Where's Powe's Jump Shot?

[nickagneta]

When Baby plays with the same unit, it is Davis that takes the larger, more imposing opposing player. In other words, the tougher assignment physically where he is guarding men often much taller than himself whereas Powe often guards people more his own size.

Also, I don't agree with this logic. If it were advantageous to have Baby guard the bigger player, that would show up in that unit's +/-. If Powe and Davis were interchangeable defensive players, then there should be no difference between Powe/KG and Davis/KG as once on the court each would guard the opponent that best suits them. In other words, if Powe takes the 4 and KG the 5 in the Powe/KG tandem, why wouldn't Davis take the 4 again in the Davis/KG tandem? If Davis/KG defend better with Davis at the 4, then why would they play Davis at the 5 and hurt the overall team numbers? Doesn't make sense.

Except your logic isn't reality.

When KG and Powe are in the game defensively KG takes the center and Powe the PF. When it is Baby and KG KG takes the PF and Baby the center. I could care less about your logic. My eyes know what they see and what they see is what's happening not what is logical in a hypothetical debate.

[Fan from VT]

When Baby plays with the same unit, it is Davis that takes the larger, more imposing opposing player. In other words, the tougher assignment physically where he is guarding men often much taller than himself whereas Powe often guards people more his own size.

Also, I don't agree with this logic. If it were advantageous to have Baby guard the bigger player, that would show up in that unit's +/-. If Powe and Davis were interchangeable defensive players, then there should be no difference between Powe/KG and Davis/KG as once on the court each would guard the opponent that best suits them. In other words, if Powe takes the 4 and KG the 5 in the Powe/KG tandem, why wouldn't Davis take the 4 again in the Davis/KG tandem? If Davis/KG defend better with Davis at the 4, then why would they play Davis at the 5 and hurt the overall team numbers? Doesn't make sense.

Except your logic isn't reality.

When KG and Powe are in the game defensively KG takes the center and Powe the PF. When it is Baby and KG KG takes the PF and Baby the center. I could care less about your logic. My eyes know what they see and what they see is what's happening not what is logical in a hypothetical debate.

Right, and my point is that the reason they would have Davis guard the 5 and KG the 4 is because Davis at the 4 and KG at the 5 would be even worse than KG at the 4 with Davis at the 5, which is worse Powe guarding the 4 and KG guarding the 5

[Scalablob990]

I thought Roy brought some interesting stats to the table regarding who was the better defender statistically but some of the stats he pointed to that confirmed his thesis came from team derived stats such as points per 100 possessions, points per 48 minutes, team point differential and +/-. These are all stats that help but are also reliant upon who you are playing with and when.

So I went to 82games.com and decided to look at exactly who these guys were playing with. Obviously the above stats are the same when they are both on the floor together and unchanged when both are off the floor together. So what needs to be looked at is who are they playing with the most when either one is on the floor without the other that would turn the stats in Powe's favor.

#   Unit                                                                     Min      Off         Def        +/-      W      L    Win%
 1 Rondo-R.Allen-Pierce-Garnett-Perkins        813     1.13     0.99     +216   26      10    72.2
 2 House-T.Allen-Pierce-Powe-Davis               114     1.06     0.97     +23      13      10   56.5
 3 House-T.Allen-R.Allen-Powe-Davis                82     1.09     0.96     +18      10      10   50.0
 4 Rondo-R.Allen-Pierce-Garnett-Davis             59     1.11     1.00     +13        9        7    56.2
 5 Rondo-R.Allen-Pierce-Powe-Garnett             57     1.17     0.80     +40        7        2    77.7
 6 Rondo-T.Allen-R.Allen-Garnett-Perkins         44     1.10     1.12       -3         8        8    50.0
 7 House-R.Allen-Pierce-Garnett-Perkins          43     1.11     1.01      +9        8         9    47.0
 8 Rondo-R.Allen-Pierce-Garnett-Scalabrine    41     1.25     1.09     +15       3         0    100
 9 House-T.Allen-Powe-Scalabrine-Davis         40     1.09     1.15       -4         4        4    50.0
10 House-Pruitt-T.Allen-Powe-Davis                  26     0.91     0.96      -1          3        2    60.0
11 Rondo-R.Allen-Pierce-Davis-Perkins         20     0.96     1.32      -14        3        6    33.3
12 House-R.Allen-Pierce-Powe-Davis              19     0.71     1.07      -12         1        3   25.0
13 Rondo-R.Allen-Pierce-Scalabrine-Perkins 19     1.23     0.88      +12        2        0   100
14 House-Pruitt-T.Allen-Scalabrine-O'Bryant    18     0.77     0.83       -1         3        1    75.0
15 House-Pruitt-Pierce-Powe-Davis                  18     1.15     1.25        -9         1        3    25.0
16 Pruitt-T.Allen-Scalabrine-Davis-O'Bryant  17     1.05     1.18        +0        2        2   50.0
17 House-Pruitt-T.Allen-Powe-Scalabrine      17     1.06     1.17        -7          1        3   25.0
18 House-T.Allen-R.Allen-Garnett-Davis          16     1.18     0.93        +9        6        3   66.6
19 Rondo-House-R.Allen-Pierce-Garnett         16     0.76     1.39        -14       5        8   38.4
20 Rondo-T.Allen-Pierce-Garnett-Perkins        15     1.01     1.19        -2         5        6   45.4


So what light does this shed on the subject?

Well first that it appears that of the most common rotations, BBD is used more often and in a larger number of rotations as he is in 5 rotations in the top 20 that don't have Powe while Powe is in only 2 that don't have Davis.

The combos that they are in most often without each other is when each one plays with all the starters. Here Powe has a decided difference which probably accounts for his numbers being better than Baby's. When each substitutes for Perk the breakdown is as follows:


4 Rondo-R.Allen-Pierce-Garnett-Davis             59     1.11     1.00     +13        9        7    56.2
 5 Rondo-R.Allen-Pierce-Powe-Garnett             57     1.17     0.80     +40        7        2    77.7

All the numbers that Roy quoted are significantly in Powe's favor. Does that mean he's a better defender than Baby? I don't think so. I think what it tells us is that he is a better fit for that combination. In the starting rotation Perk, the center,  is the post man and KG, the PF, is the big on the perimeter. On the second team Powe, the PF, is the man in the post and Baby, the C, is the man on the perimeter on the offensive end. When Powe plays with the starters on offense he is substituting for Perkins role, a role he is familiar with because that is what he plays on second team. But when Baby plays with that group, both he and KG are used to playing on the perimeter and hence the offense doesn't work as well. That would account for the discrepancy offensively.

On the defensive side though this combo also favors Powe. When Powe is in the game with that unit of RA, PP, KG, and RR, Powe takes the smaller, less physical of the two opposing big men as KG is better suited to guard the centers. When Baby plays with the same unit, it is Davis that takes the larger, more imposing opposing player. In other words, the tougher assignment physically where he is guarding men often much taller than himself whereas Powe often guards people more his own size.

Other info in the 5 man units data show that when they play together Powe and Baby are actually pretty effective when playing together and with House and TAllen. Familiarity of playing with the same people and playing the same role shows itself here. When the less played combos are on the floor it is pretty easy to see that the unfamiliarity of playing with the same people shows up as 8 of the 11 least used combos all are in in the minus, some in double digits so.

Anyway, I'm not sure that any of the data I supplied sheds any more or less light on who is the better defender. There is data that goes both ways. But when comparing the two defensively we need to note that both play different positions and different roles. So statistical data could be misleading and therefore it is incumbent upon us to observe the two and decide who is the better man defender and who is the better team defender. In both cases I see Baby being a better defender. But not by a ton. 

Very good post Nick, TP. Alot of info which is true in there.

[expobear]

Davis is probably the better defender and rebounder. After hearing all the complaints about Powe missing rotations, being late on switches etc, his defensive weaknesses stem from the fact that he has poor vision on the ball and his man (when his man does not have the ball). Many times, I've noticed Powe tries to find his man on the weak side and he'll lose sight of the ball. I don't think you can rotate well defensively when you're in this position. Rebounding is also difficult when you don't know when the ball is shot because you're trying to locate your man. If the man Powe is guarding is bigger, then it's even more difficult to box him out. I think Davis has been taught better defensive fundamentals throughout his basketball career.

That said, I'm still a Powe fan. Powe can score and still rebound through sheer determination and guts. I think if Powe went to a big man camp or something similar where he can learn a few basics about defense and perfect his jumper, he can be a pretty good player. I go with Powe because he's a winner.  

[Scalablob990]

Davis is probably the better defender and rebounder. Throwing out missed rotations, being late on switches etc, Powe's defensive weakness is based on the fact that he has poor vision on the ball and his man (when his man does not have the ball). Many times, I've noticed Powe tries to find his man on the weak side and he'll lose sight of the ball. I don't think you can rotate well defensively when you're in this position. Rebounding is also difficult when you don't know when the ball is shot because you're trying to locate your man. If the man Powe is guarding is bigger, then it's even more difficult to keep box him out. I think Davis has been taught better defensive throughout his career.

That said, I'm still a Powe fan. Powe can score and still rebound through sheer determination and guts. I think if Powe went to a big man camp or something similar where he can learn a few basics about defense and perfect his jumper, he can be a pretty good player. I go with Powe because he's a winner.  

If Powe came to training camp next year with a mid range jumper, and more smooth on defensive rotations, I think BB will have trouble trying to keep up with him, but I guess the same can be for big baby is he became a better post player.

[nickagneta]

When Baby plays with the same unit, it is Davis that takes the larger, more imposing opposing player. In other words, the tougher assignment physically where he is guarding men often much taller than himself whereas Powe often guards people more his own size.

Also, I don't agree with this logic. If it were advantageous to have Baby guard the bigger player, that would show up in that unit's +/-. If Powe and Davis were interchangeable defensive players, then there should be no difference between Powe/KG and Davis/KG as once on the court each would guard the opponent that best suits them. In other words, if Powe takes the 4 and KG the 5 in the Powe/KG tandem, why wouldn't Davis take the 4 again in the Davis/KG tandem? If Davis/KG defend better with Davis at the 4, then why would they play Davis at the 5 and hurt the overall team numbers? Doesn't make sense.

Except your logic isn't reality.

When KG and Powe are in the game defensively KG takes the center and Powe the PF. When it is Baby and KG KG takes the PF and Baby the center. I could care less about your logic. My eyes know what they see and what they see is what's happening not what is logical in a hypothetical debate.

Right, and my point is that the reason they would have Davis guard the 5 and KG the 4 is because Davis at the 4 and KG at the 5 would be even worse than KG at the 4 with Davis at the 5, which is worse Powe guarding the 4 and KG guarding the 5

Or the logic could be, and I've heard Doc say this, they want KG playing the 5 as little as possible because they are worried about the physical toll that playing 5 will have on KG. Powe just can not play the 5, he's way to small and can't do what is physically necessary there. So Powe has to play the 4 and KG has to play the 5 when they are in together. Baby, however, can play the 5 and can take the hysical beating that that position entails. Therefore they can play KG in the preferred less physically taxing position of PF.

You are reading a logic into the decision making for why each plays the position they do with KG based on some misplaced ideology that it has anything to do with who can play better defense at the 4 when the reality is that each of the three players play the positions they do with each other because of the necessity of the physical realities and the fear of unduly taxing the team's most important player.In the KG/Powe tandem the need for KG to play the 5 outweighs the need to have him play that position as little as posible due to Powe's physical limitations in being able to play the 5. In the KG/Baby tandem the need to physically limit KG outweighs the need for him to play the 5 because Baby can play the five and do what is physically necessary to still have the team succeed.

[Fan from VT]

Or the logic could be, and I've heard Doc say this, they want KG playing the 5 as little as possible because they are worried about the physical toll that playing 5 will have on KG. Powe just can not play the 5, he's way to small and can't do what is physically necessary there. So Powe has to play the 4 and KG has to play the 5 when they are in together. Baby, however, can play the 5 and can take the hysical beating that that position entails. Therefore they can play KG in the preferred less physically taxing position of PF.

You are reading a logic into the decision making for why each plays the position they do with KG based on some misplaced ideology that it has anything to do with who can play better defense at the 4 when the reality is that each of the three players play the positions they do with each other because of the necessity of the physical realities and the fear of unduly taxing the team's most important player.In the KG/Powe tandem the need for KG to play the 5 outweighs the need to have him play that position as little as posible due to Powe's physical limitations in being able to play the 5. In the KG/Baby tandem the need to physically limit KG outweighs the need for him to play the 5 because Baby can play the five and do what is physically necessary to still have the team succeed.

Yes, and I understand this, and as I've said before, this is not Davis's fault and just shows that we really need a real backup center; Davis is just a stop-gap at that position and Powe clearly outclasses him at the 4. I just can't get around the fact that if Davis were as good at defense as some say they observe, it would show up somewhere in the actual results of games, but it doesn't. He's a mediocre 4 forced to play out of position.

[nickagneta]

Or the logic could be, and I've heard Doc say this, they want KG playing the 5 as little as possible because they are worried about the physical toll that playing 5 will have on KG. Powe just can not play the 5, he's way to small and can't do what is physically necessary there. So Powe has to play the 4 and KG has to play the 5 when they are in together. Baby, however, can play the 5 and can take the hysical beating that that position entails. Therefore they can play KG in the preferred less physically taxing position of PF.

You are reading a logic into the decision making for why each plays the position they do with KG based on some misplaced ideology that it has anything to do with who can play better defense at the 4 when the reality is that each of the three players play the positions they do with each other because of the necessity of the physical realities and the fear of unduly taxing the team's most important player.In the KG/Powe tandem the need for KG to play the 5 outweighs the need to have him play that position as little as posible due to Powe's physical limitations in being able to play the 5. In the KG/Baby tandem the need to physically limit KG outweighs the need for him to play the 5 because Baby can play the five and do what is physically necessary to still have the team succeed.

Yes, and I understand this, and as I've said before, this is not Davis's fault and just shows that we really need a real backup center; Davis is just a stop-gap at that position and Powe clearly outclasses him at the 4. I just can't get around the fact that if Davis were as good at defense as some say they observe, it would show up somewhere in the actual results of games, but it doesn't. He's a mediocre 4 forced to play out of position.

Well it shows up here:

Opponent's production (primary position):

Powe: .440 eFG%, 19.4 pts, 10.3 reb, 15.5 PER

BBD: .497 eFG%, 15.7 pts, 11.2 reb, 13.4 PER

BBD holds his opposition to 3.7 less points per game and holds his opposition to over 2 points less in over efficiency as BBD holds his opposition to slightly over the league average of a 14 PER while Powe's opposition's efficiency is 15.5 PER. 2 poins in PER isn't a small amount.

So stats show that BBD's opposition is held to less points than Powe's direct competition and plays less efficiently overall than Powe's direct opposition.

So why aren't these two very significant numbers being avoided in the argument? They do kinda prove that BBD is better than Powe defensively as opposed to the other stats where cohesion amongst over players is what is being measured and where how other players perform on his own team effects Powe's and BBD's stats.

[Fan from VT]

Or the logic could be, and I've heard Doc say this, they want KG playing the 5 as little as possible because they are worried about the physical toll that playing 5 will have on KG. Powe just can not play the 5, he's way to small and can't do what is physically necessary there. So Powe has to play the 4 and KG has to play the 5 when they are in together. Baby, however, can play the 5 and can take the hysical beating that that position entails. Therefore they can play KG in the preferred less physically taxing position of PF.

You are reading a logic into the decision making for why each plays the position they do with KG based on some misplaced ideology that it has anything to do with who can play better defense at the 4 when the reality is that each of the three players play the positions they do with each other because of the necessity of the physical realities and the fear of unduly taxing the team's most important player.In the KG/Powe tandem the need for KG to play the 5 outweighs the need to have him play that position as little as posible due to Powe's physical limitations in being able to play the 5. In the KG/Baby tandem the need to physically limit KG outweighs the need for him to play the 5 because Baby can play the five and do what is physically necessary to still have the team succeed.

Yes, and I understand this, and as I've said before, this is not Davis's fault and just shows that we really need a real backup center; Davis is just a stop-gap at that position and Powe clearly outclasses him at the 4. I just can't get around the fact that if Davis were as good at defense as some say they observe, it would show up somewhere in the actual results of games, but it doesn't. He's a mediocre 4 forced to play out of position.

Well it shows up here:

Opponent's production (primary position):

Powe: .440 eFG%, 19.4 pts, 10.3 reb, 15.5 PER

BBD: .497 eFG%, 15.7 pts, 11.2 reb, 13.4 PER

BBD holds his opposition to 3.7 less points per game and holds his opposition to over 2 points less in over efficiency as BBD holds his opposition to slightly over the league average of a 14 PER while Powe's opposition's efficiency is 15.5 PER. 2 poins in PER isn't a small amount.

So stats show that BBD's opposition is held to less points than Powe's direct competition and plays less efficiently overall than Powe's direct opposition.

So why aren't these two very significant numbers being avoided in the argument? They do kinda prove that BBD is better than Powe defensively as opposed to the other stats where cohesion amongst over players is what is being measured and where how other players perform on his own team effects Powe's and BBD's stats.

That's a good find, and I acknowledge that. However, couldn't one argue that a team's overall defense is more important? Additionally, those individual defensive PER's are only half the story; they go hand in hand with the offensive PER:
Powe is at 18.8 for a +3.3 at his position, while Davis is at 9.5 for a -3.9.

In addition, you just looked at Powe at PF and Davis at Center. Overall, Powe's total defensive efficiency for all minutes played is 16.6 and Davis's is at 18.2 if you count both his PF and C minutes. Overall, then, for the season Powe is at +2 vs. his opponent and Davis is at -9.5

[cordobes]

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. You are trying to read too much in a merely descriptive metric. 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. My guess is it'd be worse.

For example, just last season these same numbers were inverted. What's the theory here? That Davis was the superior defender last season but Powe is the better one this season? Why is that?

I'd recommend the reading of Dean Oliver's book. He talks about the dangers of using Defensive Rating (basically the same measurement you're trying to use here, points allowed per 100 possessions when the player is on the floor - he was the one who come up with it) as a measurement of a player's defensive quality.

Others have provided explanations for why the numbers are what they are. IMO, it's very simple: Davis is the more versatile defender, with Powe being a much bigger liability against a wide number of matchups. In some situations where Davis is subpar, Powe would be even worse. This, as you certainly understand, skews the numbers on Powe's favor. The sad truth is that there are no metrics that are remotely useful as ways of rating the defense of a basketball player. Counterparts production (like opponent PER) is an even worse idea: this is basketball, not tennis.

I think Davis is a vastly superior post defender. He's also slightly better on the pick'n'roll and generally shows more awareness and concentration on the defensive side.

[nickagneta]

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. You are trying to read too much in a merely descriptive metric. 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. My guess is it'd be worse.

For example, just last season these same numbers were inverted. What's the theory here? That Davis was the superior defender last season but Powe is the better one this season? Why is that?

I'd recommend the reading of Dean Oliver's book. He talks about the dangers of using Defensive Rating (basically the same measurement you're trying to use here, points allowed per 100 possessions when the player is on the floor - he was the one who come up with it) as a measurement of a player's defensive quality.

Others have provided explanations for why the numbers are what they are. IMO, it's very simple: Davis is the more versatile defender, with Powe being a much bigger liability against a wide number of matchups. In some situations where Davis is subpar, Powe would be even worse. This, as you certainly understand, skews the numbers on Powe's favor. The sad truth is that there are no metrics that are remotely useful as ways of rating the defense of a basketball player. Counterparts production (like opponent PER) is an even worse idea: this is basketball, not tennis.

I think Davis is a vastly superior post defender. He's also slightly better on the pick'n'roll and generally shows more awareness and concentration on the defensive side.

I also think he has better lateral movement and footwork than Powe does when out of the post and that he moves players out of the low block better when it comes to rebounding. Proper footwork and learing space are very under rated skills that just don't seem to get noticed but that can make the difference between a mediocre player and a good player and a good player and a great player.

Powe definitely has better explosion and elevation than BBD but Baby's footwork and lteral speed allow him to be more versatile than Powe.

[Fan from VT]

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.

[Fan from VT]

 The problem is that you don't know how would the team perform if it was Powe playing those minutes, not Davis. My guess is it'd be worse.

Again, we have little to work with, but look at the one scenario where one's minutes are replaced by the others:

Rondo-R.Allen-Pierce-Garnett-Davis:
Min.    Off.     Def.   +/-     
59     1.11     1.00     +13      

Rondo-R.Allen-Pierce-Powe-Garnett    
Min.    Off.     Def.   +/-
57     1.17     0.80     +40      

[cordobes]

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.

[Fan from VT]

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.