I can't speak for Roy, but I'm not standing here claiming that Powe is a better defender than Davis. I'm simply agreeing that IF Davis is a better defender, then that fact does not manifest itself in the team results. Thus I'm suspicious that when comparing the two, if Powe is better at everything, except that Davis's team defense swings the pendulum in his favor, but there are no indications that his team defense actually is any better, the argument begins to sound specious.
But you are wrong.
I'm not sure how the underlined statement was wrong. Read it again:
IF Davis is a better defender, then that fact does not manifest itself in the team results.
The numbers do not show that Davis is a better defender. So IF he is a better defender, the numbers do not show it.
I never said he definitely wasn't a better defender.
Read my next sentences: The fact that the team allows the opponents to score more when Davis is on the court doesn't mean that Davis superior defense isn't translated to the score. The problem is that you don't know how would the team perform if it was Powe playing those minutes, not Davis.
Why are you assuming that the Davis being a better defender doesn't manifest itself on the teams result? Of course it does! See it this way: when Davis is on the floor the opponents scored 100 points; if Powe was playing Davis role - the exact same minutes -, the opponents would have scored 110.
This on/off plus/minus numbers are very misleading because teammates and opponents produce externalities, so they are supposed to be strictly descriptive. For example, answer me this question Is Perkins a better offensive player than Pierce?
One could easily say that IF Pierce is a better offensive player, then that fact does not manifest itself in the team results. Would that be true? Absolutely not - sit Pierce and Perkins out for a month each to find it out. Those numbers simply don't capture what you think they do. They only thing one can conclude is that the team scored more points per 100 possessions when Perkins was on the floor than when Pierce was on the floor. That's all - any conclusion about the relative quality of the players or how they impact the team results is a fallacy and a logical error.
That's a hypothetical situation you just made up. That's not in the numbers anywhere.
Look at the above example:
When Rondo, Allen, Pierce, Davis and KG are on the floor, they allow exactly 1 point for every opponent possession. In the same amount of floor time, the lineup of Rondo, Allen, Pierce, Powe, KG allows .8 points for opponent possession. That's the clearest example of Powe "taking Davis's minutes" (in your words) and ending up with better defense.
When you make any statement that says "If (blank) then (blank) would have done (blank)" it means you are working outside the numbers available; as I said I can't see any results that concretely back up the idea that Davis is a better team defender. I'm not saying he's a worse defender, I'm saying the available data does not show he's a better one.
Yeah, but those minutes aren't exactly the same: they were played against different opponents, on a different time, with everybody playing differently.
And it's a 50 minutes sample... the biggest proponent of the use of +/- productivity is Dan Rosenbaum, who works for the Cavs. But even him adverts that the metric is completely useless when is not weighted, adjusted and the sample is smaller than several seasons. I mean, what happened during the Summer, in your opinion? Has Davis regressed as a defender? Or Powe progressed? What about the Perkins/Pierce case on offense?
Saying the available data does not show he's a better one strikes me as disingenuous, as there is no data to show that (or the opposite).
Being of the opinion that Powe is overall the better player, I can't see why would anyone think that Powe is the better defender, when the opposite is so obvious, at least from my perspective. As others have said, Davis is way more versatile than Powe, can defend the post way better, and that ends hurting him, as he faces tougher matchups.
Interesting, though, is that if you look at the 82games data, Davis's defense is better against Centers than PF's. He's at around 19 defensive PER vs. PF's.
1. Don't trust 82games positional data. They don't adjust it to whom the players are indeed defending. Nobody is charting those games.
2. Production of the counterpart (even if one charts the games and the counterpart is properly and accurately defined), be it PER or another metric, is a terrible method to evaluate defense.
3. How many minutes does he have at PF, accordingly to 82games? 30, 40? Half of them on garbage time? How's that even slightly significant?