were attempting to describe everything, especially when I pointed out exactly what they described and it was never claimed they describe anything further. They have theoretical flaws, but nobody has pointed out any yet.

It may be that the provided statistics are not ideal, but they are as ideal as we have so far. Unless someone on here would actually watch all 2400+ games and count.