RealGM Top 100 List #41

[tsherkin]

Well im happy you felt the need to respond with an essay, over a point that shouldve been understood without the need to over analyze it.

Right, then for your benefit:

"Please contribute in a meaningful fashion or feel welcome to stop posting in this project."

Cheers.

[Doctor MJ]

If I recall correctly...

What issue

He's telling you. "RAPM is a process that makes it so that 1 "point" of value in its results is no longer the same thing as 1 point on the scoreboard".

Please elaborate - why do you think 1 APM point is equal to 1 scoreboard point and 1 RAPM isn't equal to 1 socreboard point?

This is correct. The machine learning component of RAPM does not conserve that relationship. When Joe Sill invented it, he actually went around asking people what they thought reasonable numbers should look like with idea of perhaps scaling the numbers based on that.

I'm not aware of anyone responding to him in a way that established any standard. If they did, it clearly didn't take hold to the point where Engelmann or others included that scaling in their methods. So this was the impetus for my normalized and then scaled spreadsheet along with the belief that the data could be presented in a more "player readable" fashion (meaning in one glance you could see a player's chronological snapshot).

[Doctor MJ]

With Miler in I'm back in an open place again, which is nice. I'm starting out by reviewing some list.

Here are the guys in the Pre-List Top 50 not in yet:

37. Paul Pierce
39. Kevin McHale

(Gervin and Mikan were the non-Top 40ers who made it in already)

42. Dikembe Mutombo
43. Willis Reed
45. Dave Cowens
(tie). Tracy McGrady
47. Elvin hayes
48. Dwight Howard
49. Alonzo Mourning
50. Ray Allen

From the 2011 list:

37. Tracy McGrady
38. Paul Pierce
39. Dwight Howard
40. Kevin McHale

(Mikan, Chris Paul, Kevin Durant, and Reggie Miller were the "new" Top 40ers)

41. Dominique Wilkins
43. Willis Reed
44. Bob Cousy
45. Dave Cowens
46. Allen Iverson
47. Bill Walton
48. Alonzo Mourning
49. Elvin Hayes
50. Chris Paul

By RPOY shares:

24. Dolph Schayes
29. Bob McAdoo
30. Bill Walton
33. Bob Cousy
34. Dwight Howard
35. Neil Johnston
37. Paul Arizin
38. Alonzo Mourning
40. Willis Reed

(MIkan, Gilmore, Havlicek, Drexler, Pippen, Isiah, Kidd, Stockton, and Reggie were the novel Top 40ers.)

[D Nice]

Spoiler:

(Modified post from last thread, removing Reggie Miller and focusing on the bigs vs Pierce)

Alonzo Mourning, Dikembe Mutombo, Rasheed Wallace and/or Dwight Howard vs Pierce

Right now, I'm looking more at the dominant defensive bigs of the most recent eras, particular Zo Mourning, Mutombo and Howard (actually, I guess Sheed may just fit into this group as well, so I won't rule him out as a candidate either).

The player getting the most traction thus far in this thread has been Pierce. He's mainly here for his offensive contributions, though Pierce's well-rounded game does include solid defense on the wing as well. But without going into heavy analysis, I feel like the dominant defensive big is a more valuable commodity than a good scoring wing of similar caliber. The big is the more limited resource, he can change the way the game is played, and you could build a strong unit around them without having to bring in other All-Star caliber defenders. No one can make a defensive unit elite by themselves, but if you surround Mutombo or Mourning with 4 reasonable defenders/defensive role players in a good defensive system you should get one of the better defenses in the league. On the other hand, even with a good system, if Pierce is your offensive anchor but your other four offensive players are role players/reasonable offensive types it's unlikely that you've got one of the best offenses in the league.

That was a rambling way to say that. Let me try again. The dominant defensive bigs can be your franchise centerpiece on that end of the floor, and this group can also contribute to some extent on offense as well (varying from offensive neutral with some valuable attributes to semi-offensive force). Pierce isn't really the franchise centerpiece type for a good team at either end...he's more of an excellent offensive lieutenant that can also contribute to the defense. Which is a very valuable thing, but not as valuable to my eyes as a true franchise, dominant defensive big.

Estimating impact
If we wanted to put numbers to it, we'd have to use RAPM as much as possible because the box score stats just don't do relative defensive impact that well. That's another reason that I'm focusing here on players that at least overlapped some of their prime with the databall era, so that we can try to quantify their relative impacts a bit. Once again, using Doc MJ's normalized PI RAPM spreadsheet form 1998 - 2012 (with disclaimers as mentioned before):

3-year peaks (best 3 1-year scores from 1998 - 2012, averaged together):

Mourning: +8.8 *
Rasheed Wallace: +8.2
Howard: +6.7 *
Mutombo: +6.7 *
Pierce: +6.3

(*I put asterisks next to Mourning and Mutombo because the 1998 cut-off missed big chunks of their primes so their best numbers could be even better. Similarly, Howard's best 3 scores come from 2010 - 2012, but 2009 was very arguably his best season and there's a famous calculation conundrum with calculating RAPM for the 2009 Magic (e.g. that Howard and Gortat played all available center minutes with 0 overlap between them) so I'd argue that it's very possible that Howard's 2009 RAPM score as listed is too low)

There's a clear difference here between the measured RAPM peaks of the best dominant defensive bigs than Pierce. To me, this helps quantify some of the qualitative descriptions that I rambled through above. Pierce was a very high impact players, but the bigs had more.

Let's also look at the offensive and defensive splits. Because say, for example, you have a player that's dominant on one end but weak on the other end. With the right coaching you could possibly embellish their strengths and hide their weaknesses enough to make them more valuable than a more rounded player (think Phoenix version of Steve Nash, or perhaps Allen Iverson). Also, keep in mind that the offensive and defensive 3-year peaks may not necessarily be the same years so combining the following 2 numbers won't necessarily result in the numbers above:

Offense 3-year RAPM peaks:
Pierce: +4.6
Mourning: +3.7 (+3.3 with 1998 correction)
Sheed: +3.5
Howard: +3.3
Mutombo: +0.0

Defense 3-year RAPM peaks:
Mutombo: +8.7 (7.9 w/ '98 correction)
Mourning: +6.0 (+5.5 w/ '98 correction)
Sheed: +5.7
Howard: +4.7
Pierce: +3.4

Among this group, Mutombo stands out as the most one-sided that a smart coach could build a unit around (to differentiate from just looking at the pure overall RAPM scores). Mutombo's defensive impact looks larger than Pierce's combined impact. And among the more 2-way players, Mourning and Sheed's defensive impact was larger than Pierce's offensive, and interestingly their offense was also a bit better than Pierce's defense.

Anyway, the point isn't to scrape the decimal points to look for differences because a) RAPM isn't that precise and b) we know of an issue with the 1998 values that could change some of these estimates by some tenths of points. But the point is to look on a more macro scale, and on that scale it seems to me that this group of bigs tends to have overall more impact and more specialized impact than the two wings, making them more valuable commodities.

Now, the longevity point is out there on the table Pierce has great longevity. Sheed and Mutombo have pretty strong longevity themselves, but by dint of health (Zo) and age (Howard) both of them have about a decade of longevity to work with. Everyone's mileage can (and has) varied widely on what to do with longevity. But for me, I don't think Pierce's longevity advantages is enough for them to shoot to the top of this particular grouping. Thus, my vote in this thread likely comes from among the 4 bigs I discussed here.

Re: Superior Estimated Impact

Still not seeing any talk about the substitution bias. If the case against Pierce for these guys with clearly inferior longevity/durability (save Deke) is going to be based on their fantastic D-RAPM scores I’d like it at least addressed. Below is some data on the 5 top Non-Shane Battier Wings vs 5 Solid-but-not-exceptional Defensive Bigs from the same era. It should also be noted that amongst the top 60 data-points for "top individual D-RAPM scores" only 3 of the players were perimeter players, and of the top 100 only 9 were perimeter players, and 4 of those 9 data points belong to Shane Battier. Also all of the data below is based on at least 25mpg of play, typically more.

Spoiler:

Prime Ron Artest (03-04, 06-07): 0.7 (1.7N), 1.5 (1.9N), 3.5 (4.0N), 2.3
Prime Luol Deng (07-11): 2.2, -0.1, 1.7, 1.9, 3.9
Prime Bruce Bowen (02-07): 0.6 (1.6N), 0.9 (2.6N), 2.0 (2.3N), 1.2 (1.5N), -0.3 (1.5N), 1.3
Prime Andrei Kirilenko (04-06, 08): 2.9 (1.2N), 2.2 (1.9N), 2.3 (1.4N), 1.1
Prime Andre Iguodala (08-12): 1.6, 2.8, 0.4, 2.5, 1.9

Vlade Divac (02-04): 2.3 (4.3N), 3.3 (2.9N), 2.0 (0.9N)
Shawn Bradley (01, 03): (5.5N), 1.6 (2.9N)
Jason Collins (03-07): 1.9 (3.6N), 2.4 (4.7N), 6.3 (5.8N), 3.4 (2.9N), 2.7
Greg Ostertag (03-04): 3.0 (4.5N), 3.1(1.9N)
Nene Hilario (04-05): 3.2 (3.8N), 3.3 (2.4N)

Note: For Bradley I couldn’t get access to most of his prime data (97-01) so I just used 01 & 03, though 03 breaks my 24mpg rule he played 21 minutes and its likely he would see a similar statistical footprint to his other campaigns if data were available, if anything he might even look a little better. He was 2.4+ the years his minutes were scaled back that we have data for and his Scaled D-RAPM according to Doc’s sheet was 5.3/5.2/6.3 from 1998-2000.

Is it not a reasonable conclusion that there is a clear positional bias going on here that a lot more with court positioning than player ability? Can we acknowledge that while it’s very possible for Big men to be the best defenders in the league, D-RAPM clearly overstates this quantitative impact in comparison to smalls? The argument that “well the bigs could just have a larger talent separation from their crop than wings do theirs” is at least somewhat undercut by the likes of the 5 bigs above trouncing the top 5 (non-Battier) perimeter defenders from their era.

And what adjustments, if any, need to be made when evaluating D-RAPM at face value here? Offense isn’t position-biased, it’s ability-biased, and since it is ability we are trying to measure/estimate/predict, I don’t think O-RAPM requires the same level of re-thinking before application D-RAPM does. Defense is clearly a function of where you are playing, if players are presented with this opportunity and others are not based on the position they play, does this inherently make them more talented/better/more effective players?

Or is Divac/Bradley/Collins/Ostertag/Nene > Prime Artest/Deng/Bowen/AK-47/Iggy an accurate representation of their defensive efficacy/impact?

Re: These guys can anchor mediocrity to higher levels than PP

The fact that these guys can take a team of “4 solid defenders” and make them elite but PP can’t take 4 solid offensive players and make elite them isn’t a fair “problem setup” because defense is completely different than offense. There’s no way of transferring primacy on defense, and a singular bad defender trays down team efficacy more on that end than a poor offensive player does. It’s why a guy like T-Mac or Kobe, can drag a team with 2+ awful offensive starters to a top 10 ORTG but if you put 1-2 awful defenders on the floor next to a great defensive big they generally won’t produce great results (Dwight with Hedo/Shard being the exception if you actually find them to be awful defenders).

This is how an ATG anchor can have defenses in the bottom 3rd of the league (because of Mugsy Bogues, Dell Curry, and to a lesser extent LJ not being great defenders). In Miami he had very solid defensive units and consistently anchored top 8 defenses. You see the same thing with KG in Minny vs. Boston. Cleveland had no real defensive anchor between 07-10. Their best defensive player was Anderson Varejao. And yet they had DRTGs ranked 4th, 11th, 3rd, and 7th over that span. Why? Because they didn’t put weak individual defensive players and they were coached defense. Team orientation and philosophy make a huge difference when evaluating team performance there.

I do think there can be an argument made for Zo/Deke/Dwight scaling better than PP. Pierce might be supremely portable but given the talent of the 08-12 Celtics they produced very underwhelming offenses. They underperformed their talent IMO, a team with Pierce/Allen/KG should have been more impressive on that end than they were. Then again, offense and defense are at least interconnected to some degree and Elgee did some cursory analysis here that shows the team’s efficacy on that end might have been somewhat stifled by team philosophy (re: offensive rebounding). See below.

Spoiler:

My attempt to "normalize" their rebounding -- a method that involved looking for a "team tendency" of OREB strategy by accounting for individual outliers -- gave the Celtics about 2 points of ORtg points. The 2008 Celtics were +2.7 and the 2009 team +2.2.

Would you mind explaining in greater depth how you did this? I'm highly curious as to the methodology applied here.

Well, I never published it so I don't exactly consider it rigorous.IIRC, the general approach was:

-look at OREB% by "position" on a team (PG, wings, bigs)
-remove the top OREB% at each team for position (intended to remove individual outliers)
-take the average per position for that team
-do this for every team
-this creates a ranking of all 30 teams of "OREB strategy." (Confounded slightly by teams who are individually filled with great O rebounders.)
-calculate the number of extra possessions a team is generating by their "OREB strategy" relative to league average (for a team who doesn't hit the offensive glass, this will be negative)
-add (or subtract) the expected value of those extra possessions to a team's ORtg

So in other words, I was trying to isolate a team's OREB strategy independent of individually brilliant rebounders like Reggie Evans, normalize it and then create an "adjusted" offensive and defensive rating (because the strategy is a tradeoff between offensive and defense). The results were that teams that hit the glass hard -- Dallas in 2012 -- added about 2.5 points of offensive, and teams that got back on D -- Boston and Chicago -- lost about 2.5 to almost 3 points off ORTG.

If you're wondering why not just look at some combination of TOV% and TS%, it's because both eFG% and free throw rate (FTR) could be impacted by OREB strategy. And of course, some players are just good offensive rebounders which is how they impart value individually. While an eFG% and TOV% only ORtg approximates what I did well (r=0.88) the misses can be significant (e.g. Memphis would be a +0.5 offense with the OREB adjustment, but a -4.8 offense based just on eFG% and TOV%).

There are a few ramifications of all this when it comes to OREB:

(1) If we normalize the playing field by isolating a crash/get-back strategy, teams who crash the offensive glass hard aren't quite as "good" as ORTG paints them out to be, and teams who always get back on defense aren't quite as good as DRTG makes them out to be. The degree of this seems to be about 2-3 points at the edges.

(2) It's important to keep OREB% in mind in general when talking about team-building and offensive efficacy. It's one of the reasons I value Rodman and Chandler (and others) as offensive players -- because individually skilled offensive rebounders contribute in a super-portable way: the skill is basically never redundant with the scorers (or the first wave of attack) and then perhaps becomes more valuable as the offense improves because scrambling defenses are more vulnerable to offensive rebounds.

Thus, when considering skillsets, portability, team-building, etc. it's important to understand where we are on the scale independent of a team's transition defense strategy when evaluating what kind of offenses players have been the "1st option" on, "2nd option," etc.

And honestly in his calculation he used team ORTG as the EV-estimation parameter, I think if he replaced that with [team 2nd chance points/team 2nd chances] that normalized score would look even better, because put-backs are some of the highest-rate PPS plays in basketball. However, again, Pierce has more top 10 seasons than any of these guys (comfortably) and a better post-prime than all of them.

Unaddressed: Longevity/Durability

Pierce was a top 10 player from '01-'06, '08-'11. That's 11 seasons of elite play. Even if you were to be conservative a couple times it comes out to something like 9 top 10 campaigns and 2 top 13 campaigns. Then he gives you 2 elite supporting star campaigns in '12 & '13 (though he missed 21 games in '12).

Being generous with Zo he's got 7 top 10 campaigns, '93-'97, '00, '02, but bear in mind he missed 17 games in '97 and 22 in '94. That's a lot. He was a solid backup (GOAT swatter) in '06 & '07, but they don't hold a candle to Pierce in '12 & '13.

Dwight has 5 top-10 level seasons ('08-'11, '14) but he definitely peaked higher, with 3 of the se campaigns being a legit top 5 player and MVP candidate ('09/'10/'11). I think that bears some mention. He was all-star good in '07, almost as good as his '07 self in '13, and missed 28 games in '12, so we'll say 2.5 years of very good #2-option caliber play and 5 franchise-caliber years with 3 of them on a level Pierce never reached. I think the higher peak for 3 straight seasons makes it debatable, but again, give me Pierce's 11.5+2 seasons (if '12 for Dwight counts then so does '07 for Pierce) over Dwight's 5+2.5. If he can stay '14 good for 2-3 more years I'd revisit the comparison.

Rasheed: Sheed is tricky because his numbers don't support him ever being a top 10 player outside of his '00-'02 stretch, and that's being generous in 2000. Maybe you can extend that to '03 but that team severely underperformed their talent IMO. Still, I'll credit him for 2003, which would give him 4 top-10 player campaigns. Outside of that he was a solid #2 level player for 5 other seasons ('97-'98, '04-'06) and a solid #3/#4 for another in '07 & '08. That's 11 years of relevance, but it doesn't nearly match up with the sustained level of excellence PP has. Even counting '04/'05 as top-10 level years, that's 6 franchise caliber seasons vs. 11 for Paul. And I think saying Sheed was a franchise player for 6 years is over-the-top generous.

Deke has the best argument IMO, I can totally see Deke over PP. I've got him as the #3 or #4 Defensive GOAT after Russell, Hakeem, and maybe Big Ben. He was a durable, top 10-12 caliber player from his rookie year to 2000, which gives him 9 franchise seasons. He was than an elite #2 for 2 more years ('01 & '02), I'd probably say '01/'02 Deke might actually be slightly more valuable than '12/'13 PP. So it becomes a question of whether '92-'00 Deke is as valuable as '01-'11 Pierce (minus '07). I think there's a good argument to be made there but I side with PP if for no other reason than the 2 extra seasons.

[Doctor MJ]

Okay so these are the guys on my mind now:

Pierce - the big, non-big in my mind right now. A fundamentally solid candidacy relative to the other guys.

Mutombo - honestly, the highest guy on my pre-project list, and my screw up in my spreadsheet has an effect on me here. Frankly that makes me wonder whether I was doing things right before, but it is what it is. I was giving him a bit of a boost based on my perception of his portability, but at this point relative to Pierce I can't argue I think he had clearly more impact and he's got a weaker longevity.

Mourning - oddly the spreadsheet error doesn't have much effect here. His peak number was so huge before that it just fell in a the "HUGE" category, and he's still there. So whereas before I had Mutombo ahead based on his longevity, I might have Zo ahead now.

Pierce vs Mourning is just a clear cut peak vs longevity thing.

Howard - Thinking a lot on Zo vs Dwight, and I just keep coming down on Zo's side. I think Zo had more impact at his best, and I can't seriously give Dwight a longevity edge despite the extra minutes. It's not a big gap, and Dwight's negative off-court impact eats that up pretty easily for me.

btw, Reed and Cowens are somewhat similar to Howard here. Not bad attitudes, but they don't have a serious longevity edge over Zo, and I look at Zo's 2-way impact with his clear cut athletic big's build, and I can't honestly say I'd draft those two over him.

Finally McHale - Am I sleeping on him? One can easily make the case for him over Zo, but I find myself skeptical. If I'm won over, then it's pretty easy to see giving him the nod over Pierce too.

So Pierce, Zo, and McHale are I think my most likely votes for me.

[ronnymac2]

Vote: Tracy McGrady

Very much open to changing this, but I'll make T-Mac my official vote for now. Perhaps the best peak left. He's got a 5-6 year stretch comparable to Kobe Bryant and nearly as good of a peak. He's got other useful seasons as well. Just a dynamic wing with a desirable skillset.

He's on the Kobe/Wade level as far as playmaking off the pick-n-roll, he can take on incredible volume/USG%, he can shoot coming off screens, he's got range, and he can play legit Point Forward. Solid defensive player, especially early in his career. 2005 was probably his defensive peak and second best season overall.

Strongly considering Alonzo Mourning here, too.

Spoiler:

Bigs: Dikembe Mutombo, Nate Thurmond, Alonzo Mourning, Dwight Howard, Willis Reed, Dave Cowens, Ben Wallace, Bob Lanier, Bob McAdoo, Kevin McHale, Robert Parish

Worms: Dennis Rodman

Wings: Paul Pierce, Vince Carter, Reggie Miller, Allen Iverson, Ray Allen, Tracy McGrady, Paul Arizen, Alex English, Dominique Wilkins, Penny Hardaway, Manu Ginobili, Sidney Moncrief

Point Guards: Nate Archibald, Kevin Johnson, Chauncey Billups, Deron Williams, Mark Price

[Moonbeam]

I'm going to get a quick vote in while I can, but I may change it.

Guys on my shortlist at this point are Paul Pierce, Dwight Howard, Alonzo Mourning, Dolph Schayes and Sam Jones (I'd like to see more about the latter two in particular). Of the other guys who have been mentioned, Mutombo is one I'd love to get behind, as he's my 2nd favorite player ever, but I felt that his amazing defensive ability, while amazing, was not able to sufficiently slow the truly elite centers of the day in the playoffs. McGrady and Walton are tempting, but I feel that their incredible peaks are overtaken by the longer sustained excellence of the players on my shortlist. All three will be on my radar soon, along with McHale, English/Dantley/Wilkins, Moncrief, McAdoo, Hayes, Unseld, Gasol and a few others. I think it's too early for Rasheed and Ginobili, but I'm intrigued.

For the time being, I'll cast my vote for Paul Pierce. I picked him over Isiah in the runoff and would have strongly considered picking him over Reggie as well (very tough call, that). His all-around offensive game has been a big boon to his teams, even during periods where they were quite talent-poor, generally putting up healthy scoring totals on league efficiency usually 4% above league average, in addition to good playmaking, spacing and rebounding. His defense is by no means as noteworthy (and clearly not in comparison with the likes of Mourning and Howard, let alone Mutombo), but I never felt he was letting down his teams there. Add to that some great playoff performances (even well into his 30s), and he seems like a worthy candidate. Dwight peaked higher, for sure, but his injury issues and off-court distractions have limited his impact of late. One more season playing at the level he did in the 2014 playoffs would be enough to have him come out ahead. Mourning is probably the most tempting candidate other than Pierce for me, with his good offense coupled with great defense making him incredibly valuable in both Charlotte and Miami. My main hesitation is due to his durability, sadly, as those great first 8 years amounted to just over 21,000 total minutes.

[drza]

Spoiler:

(Modified post from last thread, removing Reggie Miller and focusing on the bigs vs Pierce)

Alonzo Mourning, Dikembe Mutombo, Rasheed Wallace and/or Dwight Howard vs Pierce

Right now, I'm looking more at the dominant defensive bigs of the most recent eras, particular Zo Mourning, Mutombo and Howard (actually, I guess Sheed may just fit into this group as well, so I won't rule him out as a candidate either).

The player getting the most traction thus far in this thread has been Pierce. He's mainly here for his offensive contributions, though Pierce's well-rounded game does include solid defense on the wing as well. But without going into heavy analysis, I feel like the dominant defensive big is a more valuable commodity than a good scoring wing of similar caliber. The big is the more limited resource, he can change the way the game is played, and you could build a strong unit around them without having to bring in other All-Star caliber defenders. No one can make a defensive unit elite by themselves, but if you surround Mutombo or Mourning with 4 reasonable defenders/defensive role players in a good defensive system you should get one of the better defenses in the league. On the other hand, even with a good system, if Pierce is your offensive anchor but your other four offensive players are role players/reasonable offensive types it's unlikely that you've got one of the best offenses in the league.

That was a rambling way to say that. Let me try again. The dominant defensive bigs can be your franchise centerpiece on that end of the floor, and this group can also contribute to some extent on offense as well (varying from offensive neutral with some valuable attributes to semi-offensive force). Pierce isn't really the franchise centerpiece type for a good team at either end...he's more of an excellent offensive lieutenant that can also contribute to the defense. Which is a very valuable thing, but not as valuable to my eyes as a true franchise, dominant defensive big.

Estimating impact
If we wanted to put numbers to it, we'd have to use RAPM as much as possible because the box score stats just don't do relative defensive impact that well. That's another reason that I'm focusing here on players that at least overlapped some of their prime with the databall era, so that we can try to quantify their relative impacts a bit. Once again, using Doc MJ's normalized PI RAPM spreadsheet form 1998 - 2012 (with disclaimers as mentioned before):

3-year peaks (best 3 1-year scores from 1998 - 2012, averaged together):

Mourning: +8.8 *
Rasheed Wallace: +8.2
Howard: +6.7 *
Mutombo: +6.7 *
Pierce: +6.3

(*I put asterisks next to Mourning and Mutombo because the 1998 cut-off missed big chunks of their primes so their best numbers could be even better. Similarly, Howard's best 3 scores come from 2010 - 2012, but 2009 was very arguably his best season and there's a famous calculation conundrum with calculating RAPM for the 2009 Magic (e.g. that Howard and Gortat played all available center minutes with 0 overlap between them) so I'd argue that it's very possible that Howard's 2009 RAPM score as listed is too low)

There's a clear difference here between the measured RAPM peaks of the best dominant defensive bigs than Pierce. To me, this helps quantify some of the qualitative descriptions that I rambled through above. Pierce was a very high impact players, but the bigs had more.

Let's also look at the offensive and defensive splits. Because say, for example, you have a player that's dominant on one end but weak on the other end. With the right coaching you could possibly embellish their strengths and hide their weaknesses enough to make them more valuable than a more rounded player (think Phoenix version of Steve Nash, or perhaps Allen Iverson). Also, keep in mind that the offensive and defensive 3-year peaks may not necessarily be the same years so combining the following 2 numbers won't necessarily result in the numbers above:

Offense 3-year RAPM peaks:
Pierce: +4.6
Mourning: +3.7 (+3.3 with 1998 correction)
Sheed: +3.5
Howard: +3.3
Mutombo: +0.0

Defense 3-year RAPM peaks:
Mutombo: +8.7 (7.9 w/ '98 correction)
Mourning: +6.0 (+5.5 w/ '98 correction)
Sheed: +5.7
Howard: +4.7
Pierce: +3.4

Among this group, Mutombo stands out as the most one-sided that a smart coach could build a unit around (to differentiate from just looking at the pure overall RAPM scores). Mutombo's defensive impact looks larger than Pierce's combined impact. And among the more 2-way players, Mourning and Sheed's defensive impact was larger than Pierce's offensive, and interestingly their offense was also a bit better than Pierce's defense.

Anyway, the point isn't to scrape the decimal points to look for differences because a) RAPM isn't that precise and b) we know of an issue with the 1998 values that could change some of these estimates by some tenths of points. But the point is to look on a more macro scale, and on that scale it seems to me that this group of bigs tends to have overall more impact and more specialized impact than the two wings, making them more valuable commodities.

Now, the longevity point is out there on the table Pierce has great longevity. Sheed and Mutombo have pretty strong longevity themselves, but by dint of health (Zo) and age (Howard) both of them have about a decade of longevity to work with. Everyone's mileage can (and has) varied widely on what to do with longevity. But for me, I don't think Pierce's longevity advantages is enough for them to shoot to the top of this particular grouping. Thus, my vote in this thread likely comes from among the 4 bigs I discussed here.

Re: Superior Estimated Impact

Still not seeing any talk about the substitution bias. If the case against Pierce for these guys with clearly inferior longevity/durability (save Deke) is going to be based on their fantastic D-RAPM scores I’d like it at least addressed. Below is some data on the 5 top Non-Shane Battier Wings vs 5 Solid-but-not-exceptional Defensive Bigs from the same era. It should also be noted that amongst the top 60 data-points for "top individual D-RAPM scores" only 3 of the players were perimeter players, and of the top 100 only 9 were perimeter players, and 4 of those 9 data points belong to Shane Battier. Also all of the data below is based on at least 25mpg of play, typically more.

Spoiler:

Prime Ron Artest (03-04, 06-07): 0.7 (1.7N), 1.5 (1.9N), 3.5 (4.0N), 2.3
Prime Luol Deng (07-11): 2.2, -0.1, 1.7, 1.9, 3.9
Prime Bruce Bowen (02-07): 0.6 (1.6N), 0.9 (2.6N), 2.0 (2.3N), 1.2 (1.5N), -0.3 (1.5N), 1.3
Prime Andrei Kirilenko (04-06, 08): 2.9 (1.2N), 2.2 (1.9N), 2.3 (1.4N), 1.1
Prime Andre Iguodala (08-12): 1.6, 2.8, 0.4, 2.5, 1.9

Vlade Divac (02-04): 2.3 (4.3N), 3.3 (2.9N), 2.0 (0.9N)
Shawn Bradley (01, 03): (5.5N), 1.6 (2.9N)
Jason Collins (03-07): 1.9 (3.6N), 2.4 (4.7N), 6.3 (5.8N), 3.4 (2.9N), 2.7
Greg Ostertag (03-04): 3.0 (4.5N), 3.1(1.9N)
Nene Hilario (04-05): 3.2 (3.8N), 3.3 (2.4N)

Note: For Bradley I couldn’t get access to most of his prime data (97-01) so I just used 01 & 03, though 03 breaks my 24mpg rule he played 21 minutes and its likely he would see a similar statistical footprint to his other campaigns if data were available, if anything he might even look a little better. He was 2.4+ the years his minutes were scaled back that we have data for and his Scaled D-RAPM according to Doc’s sheet was 5.3/5.2/6.3 from 1998-2000.

Is it not a reasonable conclusion that there is a clear positional bias going on here that a lot more with court positioning than player ability? Can we acknowledge that while it’s very possible for Big men to be the best defenders in the league, D-RAPM clearly overstates this quantitative impact in comparison to smalls? The argument that “well the bigs could just have a larger talent separation from their crop than wings do theirs” is at least somewhat undercut by the likes of the 5 bigs above trouncing the top 5 (non-Battier) perimeter defenders from their era.

We discussed this a few times a few threads back, and I still don't think you have enough to make your case. To me, from the data that you laid out, the more logical conclusion is that the center/big man defender is the most important defensive slot on the floor and that the RAPM results simply reflect that. Again, I'm not interested in the equivalent of a "pound-for-pound" champion. If Mutombo's size and ability makes it more natural that he occupy the highest impact portions of the defense, and that partially because of that he is able to have a MUCH larger defensive impact than the wing that occupies the same defensive ranking slot as Deke is to centers...then that's just what happens.

There's a reason that height is a sought after commodity in basketball. I don't think you're identifying a bias in Defensive RAPM, I think you're identifying a bias in basketball. Bruce Bowen's defensive rank among wings might be the same as Ben Wallace's defensive rank among big men. But Wallace is able to make a much bigger impact with his defense than Bowen does with his. In fact, Shawn Bradley and Nene may, as well. Barring more evidence of a bias besides an alternate interpretation of the data, I just don't see that your case is that strong.

And what adjustments, if any, need to be made when evaluating D-RAPM at face value here? Offense isn’t position-biased, it’s ability-biased, and since it is ability we are trying to measure/estimate/predict, I don’t think O-RAPM requires the same level of re-thinking before application D-RAPM does. Defense is clearly a function of where you are playing, if players are presented with this opportunity and others are not based on the position they play, does this inherently make them more talented/better/more effective players?

Or is Divac/Bradley/Collins/Ostertag/Nene > Prime Artest/Deng/Bowen/AK-47/Iggy an accurate representation of their defensive efficacy/impact?

I think the underlined is exactly one of the points where we're disagreeing. That's what I was trying to tease out above with my "pound for pound" comment. I'm NOT trying to estimate ability in a vacuum. Chris Paul could very arguably be the most skilled basketball player that ever lived, with more ability for his size than anyone else. But he'll never have the impact that Shaq did, and a lot of that is because Shaq was friggin monstrously huge and that allowed him to dominate the highest efficiency areas of the court. Now, if we're allowed to factor size and the court areas that players occupy as part of "ability", then I revert to what I said above. Big men have more ability to affect defenses than wings or guards...they have the ability to man the highest impact defensive areas BECAUSE they are big enough to deter opponents and get rebounds on an order that smaller players just can't. Again, I don't think you're identifying a bias in RAPM, I think you're identifying a bias in basketball.

Re: These guys can anchor mediocrity to higher levels than PP

The fact that these guys can take a team of “4 solid defenders” and make them elite but PP can’t take 4 solid offensive players and make elite them isn’t a fair “problem setup” because defense is completely different than offense. There’s no way of transferring primacy on defense, and a singular bad defender trays down team efficacy more on that end than a poor offensive player does. It’s why a guy like T-Mac or Kobe, can drag a team with 2+ awful offensive starters to a top 10 ORTG but if you put 1-2 awful defenders on the floor next to a great defensive big they generally won’t produce great results (Dwight with Hedo/Shard being the exception if you actually find them to be awful defenders).

This is how an ATG anchor can have defenses in the bottom 3rd of the league (because of Mugsy Bogues, Dell Curry, and to a lesser extent LJ not being great defenders). In Miami he had very solid defensive units and consistently anchored top 8 defenses. You see the same thing with KG in Minny vs. Boston. Cleveland had no real defensive anchor between 07-10. Their best defensive player was Anderson Varejao. And yet they had DRTGs ranked 4th, 11th, 3rd, and 7th over that span. Why? Because they didn’t put weak individual defensive players and they were coached defense. Team orientation and philosophy make a huge difference when evaluating team performance there.

I do think there can be an argument made for Zo/Deke/Dwight scaling better than PP. Pierce might be supremely portable but given the talent of the 08-12 Celtics they produced very underwhelming offenses. They underperformed their talent IMO, a team with Pierce/Allen/KG should have been more impressive on that end than they were. Then again, offense and defense are at least interconnected to some degree and Elgee did some cursory analysis here that shows the team’s efficacy on that end might have been somewhat stifled by team philosophy (re: offensive rebounding). See below.

Spoiler:

Would you mind explaining in greater depth how you did this? I'm highly curious as to the methodology applied here.

Well, I never published it so I don't exactly consider it rigorous.IIRC, the general approach was:

-look at OREB% by "position" on a team (PG, wings, bigs)
-remove the top OREB% at each team for position (intended to remove individual outliers)
-take the average per position for that team
-do this for every team
-this creates a ranking of all 30 teams of "OREB strategy." (Confounded slightly by teams who are individually filled with great O rebounders.)
-calculate the number of extra possessions a team is generating by their "OREB strategy" relative to league average (for a team who doesn't hit the offensive glass, this will be negative)
-add (or subtract) the expected value of those extra possessions to a team's ORtg

So in other words, I was trying to isolate a team's OREB strategy independent of individually brilliant rebounders like Reggie Evans, normalize it and then create an "adjusted" offensive and defensive rating (because the strategy is a tradeoff between offensive and defense). The results were that teams that hit the glass hard -- Dallas in 2012 -- added about 2.5 points of offensive, and teams that got back on D -- Boston and Chicago -- lost about 2.5 to almost 3 points off ORTG.

If you're wondering why not just look at some combination of TOV% and TS%, it's because both eFG% and free throw rate (FTR) could be impacted by OREB strategy. And of course, some players are just good offensive rebounders which is how they impart value individually. While an eFG% and TOV% only ORtg approximates what I did well (r=0.88) the misses can be significant (e.g. Memphis would be a +0.5 offense with the OREB adjustment, but a -4.8 offense based just on eFG% and TOV%).

There are a few ramifications of all this when it comes to OREB:

(1) If we normalize the playing field by isolating a crash/get-back strategy, teams who crash the offensive glass hard aren't quite as "good" as ORTG paints them out to be, and teams who always get back on defense aren't quite as good as DRTG makes them out to be. The degree of this seems to be about 2-3 points at the edges.

(2) It's important to keep OREB% in mind in general when talking about team-building and offensive efficacy. It's one of the reasons I value Rodman and Chandler (and others) as offensive players -- because individually skilled offensive rebounders contribute in a super-portable way: the skill is basically never redundant with the scorers (or the first wave of attack) and then perhaps becomes more valuable as the offense improves because scrambling defenses are more vulnerable to offensive rebounds.

Thus, when considering skillsets, portability, team-building, etc. it's important to understand where we are on the scale independent of a team's transition defense strategy when evaluating what kind of offenses players have been the "1st option" on, "2nd option," etc.

And honestly in his calculation he used team ORTG as the EV-estimation parameter, I think if he replaced that with [team 2nd chance points/team 2nd chances] that normalized score would look even better, because put-backs are some of the highest-rate PPS plays in basketball. However, again, Pierce has more top 10 seasons than any of these guys (comfortably) and a better post-prime than all of them.

I'll address your last line in the section below. But other than that, I think you have interesting points about the difference between defense and offense. You can only go so far with a direct comparison of defense vs offense because, as you point out, the units are run differently. However, I also think that this is another part of why bigs tend to have (much) larger defensive impacts than their wing counterparts. Using your phrase, you can't transfer primacy on defense. So when a good defensive wing's man doesn't have the ball, he has fewer ways to directly influence the game. However, when a big man's man doesn't have the ball he is still able to help out on opposing penetrators, he's still able to help prevent passes into the highest efficiency areas of the court, and he's still the most likely player on his team to grab a rebound and end a possession. The big man, defensively, can just do much more than the wing.

Now, back to the examples you gave of a) Kobe or TMac dragging poor offensive casts to top-10 results while b) the 09 - 12 Celtics NOT having top-10 results despite lots of offensive talent. For a) I'd say that it matches exactly with the point I was making...I don't think Pierce, in his best days, could do offensively what Kobe or the version of TMac that you're referring to could do. Pierce was a great scorer with a solid all-around game, but Kobe and TMac were better scorers with much more ability to act as lead guards for their teams. It's a similar point to what I made last thread when comparing Pierce and Iverson...Pierce has a good handle/passing ability for a small forward, but that is NOT the same as being able to run a team. That lack of ability to be an offensive general on that level is exactly WHY I don't think that you could build a top offense with him and only offensive role players, regardless of the system. He would NEED additional offensive help, likely from a gifted point guard, to lift the offensive unit to that level.

Which brings us to b) those Celtics...much, much more-so than the offensive rebounding, the culprit behind the underwhelming offenses was Rajon Rondo and the offense running more through him. Part of the reason that Rondo was given the increased primacy was because his own game was too limited to play off the ball and he was the young talent, so the team tried to play to his strengths. But IMO, another reason for Rondo's increased primacy is that Pierce couldn't handle being the lead guard for the Celtics full time, thus requiring that it be placed elsewhere. And note, this isn't necessarily a negative on Pierce's game. He was, in fact, a small forward and so he wouldn't traditionally be EXPECTED to be the offensive engine in this way. But like defensive big men, I think there's a reason that the top of the offensive RAPM lists are dominated by either point guards or TRULY exceptional defense-warping scorers (or both). The point guard ability to control an offense is a huge way to make an offensive impact, and Pierce doesn't have that on the level of either point guards or wings like Kobe/TMac/AI. But at the same time, Pierce also isn't able to destroy defenses with his scoring the way that the other dominant scorers can. It's why Pierce's offensive RAPM is very good, but not killer. And it's why I don't believe that you could put him as the centerpiece of an offense with only offensive role players and have them produce as well as Mutombo or Howard with a team of defensive role players.

Unaddressed: Longevity/Durability

Pierce was a top 10 player from '01-'06, '08-'11. That's 11 seasons of elite play. Even if you were to be conservative a couple times it comes out to something like 9 top 10 campaigns and 2 top 13 campaigns. Then he gives you 2 elite supporting star campaigns in '12 & '13 (though he missed 21 games in '12).

Being generous with Zo he's got 7 top 10 campaigns, '93-'97, '00, '02, but bear in mind he missed 17 games in '97 and 22 in '94. That's a lot. He was a solid backup (GOAT swatter) in '06 & '07, but they don't hold a candle to Pierce in '12 & '13.

Dwight has 5 top-10 level seasons ('08-'11, '14) but he definitely peaked higher, with 3 of the se campaigns being a legit top 5 player and MVP candidate ('09/'10/'11). I think that bears some mention. He was all-star good in '07, almost as good as his '07 self in '13, and missed 28 games in '12, so we'll say 2.5 years of very good #2-option caliber play and 5 franchise-caliber years with 3 of them on a level Pierce never reached. I think the higher peak for 3 straight seasons makes it debatable, but again, give me Pierce's 11.5+2 seasons (if '12 for Dwight counts then so does '07 for Pierce) over Dwight's 5+2.5. If he can stay '14 good for 2-3 more years I'd revisit the comparison.

Rasheed: Sheed is tricky because his numbers don't support him ever being a top 10 player outside of his '00-'02 stretch, and that's being generous in 2000. Maybe you can extend that to '03 but that team severely underperformed their talent IMO. Still, I'll credit him for 2003, which would give him 4 top-10 player campaigns. Outside of that he was a solid #2 level player for 5 other seasons ('97-'98, '04-'06) and a solid #3/#4 for another in '07 & '08. That's 11 years of relevance, but it doesn't nearly match up with the sustained level of excellence PP has. Even counting '04/'05 as top-10 level years, that's 6 franchise caliber seasons vs. 11 for Paul. And I think saying Sheed was a franchise player for 6 years is over-the-top generous.

Deke has the best argument IMO, I can totally see Deke over PP. I've got him as the #3 or #4 Defensive GOAT after Russell, Hakeem, and maybe Big Ben. He was a durable, top 10-12 caliber player from his rookie year to 2000, which gives him 9 franchise seasons. He was than an elite #2 for 2 more years ('01 & '02), I'd probably say '01/'02 Deke might actually be slightly more valuable than '12/'13 PP. So it becomes a question of whether '92-'00 Deke is as valuable as '01-'11 Pierce (minus '07). I think there's a good argument to be made there but I side with PP if for no other reason than the 2 extra seasons.

I pretty strongly disagree that Pierce was a top-10 player anywhere near that number of times. I've got him in the top 10 maybe a handful of times in his career, and I think that's probably an overestimate. Quotatious, one of the biggest Pierce supporters in this project, says in his post on page 1 of this thread: "Pierce was a top 10 player in the league in a few seasons (2002, 2003, 2006, 2008)"...I think that's much closer to correct than your 9+ times, and I'd quibble a bit with Pierce being no-doubt top-10 even in those years (e.g. 2003 I'd say Duncan, KG, Shaq, Dirk, TMac, Kobe, Kidd, Sheed, Big Ben, Webber all have a pretty strong case). So no, I don't believe that Pierce was top-10 anywhere near as often as you state.

HOWEVER, I do agree with your point that Pierce was very, very good for a very long time. Maybe the case could be made that he was top-15 as often as you suggested he was top-10 (haven't really looked that in depth), and that is still extremely valuable. In fact, to me, that's the key to Pierce's candidacy at this point...that he may not have been truly elite at all in any season in his career, but that he was able to maintain that top-15 to top-20 level for an extremely long time with a few forays up to arguably top-10. And that such consistency at very good value might be worth more than other players that gave clearly better peaks but for shorter periods.

To me, that's the Pierce case in a nutshell. He's the ultimate junkyard dog...he's very good at like everything, without being elite at anything. So what I'm trying to weigh is at what point does a whole lot of very good overcome shorter periods of better. And as I pointed out with my big men post, there are still players that I think were enough better for long enough that I'm not ready to vote Pierce yet. Even among the 4 bigs I speak of, I think Mutombo's and Sheed's longevity are close enough to Pierce's that their higher impacts are compelling to me. And while Mourning and Dwight don't have the same longevity, they're enough better than Pierce at his best that right now I'm leaning towards them as well.

[john248]

My official vote is for Paul Pierce. I've always been impressed with Pierce because he looks fundamentally sound with his shooting, ball handing, and solid passing where not a super playmaker but recognized where to get the ball to. Wasn't the most explosive guy like TMac or Kobe, but he could get to his spots to get a shot off. Pretty good in the post, career FTR at .452, career 109 ORTG, career 56.9% TS. 4 seasons over 25 ppg. 04 was a down year with career lows in ORTG, WS/48. In the playoffs though, not as good as you'd expect. Even with nice numbers, I don't view Pierce as an elite offensive player in line with the very best that we've seen. He's clearly at a level below given his inconsistencies in the playoffs and in his career.

Pierce was an above average to a good defender during his time. He was a good man defender prior to Rivers being there though not elite at covering a large area of the court like Pippen. At worse, was a neutral defender during the losing seasons when he had far more responsibility. People point to his defense 08+ when Thibs got on board, but the guy was legit on that end before that when Jim O'Brien was the head coach when they were 5th and 8th in team DRTG in 02 and 03. I'm not calling him an anchor; just saying he was a good defender.

Went to the ECF in 02 with Antoine Walker as the 2nd best player then returned to the semis the year after. 02 is especially interesting since Walker was a high usage player who posted a 99 ORTG which was below the teams 103 value. Pierce as the other high usage guy was 109 ORTG to keep the Celtics offense afloat that year. Followed that up with a couple 1st round exits with Ricky Davis as the other notable player. Walker and Davis...Yikes. Didn't see success again until Ainge came on as the GM who traded a bunch of picks and draft youg guys to be traded later. In the end, he was able to fit in nicely once KG and Allen arrived to contend for a few years and winning a title in the process. Shows that he has a good all-around game and a team player. Never a superstar player but a tier below that for the duration of his career.

He's had some notable playoff moments himself. Down 21 in the 4th to the Nets in 02 to comeback and win it scoring 19 that quarter. Goes 26/9/6 against Ron Artest in 03 where the Celtics upset the Pacers. There are number of clunkers in the playoffs even in 03 and after.

I see that TMac has gotten a vote here. At their best, I'd side with TMac who's a better scorer and passer...an elite player. But his longevity hurts his value here. Once he started having back problems in 06, he was never the same player again, playing just 47 games that season. He stayed on the perimeter more where his shooting really suffered, and this includes at the FT line. Pierce's prime is pretty much double what McGrady's was.

[SactoKingsFan]

I still think Pierce’s all-around game, versatile skillset and great longevity makes him a legit top 35-40 candidate. In addition to great longevity, Pierce provides volume scoring on well above league average efficiency, solid defense, good rebounding and playmaking for a SF and high bball IQ. Generally a reliable playoff performer capable of elevating his level of play and volume scoring in big moments. After being the lone star on teams with consistently subpar supporting casts, he was able to adjust his game to fit within a co-starring role on a championship team and perennial title contender.

Vote: Paul Pierce

Sent from my LG-D800 using RealGM Forums mobile app

[penbeast0]

Spoiler:

Reed v Cowens v. Beaty

All three shared some things in common . . . undersized, mobile centers with good range and good defense. Reed is the strongest, Cowens the most physical, Beaty has a Bill Laimbeer/Vlade Divac rep for annoying opps with a lot of dirty/veteran tricks.

All had shortened careers, Reed played only 7 years over 20 games in a season, Cowens played 10 but with increasing missed games plus a half year attempted comeback in Milwaukee, Beaty played 12 seasons (2 as reserve) but 2 were cut short by his recurring knee injuries; it's not a coincidence that his best season (72 in Utah) came after being forced to sit out a year. Beaty is considered a step down from Reed and Cowens but seems a reasonable comp to show their strengths.

Prime numbers

Reed (67-71) 20.5ppg 13.5reb 2ast .540ts%
Cowens (72-76) 19.5ppg 15.5reb 4ast .495ts%
Beaty* (65-69) 20ppg 12reb 1.5ast .535ts%
*(using NBA numbers only)

As the numbers show, Cowens was the best passer and rebounder, but the worst shooter. Reed is slightly better than Beaty in all areas but only slightly (and the NBA had expansion in these years so Reed was playing against slightly inferior competition). Beaty has the longevity edge. Very comparable players if you consider their defensive abilities roughly equal. I didn't use Beaty's ABA years because I don't want to get into a debate about how much to discount them.

The main non-big prospect seems to be Paul Pierce. I don't have him up there with the top bigs, never got the impression he had that kind of impact on a game. I have him roughly equal with Alex English and a bit above Dantley, Nique, or Richmond. I am open to arguments for Sam Jones or Paul Arizin but the comps we had 2 threads ago had them looking like a bit below Dolph Schayes who I don't have up with the best bigs left, though I am open to being convinced.

There are also a lot of short career prospects. The shortest and greatest peak would be Walton; others would include Connie Hawkins or Sidney Moncrief on the wing, Nate Archibald (short peak anyway) or Kevin Johnson for point guards.

Willing to listen . . . .

Having seen very VERY little footage of him playing, I'd assumed Zelmo wasn't all that good defensively (based on something I'd read once). After reading your post, though, I was reading some other stuff online that indicated he was actually decent on that end.
Why, do you think, is this guy repeatedly passed over historically? Casual fans I'm sure have never even heard of him. A couple years ago I figured him only around the ~#175 rank; and I noted he's never made the top 100 here in the past.

But recently I'm coming to the conclusion that he does indeed deserve some consideration (though not in the peri-50 range, imo).

He was good but never spectacular, on either end. He played during the era of Wilt and Russell when everyone else sufferred by comparison. Bellamy in Chicago had seasons that were similar to TMac in Orlando though pace is an issue; Reed got his rings and fame after Russell retired when Beaty went to the ABA; ABA years get scoffed at no matter how good by many people and again . . . not spectacular numbers like a Spencer Haywood or Julius Erving.

[penbeast0]

Paul Pierce -- Quotatious, trex_8063, Moonbeam, john248, SactoKingsFan,

Dwight Howard -- penbeast0, tsherkin,

Bill Walton -- JordansBulls

Tracy McGrady -- ronnymac2

[D Nice]

I think the underlined is exactly one of the points where we're disagreeing. That's what I was trying to tease out above with my "pound for pound" comment. I'm NOT trying to estimate ability in a vacuum. Chris Paul could very arguably be the most skilled basketball player that ever lived, with more ability for his size than anyone else. But he'll never have the impact that Shaq did, and a lot of that is because Shaq was friggin monstrously huge and that allowed him to dominate the highest efficiency areas of the court. Now, if we're allowed to factor size and the court areas that players occupy as part of "ability", then I revert to what I said above. Big men have more ability to affect defenses than wings or guards...they have the ability to man the highest impact defensive areas BECAUSE they are big enough to deter opponents and get rebounds on an order that smaller players just can't. Again, I don't think you're identifying a bias in RAPM, I think you're identifying a bias in basketball.

I’m asking the questions for people to arrive at their own answers. Do you really believe that Greg Ostertag, Shawn Bradley, Vlade Divac, etc are better defensive players than those guys? If so then fine, take the data at face value, if not, you need to re-assess any kind of “1-to-1” rating of DRAPM for bigs. I personally find that utterly ridiculous, but if you don’t, it would at least give me a clearly picture of how you’re thinking about things. It has nothing to do with “pound for pound.” It has everything to do with how much better these guys actually are than baseline replacement level talent that would otherwise (on average) be theoretically occupying their position. That’s the only way to make cross-position debates meaningful in any way.

And I wasn’t trying to distinguish between impact and ability, those were different words for the same thing. I would have used the word “impact” if the data we were discussing weren’t called “impact stats,” but they are, so I didn’t. Obviously I’m not just listing guys based on basketball skills, I think you’re aware of that, my top 15 players of all time would look something like 11 PGs + Kobe/Jordan/West/Bird if that were the case and I think you’re aware it isn’t.

And if you actually do believe guys like Ostertag and Divac are more valuable than GOAT wing defenders based on RAPM consistency would ask that D-Rob be rated (for example) ahead of Larry Bird. If we had RAPM data for both guys Robinson would almost assuredly smash Larry because of his perceived 2-way play and the fact he’d probably have very good O-RAPM numbers given the offensive bootstrapping he was doing in addition to his 6+ D-RAPM numbers (matching what old fogee Robinson was putting up).

We discussed this a few times a few threads back, and I still don't think you have enough to make your case. To me, from the data that you laid out, the more logical conclusion is that the center/big man defender is the most important defensive slot on the floor and that the RAPM results simply reflect that. Again, I'm not interested in the equivalent of a "pound-for-pound" champion. If Mutombo's size and ability makes it more natural that he occupy the highest impact portions of the defense, and that partially because of that he is able to have a MUCH larger defensive impact than the wing that occupies the same defensive ranking slot as Deke is to centers...then that's just what happens.

There's a reason that height is a sought after commodity in basketball. I don't think you're identifying a bias in Defensive RAPM, I think you're identifying a bias in basketball. Bruce Bowen's defensive rank among wings might be the same as Ben Wallace's defensive rank among big men. But Wallace is able to make a much bigger impact with his defense than Bowen does with his. In fact, Shawn Bradley and Nene may, as well. Barring more evidence of a bias besides an alternate interpretation of the data, I just don't see that your case is that strong.

What more evidence would/could I possibly need/uncover? And again, you aren’t really seeing this either. If it were merely a “more important spot on the floor that affects all players at the position equally” there wouldn’t be such pronounced asymmetry in the distribution because that effect would be filtered out already, as bigs are replaced (by and large) by other bigs. Period.

And I deliberately avoided Ben Wallace because he’s generally a top 5 defensive player of all time and Bruce Bowen is nowhere near that. I’m comparing apex-level perimeter defenders to “pretty good at best” defensive bigs to underscore that when you are comparing Ben Wallace’s D-RAPM numbers to Tracy McGrady’s O-RAPM numbers there might be bias present in Ben’s data (for example) that isn’t going to show up in Tracy’s.

I also don’t think this necessarily is true of every single big. Just because there is a certain data-related flaw does not mean it extends to every single data point, merely that it is much more likely to occur. Deke, for example, his RAPM scores make sense. Having watched essentially his entire career, he always felt like one of the most dominating/game-changing forces on the floor, so I do take his D-RAPM data at face value despite the fact that I do think there’s a positional skew there. However just because I think Deke might be exempt/less affected by the bias does not mean I don’t think it doesn’t affect the +/- of some other players.

Building off of this though an interesting sidebar worth exploring is that these bigs seem to lose some of their defensive importance come the PS. The game slows down, teams get scouted more thoroughly, guys are rested so fatigue is less of a factor, and despite the fact that team defensive ratings fall across the board (it’s pretty well-established teams universally exhibit better defense in the playoffs) the DRTGs of these centers go the opposite way. Below is an examination of 11 ATG bigs, some of whom are already in and some of whom are coming up (primes/extended primes only).

Spoiler:

Hakeem 85-97 RS DRTG: 97
Hakeem 85-97 PS DRTG: 102

D-Rob 90-01 RS DRTG: 96
D-Rob 90-01 PS DRTG: 97

Ewing 86-97 RS DRTG: 99
Ewing 86-97 PS DRTG: 101

Kareem 74-86 RS DRTG: 98
Kareem 74-86 PS DRTG: 101

Garnett 00-12 RS DRTG: 98
Garnett 00-12 PS DRTG: 98

Duncan 98-07 RS DRTG: 94
Duncan 98-07 PS DRTG: 97

Big Ben 00-06 RS DRTG: 93
Big Ben 00-06 PS DRTG: 91

Sheed 99-08 RS DRTG: 100
Sheed 99-08 PS DRTG: 100

Mutombo 92-02 RS DRTG: 100
Mutombo 92-02 PS DRTG: 102

Mourning 93-02 RS DRTG: 99
Mourning 93-02 PS DRTG: 99

Howard 07-14 RS DRTG: 97
Howard 07-14 PS DRTG: 100

I might do a more granular series-by-series/team-by-team analysis at some point in the future but it is KNOWN that team defenses show global improvement in the PS yet the DRTGs of the guys we’re discussing get worse in the playoffs or stay the same, they don’t improve. The very consistent trend is that the teams exhibit lower DRTGs while the Bigs hold constant or have worse DRTGs, meaning their importance/“impact” is degrading. So if there is a “regular season factor” where these guys are more valuable, it’s very clearly the offensive players whose games translate to the PS who are more valuable (in terms of relative rise/fall) then, but because we don’t have PS RAPM they don’t get the same benefit these guys do even though playoffs >> RS in terms of importance. Go figure.

Spoiler:

I'll address your last line in the section below. But other than that, I think you have interesting points about the difference between defense and offense. You can only go so far with a direct comparison of defense vs offense because, as you point out, the units are run differently. However, I also think that this is another part of why bigs tend to have (much) larger defensive impacts than their wing counterparts. Using your phrase, you can't transfer primacy on defense. So when a good defensive wing's man doesn't have the ball, he has fewer ways to directly influence the game. However, when a big man's man doesn't have the ball he is still able to help out on opposing penetrators, he's still able to help prevent passes into the highest efficiency areas of the court, and he's still the most likely player on his team to grab a rebound and end a possession. The big man, defensively, can just do much more than the wing.

Now, back to the examples you gave of a) Kobe or TMac dragging poor offensive casts to top-10 results while b) the 09 - 12 Celtics NOT having top-10 results despite lots of offensive talent. For a) I'd say that it matches exactly with the point I was making...I don't think Pierce, in his best days, could do offensively what Kobe or the version of TMac that you're referring to could do. Pierce was a great scorer with a solid all-around game, but Kobe and TMac were better scorers with much more ability to act as lead guards for their teams. It's a similar point to what I made last thread when comparing Pierce and Iverson...Pierce has a good handle/passing ability for a small forward, but that is NOT the same as being able to run a team. That lack of ability to be an offensive general on that level is exactly WHY I don't think that you could build a top offense with him and only offensive role players, regardless of the system. He would NEED additional offensive help, likely from a gifted point guard, to lift the offensive unit to that level.

Which brings us to b) those Celtics...much, much more-so than the offensive rebounding, the culprit behind the underwhelming offenses was Rajon Rondo and the offense running more through him. Part of the reason that Rondo was given the increased primacy was because his own game was too limited to play off the ball and he was the young talent, so the team tried to play to his strengths. But IMO, another reason for Rondo's increased primacy is that Pierce couldn't handle being the lead guard for the Celtics full time, thus requiring that it be placed elsewhere. And note, this isn't necessarily a negative on Pierce's game. He was, in fact, a small forward and so he wouldn't traditionally be EXPECTED to be the offensive engine in this way. But like defensive big men, I think there's a reason that the top of the offensive RAPM lists are dominated by either point guards or TRULY exceptional defense-warping scorers (or both). The point guard ability to control an offense is a huge way to make an offensive impact, and Pierce doesn't have that on the level of either point guards or wings like Kobe/TMac/AI. But at the same time, Pierce also isn't able to destroy defenses with his scoring the way that the other dominant scorers can. It's why Pierce's offensive RAPM is very good, but not killer. And it's why I don't believe that you could put him as the centerpiece of an offense with only offensive role players and have them produce as well as Mutombo or Howard with a team of defensive role players.

Why restrict blame to Pierce here? I think it should call into question the offensive portability/scalability of Allen and Garnett as well. Yes, Paul was the best offensive player here and he’s the one we’re discussing now so he shoulders some blame but KG and Allen are supposed to be all-time level players in terms of portability at their position because of their multi-functional skill-sets and shooting ability. I do agree that in comparison to Kobe/T-Mac he isn’t the same level of playmaker, but we’re talking about a top 10-15 player and a guy who would’ve been top 20-25 had he stayed healthy. The guys being discussed now aren’t on that level so he shouldn’t be held to the Kobe/T-Mac/Wade standard, but rather the standard of those he’s being compared to. And if the team ORTG was really suppressed by the coaching strategy, that is something that needs to be investigated further before we automatically proclaim Pierce to being scalable to a certain degree solely on the basis of team ORTG.

And I do think there’s definitely something to what Elgee was getting at, offense and defense are at least somewhat interconnected and certain strategies (i.e. the triangle, electing to forgoe crashing the boards) affect the outcome more than we give them credit for, and it’s a perfect explanation to a couple things

a) Was there anything besides Thibs Magic and Posey/Allen/KG/Perkins causing this team to be off-the-charts defensively (especially since Allen and Posey were bench players and only around for the first year)?

b) Why was a team with 3 all-time-greats relatively pedestrian offensively considering talent/fit?

Considering how strong a team the 08-10 Celtics were when healthy in spite of their ORTG does make me wonder if there was some “tricks” of sacrificing some offensive efficacy for better defense, like (for example) the SSOL Suns in reverse. It’s very theoretical but I think it bears mention. Particularly since, in spite of the high offensive talent level, Pierce still retained a very respectable ORAPM in ’08 (4.2) despite having KG and Ray still close enough to their apex.

Spoiler:

I pretty strongly disagree that Pierce was a top-10 player anywhere near that number of times. I've got him in the top 10 maybe a handful of times in his career, and I think that's probably an overestimate. Quotatious, one of the biggest Pierce supporters in this project, says in his post on page 1 of this thread: "Pierce was a top 10 player in the league in a few seasons (2002, 2003, 2006, 2008)"...I think that's much closer to correct than your 9+ times, and I'd quibble a bit with Pierce being no-doubt top-10 even in those years (e.g. 2003 I'd say Duncan, KG, Shaq, Dirk, TMac, Kobe, Kidd, Sheed, Big Ben, Webber all have a pretty strong case). So no, I don't believe that Pierce was top-10 anywhere near as often as you state.

HOWEVER, I do agree with your point that Pierce was very, very good for a very long time. Maybe the case could be made that he was top-15 as often as you suggested he was top-10 (haven't really looked that in depth), and that is still extremely valuable. In fact, to me, that's the key to Pierce's candidacy at this point...that he may not have been truly elite at all in any season in his career, but that he was able to maintain that top-15 to top-20 level for an extremely long time with a few forays up to arguably top-10. And that such consistency at very good value might be worth more than other players that gave clearly better peaks but for shorter periods.

To me, that's the Pierce case in a nutshell. He's the ultimate junkyard dog...he's very good at like everything, without being elite at anything. So what I'm trying to weigh is at what point does a whole lot of very good overcome shorter periods of better. And as I pointed out with my big men post, there are still players that I think were enough better for long enough that I'm not ready to vote Pierce yet. Even among the 4 bigs I speak of, I think Mutombo's and Sheed's longevity are close enough to Pierce's that their higher impacts are compelling to me. And while Mourning and Dwight don't have the same longevity, they're enough better than Pierce at his best that right now I'm leaning towards them as well.

Don’t see how you could exclude Pierce from the top 10 in 2001, 2002, 2003, 2005, 2006, or 2008. Perhaps top 12 or 13 would have been more accurate for 2009 & 2010 and I probably overrated him in 2004. I don’t think it changes too much but the correction is accepted.

Top 10 Seasons: 01, 02, 03, 05, 06, 08
Top 13 Seasons: 04, 09, 10, 11
Top 20 Seasons: 12, 13

And he was pretty good when he was on the floor in 07. Still a fairly large longevity edge over any of the guys being discussed here.

[Doctor MJ]

Vote: Paul Pierce

I'm going to add my vote to the Pierce blow out, and hopefully we can skip the runoff. As mentioned in my post where I went through the strong candidates as I saw them in detail, I don't see Pierce as the clear cut choice here, but he's a solid pick at the spot. His combination of quality and quantity of play is at the very least very rare when we get down into the 40s, and in comparison to the other guys getting votes right now, he clearly stands out as the only one with a truly complete career.

[drza]

Vote: Dwight Howard

All 4 of the big men that I've been analyzing had stronger impacts than Paul Pierce. The only one of the group to have any votes is Howard, so that's influencing me to vote for him here as opposed to one of the other three. Howard reached a level from 2008 - 2012 that Pierce never reached, and he's played a solid decade now so I'm not as worried about longevity. As a legit franchise caliber 2-way big Howard has played at the level to be the best player on a championship team several times, which is something that I don't believe that Pierce could ever do.

[DQuinn1575]

Pierce has one season top 10 mvp and one rpoy

Jones has multiple top mvp seasons and is much more impactful on championships than anyone left


Vote for Sam jones

[SactoKingsFan]

Pierce vs Howard run-off?

Sent from my LG-D800 using RealGM Forums mobile app

[Doctor MJ]

Pierce vs Howard run-off?

Sent from my LG-D800 using RealGM Forums mobile app

By my count that was 6 votes for Pierce and 5 votes for everyone else after 2 days of voting. The last vote makes it 6 to 6, and so if another non-Pierce vote gets in there, maybe it will make most sense to run it off, but from what I see it seems like Pierce won the thing.

[penbeast0]

Paul Pierce -- Quotatious, trex_8063, Moonbeam, john248, SactoKingsFan, Doctor MJ, penbeast0

Dwight Howard -- tsherkin, drza

Bill Walton -- JordansBulls

Tracy McGrady -- ronnymac2

Sam Jones -- DQuinn1575

I didn't get in at the beginning of the hour but was there a few minutes before; it was indeed 6 for Pierce v. 5 for the world. I will change my vote to Pierce to avoid the runoff.

[SactoKingsFan]

[youtube]http://www.youtube.com/watch?v=OU_1f2rC7Gg[/youtube]