RealGM

Bill Russell has been enshrined in the 3rd spot of our list.

The #4 thread will take place over just under 3 days, and after that thread we'll give 2 days a try.

Re-pasting my vote from the previous thread. I'll be sticking with it unless I'm swayed:

Doctor MJ wrote:Vote: LeBron James '09

Here's the baseline: Jordan's in at #1. How much better do you think Jordan is in a season where he does all the right things than LeBron doing all the right things? For me, I don't see it as a significant margin. What puts Jordan clearly over the top is the fact that he ALWAYS brought it, whereas LeBron is more mentally fragile. The fragility though isn't something that is always there, it just rears its head when things break the wrong way.

In '08-09, LeBron was without question my Offensive Player of the Year, and he could add strongly on the defensive side of the ball. Dude was lifting a team by 40 wins. I don't know how anyone can feel all that confident that he's tier below Jordan when he's doing that.

I get the lack of portability. There is a part of me who prefers other types of players because of this, but I sided with Jordan (who is even less portable than LeBron) over those other guys before, and if you were to ask me how I could place the Birds and Garnetts of the world below Jordan but above LeBron, I couldn't really give you a good answer.

In comparison with Wilt, well there's the issue that in '67 he really was quite easily offense more than defense, and his offense isn't something I think could work to anywhere near the same degree later on.

In comparison with Russell, well I go back to the Jordan line of thought. Any argument for Russell over LeBron also seems to work with Jordan. Those arguments were not as strong to me as the Jordan over Russell arguments, and it seems like I can make the same arguments for LeBron.

I'll repost my thoughts on LeBron from the last thread, but I added some new stats

LeBron led the team in pts, reb, ast, blk, and stl. He led a mediocre team to 66 wins. From 08-10, the Cavs were 1-13 in games LeBron missed (+13.7 difference). LeBron had a ridiculous +21.83 in plus/minus which led the NBA. He was +13 on offense and -8 on defense. The Cavs were +14.9 with LeBron on the court which is better than what the 96 Bulls did (+13.4). When LeBron was off the court, they were -7 which is on the level of last year's Cavs and Nets. On offense, the Cavs were 115.1 O rating with LeBron on the court which is around what the 96 Bulls and 10 Suns did. When LeBron was off the court, they were 101 O rating which would have been the worst in the NBA that season (Clippers were the worst at 102.3). They went from historical offense with LeBron to worst offense in league without. On defense, the Cavs were 100.2 with LeBron and 108.31 without. The 100.2 is -8.1 from the league average. That -8.1 is so good that it is better than the 04 Pistons (-7.5). So they went from all-time defense with LeBron on the court, to League average without. The Cavs 8.68 SRS was tied for 7th best ever in the 3 point era ahead of teams like the 87 Lakers, 91 Bulls, 07 Spurs, and 00 Lakers.

From 08-10, LeBron missed 14 games. Here are the results:
With: .737 win% +5.81 MOV
Without: 1-13 (.071 win%) (-7.68 MOV
Difference: .666 win%, +13.67 MOV

Biggest SRS dropoff in history:
1. 99 Bulls -15.82 (MJ/Pippen/Rodman)
2. 11 cavs -15.05 (LeBron)
3. 97 Spurs -13.91 (Drob injured)
4. 91 Nuggets -11.88 (English)
5. 83 Rockets -10.73 (Moses)


Next up, I'll talk about LeBron's clutch season. Arguably one of the clutchest ever. Per 48 minutes he averaged 56-14-13, .693 TS%. He was 2nd in scoring and 1st in assists during the clutch. The Cavs outscored opponents by 45 points per 48 minutes with LeBron on the court in the clutch .


Now onto the playoffs. In fact the Cavs played almost the exact same in the playoffs as they did in the regular season. They had a 112 O rating and 102 D rating in both the regular season and postseason. Their SRS improved from 8.68 to 11.67 in the playoffs. LeBron averaged 35-9-7, .618 TS%, 8.7 TOV%. His PER and WS/48 are easily the best in playoff history. The crazy part is that not one of LeBron's teammates had a PER better than 15 in the playoffs which is the league average. LeBron had a 128 O rating and 100 D rating. MJ never beat 128 O rating or 100 D rating in any of his playoff seasons

In the first two rounds, LeBron averaged 33-9-7, .644 TS%, 38.0% usage% and an insane 6.0 TOV%. That TOV% is so insane that if he did that in the regular season, it would be #2 overall in the league, just 0.1% behind the leader (Dequan Cook). That led to him having a 139 O rating and 90 D rating. To put that into perspective, the all-time #1 O rating is 92 Horace Grant at 132, LeBron would have smashed that. That 90 D rating would have led the league in 2009 ahead of the #1 guy (Howard) who had a 94.6 D rating. This might have been the best two round stretch in NBA history. The Cavs won their games by 16.75 PPG with the closest game being a 10 point win. The Cavs had a 115.95 O rating (similar to 92 Bulls) and a 95.48 D rating (04 Pistons).

Against Orlando, LeBron averaged 39-8-8, .591 TS% against the team with the #1 ranked defense in the league. More amazing, LeBron only had a 11.3 TOV% (38.3 USG%) in that series. Some blame the loss on LeBron being ball dominant but the Cavs had a 112.60 O rating in the series against a team that had a 101.9 D rating in the regular season. That makes the offense +10.70. That is on the level of the 80's Lakers and 00s Suns. For some reason, LeBron gets hate on for his defense in this series but the guys he was guarding (Turkoglu and Alston) didn't do much. Turk averaged 17.2 PPG on 39.0 FG% and Alston 12.5 PPG on 37 FG%. LeBron's sidekicks failed him though. Mo Williams did average 18.3 PPG but on 37 FG%. LeBron only had 3 teammates average 10+ PPG and those teammates did it on .505 TS%. This series was completely on his teammates failing offensively and not being able to guard Lewis and Howard.

The difference between 09 LeBron's WS/48 and 91 MJ's playoff WS/48 is bigger than the difference between 91 MJ and 97 Steve Smith.


Of course I have to mention LeBron's elite defense too. As mentioned above, LeBron shut down Turkoglu and Alston in the Orlando series. In the 1st round, Tayshaun Prince averaged 3.8 PPG and a .278 TS% despite playing 32 MPG. LeBron completely shut him down. In the next round, Joe Johnson averaged 15 PPG on .452 TS%. In the regular season, Carmelo averaged 15.5 PPG on .525 TS% vs. Cleveland. Pierce averaged 20.3 PPG on .490 TS%. Durant averaged 19.5 PPG on .528 TS%. Rudy Gay averaged 11.5 PPG on .350 TS%. Caron Butler who averaged 20.8 PPG on .552 TS% in the regular season dropped down to 15 PPG on .458 TS% vs. Cleveland.

He led the Cavs that season in steals and blocks that season. LeBron was 3rd in the league in D rating and 2nd in the league in Defensive win shares (behind Howard). With LeBron on the court, the defense had a 100.2 D rating which was -8.1 relative to league average. By comparison that is better than the 04 Pistons (-7.5) and 99 Spurs. The Cavs were 8.12 points/100 better with LeBron on the court on defense than without. That was the 2nd best total in the league behind Joel Przybilla. 82games had opponents had a PER vs. LeBron at 10.4 (which is the level of Fredette, Joel Anthony, and Johan petro). According to 82games, his opposing SF had 12.8 pts/36 on .525 TS% and PF had 13.3 pts/36 on .484 TS% when LeBron played PF. He had a +2.8 Defensive RAPM which was 2nd among qualifying perimeter players behind Ron Artest.

Some defensive stats for his position:
17.3 pts/game allowed (1st in league)
41.2 FG% allowed (1st)
15.1 FGA allowed (2nd fewest)
16.6 Efficiency allowed (1st)
1.3 Offensive rebounds allowed (3rd)

Vote LeBron 09

This is a post looking at the best players since 2002 using advanced stats.

Best RAPM since 2002:
1. 10 LeBron 9.8
2.09 LeBron 9.3
3. 07 Duncan 8.8
4..11 Dirk 8.6
5.10 Wade 8.4

Best on court Net +/- since 2002 minimum 2400 MP (80 games*30 MPG):
1. 09 LeBron +15
2. 08 Pierce +14.2
3. 07 Duncan +13.8
4. 07 Bowen +13.7
5. 05 Nash +12.7

Best PER:
1. 09 LeBron 31.7
2. 10 LeBron 31.1
3. 12 LeBron 30.7
4. 09 Wade 30.4
5. 03 T-Mac 30.3

Best WS/48 since 2002:
1. 09 LeBron .318
2. 10 LeBron .299
3. 12 LeBron .298
4. 09 CP3 .292
5. 08 CP3 .284

Best team SRS since 2002:
1. 08 Celtics/KG 9.31
2. 09 Cavs/LeBron 8.68
3. 07 Spurs/Duncan 8.35
4. 03 Mavs/Dirk 7.31
5. 05 Spurs/Duncan 7.84

Best O rating since 2002, minimum 25 usage%:
1. 08 CP3 125
2. 09 CP3 124
3. 08 Amare 124
4. 07 Dirk 123
5. 06 Dirk 123

Playoffs, minimum 400 MP

Best PER since 2002:
1. 09 LeBron 37.4 (best in history)
2. 08 CP3 30.7
3. 03 Shaq 30.6
4. 06 Duncan 30.4
5. 12 Lebron 30.3
6. 10 LeBron 28.6

Best on court Net +/- since 2002, minimum 400 MP:
1. 09 Varejao +16.4
2. 03 Ginobili +15.9
3. 09 Ilgauskas +15.5
4. 10 Lewis +14.9
5. 12 Duncan +14.5

Best playoff WS/48 since 2002:
1. 09 LeBron .399 (best in history)
2. 08 CP3 .289
3. 12 LeBron .284
4. 03 Duncan .279
5. 06 Dirk .263

Best playoff O Rating, minimum 25 Usage%:
1. 09 LeBron 128
2. 08 CP3 126
3. 07 Baron 124
4. 05 Ginobili 124
5. 06 Dirk 124


Based on that, from the guys of the past decade, I would rank it like:
1. LeBron 09
2. Duncan 03
3. Dirk 11

After that you have a mixture of 04 KG, 09 Wade, 08 CP3, 08 Kobe, 03 T-Mac

EDIT

I'm changing my vote to: LeBron '09

-------------


I'm shocked Russell won previous thread - wasn't much better defensively than Hakeem (adjusting for era) and was WAY, WAY worse offensively...

Vote: 09 LeBron

I'm going with LeBron because of the previous posts I've made, and the recents posts that other people have made about him (colts18, Doctor MJ). I've been re-convinced to take him over Wilt.

@ElGee

Well, as colts18 pointed out, you can see that LeBron in Miami is still having serious offensive impact even though he is playing next to someone that doesn't fit him that well. No he's not the greatest at playing off ball and thus blending in with a lot of other stars (KG or Allen would be perfect though), but why would you want him to? It's like taking the ball out of Nash's hands. He could play off-ball, but you'll get the best results if you just give him the ball.

And in 2011 and 2012 combined, Miami is a -3.2 team without LeBron or Wade out there. Put Wade in, and they're +1.2, for a lift of +4.4. Take Wade out and put LeBron in, and the Heat are +5.3, for a lift of +8.5. And when it's just Wade and LeBron, we see a team that's +13.1 with a 114.8 ORating. That's all-time great offense, and that's an all-time great team. So I don't think the less than ideal fit of LeBron and Wade is as crippling as some people think, and I actually think it's quite impressive how he's blended in with Wade and Bosh. And we saw him play quite well with his cuts to the basket and his screens when he did play off-ball in the 2012 playoffs, and that was with a much weaker jump shot than he had in 09.

And similarly to how you said Magic is a guy you just give the ball to and get out of the way, I think that's the same case with LeBron. And just like Magic, LeBron was still a high impact player without having the ball in his hands as much and playing with other talented ball handlers. I think because 09 LeBron was a superior shooter to 12 LeBron, I don't think his portability is as big of an issue as you seem to think.

And the redundancy of a LeBron/Wade pairing is what really holds the duo back. If you paired Magic with Wade, I honestly don't think you'd see spectacular (relatively speaking) results either. Maybe better than LeBron/Wade, but again, you'd be forcing Wade to play off-ball, which as we've seen with LeBron/Wade, he isn't that great at. And if you gave the ball to Wade, you'd be forcing Magic off-ball, which would be totally wasting Magic's talent, and the offense wouldn't be as great as people would think it should be on paper. And that wouldn't be a knock on Magic imo. Even then, I agree it would work better than LeBron/Wade, but only because Magic and Wade aren't practically carbon copies of each other like LeBron/Wade are. LeBron/Wade play very similar games, so the redundancy there is more than we've seen from any other great duo imo.

Also, Paul vs LBJ offensively according to RAPM:

2008: +5.3 (LBJ) vs +3.8 (Paul)
2009: +6.6 (LBJ) vs +4.5 (Paul)
2010: +7.1 (LBJ) vs +3.1 (Paul)
2011: +4.1 (LBJ) vs +4.1 (Paul)
2012: +4.2 (LBJ) vs +5.3 (Paul)

They've only been comparable the last two years when LeBron paired up with Wade, and even then they're about the same level, with a slight edge to Paul last year.

And compare LeBron to Nash (who has been compared favorably to Magic/Jordan/Bird offensively) using RAPM and compare their best RAPM finishes to each other (05-12):

+7.9 (2007 Nash) vs +7.1 (2007 LeBron)
+7.4 (2008 Nash) vs +7.1 (2010 LeBron)
+6.3 (2010 Nash) vs +6.6 (2009 LeBron)
+6.2 (2009 Nash) vs +5.3 (2008 LeBron)
+6.2 (2011 Nash) vs +4.2 (2012 LeBron)
+4.7 (2012 Nash) vs +4.1 (2011 LeBron)
+4.4 (2006 Nash) vs +3.9 (2006 LeBron)
+4.0 (2005 Nash) vs +2.0 (2005 LeBron)

From 06-12, LeBron and Nash are quite comparable offensively, with a slight edge to Nash probably. At their peaks, pretty identical.

And this is supported by the 4-year RAPM from 07/08-10/11:

+6.6 (LeBron) vs +7.7 (Nash)...and LeBron is far and away the best in the league in terms of overall RAPM at +10.2. FTR, Paul is at +5.5 offensively.

Yes, but like I did before, if we compare Nash and LeBron offensively using RAPM over the years, including the 4 year study, we're not seeing a big difference between them at all (and keep in mind that they're both playing the role of ball-dominant playmaker):

2006: +4.4 (Nash) vs +3.9 (LeBron)
2007: +7.9 (Nash) vs +7.1 (LeBron)
2008: +7.4 (Nash) vs +5.3 (LeBron)
2009: +6.2 (Nash) vs +6.6 (LeBron)
2010: +6.3 (Nash) vs +7.1 (LeBron)
2011: +6.2 (Nash) vs +4.1 (LeBron)
2012: +4.7 (Nash) vs +4.2 (LeBron)

LeBron's best finishes are very comparable to Nash's. Looking at the 07/08-10/11 RAPM study (LeBron was not having optimal impact in 08 or 11, and Nash was not having optimal impact in 09 or 11...so it seems to even out, with 2 "peak" offensive seasons between them):

+7.7 (Nash) vs +6.6 (LeBron)...and LeBron was far and away the best overall player in this study, at +10.2. The next closest player was ironically Nash, at +7.8.

It seems to tell me two things: LeBron's "winginess" is being used to detract from him a bit too much...first of all, his ability to score is a good thing; second of all, RAPM is not bearing it out that his penchant for volume scoring is really holding him back all that much as an offensive anchor compared to the greats. Nash ranks better, but not by much, even in his best offensive seasons. I understand RAPM is not perfect, and that on/off data and LeBron's somewhat disappointing results in Miami seem to tell us that he's not on the Magic/Nash level...but I think analyzing his roster situation both in Cleveland and Miami is pretty telling in terms of why he's not leading a +7 offense.

The second thing it tells me is that LeBron's defensive impact, although not on the same level as his offense, is not small, and is not really "a cherry on top" in terms of his contributions. His defense is a pretty big deal for a non-big, and it definitely contributes in a major way to his overall impact. I mean, in that 4-year RAPM study, his defense is coming out close to +4. That actually is big man level.

therealbig3 wrote:Like I asked before, why is Hakeem's peak better than Duncan's? Where is the evidence for that?

Because based on production and efficiency (using mysticbb's SPM), Duncan's clearly superior. Unfortunately, we don't have RAPM for Hakeem, but from the prime years we have for Duncan, he's consistently near or at the top of the list. And we don't even have complete data for his 98-03 run, which was definitely his physical peak.

The SPM really doesn't encapsulate defense, but again, I haven't seen any actual data that supports Hakeem being a better defender than Duncan. We know Duncan in his prime was an elite defender, arguably the best in the league. He's anchored much better defenses than Hakeem. Yes, he has Greg Popovich, and yes, he's had better defensive role players, but where is the cutoff in terms of how much credit we give Duncan and how much credit we give the other players/coach? When Duncan is still a very good defender, but is no longer "best in the game" caliber, the Spurs defense has fallen off significantly.

In terms of impact, including defense, it's hard to say that Hakeem was not only better, but so much better that it compensates for Duncan's clear advantage using SPM (and I'm not trying to use SPM only, if we just compare their statlines against each other it's again very similar). Duncan is very clearly a superstar, MVP-type player according to RAPM.

Anyway, like I said before, I don't see a reason why Hakeem should go now, because imo, there's clearly better players left to go.

Figured it most productive to respond to this here.

Your mentioning of productivity here as a clear thing for Duncan is unexpected. From my perspective, these guys are in about the same ballpark. Maybe you want to elaborate?

I'll also admit that I'm not in the habit of using mystic's SPM, maybe now would be a good time for him and others to do a sales job there. I'll say up front that I don't like combining +/- and box score stats into one number and that I have concerns about SPM based on some of the weights I typically see. Ex: Steals tend to be valued highly in SPM but not correlate well at all to (R)APM. This could well be because steals are very valuable when you get them, but there's a serious cost to employing a thief on days when he doesn't thieve successfully. But all that said, I'm interested. In theory, it should be better than the hand-wavy box score based metrics we typically see.

Re: Duncan doing well by RAPM. What specifically are you seeing that you like? I'm not saying he's bad by any means, but from what I see, his offensive impact in his peak according to that metric is very underwhelming. I don't know how you can take that seriously, and also take his productivity that seriously. In fact, I'll say that Garnett's clear edge in APM & RAPM stats is part of what made me reconsider Duncan vs Garnett, Hakeem, and other people.

Re: Anchored better defenses. Right, but if you're using APM (I won't say RAPM here), it's pretty clear that the impact of a defensive player isn't going to turn water into wine for the team. These arguments were key (still are really) to Duncan being considered a better defender than Garnett before he went to Boston, and when Garnett went to Boston that sure seemed pretty silly. (Although I don't really have a problem with the idea that Duncan impacted defense more than offensive-star-Garnett did in minny.)

What we do know is that Hakeem blocked shots more than Duncan, and stole the ball more than Duncan, and his peak rebounding was similar, and he was clearly quicker with unreal coordination, and he showed remarkable ability to outsmart opponents and make use of the rest of the guys on the court. So what I keep saying is this:

There are very clear indicators that Hakeem was doing more on defense than Duncan, so isn't a pro-Duncan argument really saying that a lot of that "more" Hakeem was doing wasn't as valuable as you'd first think? If that's the case, and we have people who are clearly siding with Duncan, shouldn't they be able to tell us what it is that Hakeem was doing wrong?

Not saying you have to side with me with Hakeem if you can't rebut that of course, but really, shouldn't you be able to answer that in some way before you dismiss a guy putting up defensive numbers we've NEVER actually seen by ANYONE since the time official records have been kept while gaining a parallel reputation of having remarkable instincts and team savvy?

I'm not even clearly siding with Duncan lol. I just don't think Hakeem has separated himself from the Duncan/KG tier, so I'm objecting to his name being thrown around, while those two are being ignored. I haven't seen anything that clearly gives him the edge over those two. If Hakeem is a legitimate contender at this point, so are Duncan/KG. I don't think they are, and so I don't think Hakeem is.

Doctor MJ wrote:Re: Duncan doing well by RAPM. What specifically are you seeing that you like? I'm not saying he's bad by any means, but from what I see, his offensive impact in his peak according to that metric is very underwhelming. I don't know how you can take that seriously, and also take his productivity that seriously. In fact, I'll say that Garnett's clear edge in APM & RAPM stats is part of what made me reconsider Duncan vs Garnett, Hakeem, and other people.

Year by year, RAPM from 04-08:

04 Duncan: +4.9 (+0.9 offense, +4.1 defense)
04 Garnett: +8.0 (+4.5 offense, +3.6 defense)

05 Duncan: +6.0 (+2.2 offense, +3.8 defense)
05 Garnett: +4.4 (+3.1 offense, +1.3 defense)

06 Duncan: +6.1 (+2.4 offense, +3.6 defense)
06 Garnett: +4.4 (+2.5 offense, +1.9 defense)

07 Duncan: +8.8 (+6.3 offense, +2.5 defense)
07 Garnett: +7.0 (+2.7 offense, +4.3 defense)

08 Duncan: +6.3 (+3.2 offense, +3.1 defense)
08 Garnett: +8.1 (+3.0 offense, +5.2 defense)


Where is KG's clear advantage here? And like I said, Duncan's 98-03 run isn't fully covered by RAPM (partially in 02 and 03), and that's when he was at his physical peak, so his defense may have been even better. Duncan comes out ahead of Garnett from 05-07, and is actually pretty close in 08. The only time KG has clearly outclassed Duncan according to RAPM is 04, which was his peak year.

Outside of 04, I'm seeing Duncan in the +2 to +3 range, with a huge +6 in 07. His 04 season was down, but he was injured that year. For a big man, he's actually ranking quite well offensively. An interesting thing to me is that he was clearly better defensively according to RAPM than Garnett until 07...including in Garnett's 04 peak. And this is even more interesting when you consider that Duncan being on good teams should mean his impact should be less than Garnett's...but it's not, outside of Garnett's peak year and outside of post-07 seasons when Duncan was no longer in his prime. This is another reason I would really love to have complete RAPM data from 98-03, because I think Duncan was better in those years than he was in 04-07.

And I know mysticbb made some posts regarding SPM, its superiority over APM, and the issues with long-term studies regarding RAPM in the past that really convinced me. That's why I've personally put more stock in SPM, I don't really take APM seriously, and I use the 1-year studies a lot more. And I really don't think it's a case of me looking at what I want to see, it's just that mystic's posts really made sense to me from what I understood. I'll see if I can find them later, or if mystic actually can chime in.

Doctor MJ wrote:Re: Anchored better defenses. Right, but if you're using APM (I won't say RAPM here), it's pretty clear that the impact of a defensive player isn't going to turn water into wine for the team. These arguments were key (still are really) to Duncan being considered a better defender than Garnett before he went to Boston, and when Garnett went to Boston that sure seemed pretty silly. (Although I don't really have a problem with the idea that Duncan impacted defense more than offensive-star-Garnett did in minny.)

So I'm not seeing why it's silly to not think Garnett in 08 is better defensively than anything we've seen from Duncan, because Duncan wasn't even in his defensive peak from 04-07, and he was still better than Garnett, including Garnett's peak year.

Doctor MJ wrote:What we do know is that Hakeem blocked shots more than Duncan, and stole the ball more than Duncan, and his peak rebounding was similar, and he was clearly quicker with unreal coordination, and he showed remarkable ability to outsmart opponents and make use of the rest of the guys on the court. So what I keep saying is this:

There are very clear indicators that Hakeem was doing more on defense than Duncan, so isn't a pro-Duncan argument really saying that a lot of that "more" Hakeem was doing wasn't as valuable as you'd first think? If that's the case, and we have people who are clearly siding with Duncan, shouldn't they be able to tell us what it is that Hakeem was doing wrong?

Not saying you have to side with me with Hakeem if you can't rebut that of course, but really, shouldn't you be able to answer that in some way before you dismiss a guy putting up defensive numbers we've NEVER actually seen by ANYONE since the time official records have been kept while gaining a parallel reputation of having remarkable instincts and team savvy?

Ok, so my rationale is, even if I can't explain it, if the numbers are showing me that one guy is better, he must be doing something better than the other guy which ultimately has him having higher impact. Just because I can't really explain HOW he's doing it shouldn't mean I should punish him in light of the fact that he IS doing it. What we know is that Duncan is measuring out to have anchored better defenses (and their rise and fall correlate with Duncan's rise and fall), we do know that his overall RAPM is excellent and is always among the best, we do know his defensive RAPM is among the best, and we do know his SPM/PER is clearly superior to Hakeem's.

From what I've seen, I think Duncan's defensive IQ in terms of help defense and positioning is GOAT-worthy in those departments. So for my money, even though he's not blocking as many shots, pickpocketing as many people, and isn't as quick/athletic as Hakeem, I think his defensive savvy (even better than Hakeem's imo) compensates for it. All in all, I don't have a problem at all saying that Hakeem is better defensively, but I wouldn't say it's anything significant.

Their SPM in certain seasons btw:

94 Hakeem: +5.37 RS, +6.63 PS
95 Hakeem: +4.89 RS, +6.01 PS

02 Duncan: +6.66 RS, +9.11 PS
03 Duncan: +6.47 RS, +8.05 PS

And Duncan has more post seasons than just 02 and 03 in which he clearly surpasses Hakeem's 94 and 95 post seasons. At the very least, I'm seeing a guy in Duncan that's elevating his game from the regular season to the playoffs even more than Hakeem.

I would take 2012 Lebron before 2009 Lebron unless you guys have already decided that 2009 Lebron was better at his actual peak. I would disagree and I've made numerous posts explaining why I believe 2012 Lebron is better.

^lol, then vote for 2012 LeBron if you think he should go here. Nobody's saying you have to agree on the year.

Doctor MJ wrote:I'll also admit that I'm not in the habit of using mystic's SPM, maybe now would be a good time for him and others to do a sales job there. I'll say up front that I don't like combining +/- and box score stats into one number and that I have concerns about SPM based on some of the weights I typically see. Ex: Steals tend to be valued highly in SPM but not correlate well at all to (R)APM. This could well be because steals are very valuable when you get them, but there's a serious cost to employing a thief on days when he doesn't thieve successfully. But all that said, I'm interested. In theory, it should be better than the hand-wavy box score based metrics we typically see.

Last season, 374 qualified players, correlation between SPM and STL% was 0.12, correlation between RAPM and STL% was 0.15. We can expect a similar result for other seasons.

SPM is basically just a conversion from a boxscore metric to a +/- numbers. The conversion is based on a regression of the average team boxscore metric on the MOV. Then I use that model to convert the individual boxscore metric into an unadjusted SPM. The last step is the adjustment for the strength of schedule. It is just a way better approach to say a player is +3.5 per 100 possession based on boxscore than say a player has a PER of 20 based on boxscore. Essentially both mean the same. I can use SPM much easier to predict outcome of games for example, because it is already converted into a suitable number. When I have a team with an average SPM of +2 and another with -1, I can predict that the +2 team will likely win by +6 at home (2-(-1)+3=6; 3 is the HCA). That is a pretty simple way, if you would take the PER you would probably get something like this: average PER of 16 vs. average PER of 14.5. As you can see, it is rather difficult to make a predicition based on that.

My SPM was tested by Engelmann (same approach as the one for RAPM) and the result was that SPM was actually slightly better at predicting the outcome of future games than RAPM in terms of scoring margin. It was worse at predicting offensive and defensive efficiency. The reason is simple: The boxscore is heavily biased towards offense and important defensive informations are missing. Take Carlos Boozer for example. In terms of boxscore stats he looks like a great defender, but his help and team defense as well as his man-to-man defense are not good at all. Boozer is a bad defender. Well, but to make the split between offense and defense I have to use the boxscore somehow, and that split is doing well for a lot of players, but is pretty bad for other. That completely explains why SPM is worse than RAPM in terms of predicting offensive and defensive efficiency. But the main feature is the overall value, the overall scoring margin, which determines the outcome of the game. The split between offense and defense can be seen as some sort of gimmick, something which has to be taken with a grain of salt. Also, SPM is just telling you what kind of combination of production and efficiency a player has. It does not tell you anything which is not captured by the boxscore. That goes not just for the defensive part, but also for the offensive part.

The underlying boxscore metric is my PRA. That is a descriptive non-linear model. It has a 0.96 correlation coefficient to win%. There is an offensive part (points, assists, offensive rebounds, turnovers) and a defensive part (steals, blocked shots, defensive rebounds, fouls). The offensive part is extremely good. I get the exact amount of points per possessions for each team. Offensive rebounds and assists are handled in the same fashion, but the coefficients are differently for each team. For example, an assist by Steve Nash on the Suns is more valuable than an assist by John Stockton. The team overall assist-rate is taken into account in order to determine whether assists are generated rather by the offensive setting than by the individual player. If we look at the 2007 Nash vs. 1990 Stockton, Stockton is seen as the better player by a combination of WS/48 and PER by about 3%, while SPM has Nash being better by about 10%.
A similar thing happens with offensive rebounds. In that way I can capture the impact of offensive rebounds on the overall outcome of games much better than a static value.

SPM is easily beating out PER in terms of explaining and predicting, it is better than WS/48, even though not by a huge margin. It is a far better metric in order to predict than WP48. The only other boxscore based metric I would consider is DsMok's ASPM, a statistical plus minus approach based on the regression of some boxscore based values on a multiyear RAPM.

Right now I prepare some sort of relaunch of the numbers on my blog, because I found an error in the subroutine for splitting between offense and defense. I will also add further years (all years between 1978 and 2012).

It is kind of disturbing to me that Russell won the previous thread when the only argument being put forth was '-10 on defense!!!! -10!!!'.

But whatever, my vote isn't changing.

Vote: 1967 Wilt Chamberlain

I'm looking forward to the Magic/Bird talk though. IMO it's too early to be considering LeBron, but the discussion will be interesting.

therealbig3 wrote:And I know mysticbb made some posts regarding SPM, its superiority over APM, and the issues with long-term studies regarding RAPM in the past that really convinced me.

While SPM is clearly superior to APM in terms of predicting and explaining team overall success, we should also keep in mind that they are based on two rather different things. And then again, we have RAPM values from 2003 to 2012, there is really no need to argue with APM based numbers duing those years. RAPM is clearly superior to APM, and the best RAPM values are those prior-informed 1yr RAPM values. Someone trying to argue with APM in one of those years or with multi-year APM/RAPM studies is either not understanding the underlying theory or simply has an agenda.

For 2000, 2001 and 2002 we don't have reliable RAPM data, because the raw data is not easily available. bbr.com is not willing to give the 2001 and 2002 raw data out for free (which is completely understandable, they put the effort in and had the costs in the first place, earning money with that is absolutely deserved!) and Wayne Winston has also not shown interest in sharing his raw data, which goes at least back until the 1999/2000 season. Winston has the APM data for Bryant, Garnett, Nowitzki and Duncan posted on his blog for 2000, 2001 and 2002. We can use bbr.com in order to get the Net+/- per 100 possessions.

If I want to put a number on a player, I use a combination of SPM and RAPM (coefficients found via regression). In that way I account for production, efficiency and impact. But that was also never tested in an out-of-sample test, I can just assume that it is at least as good as SPM and RAPM individually.


And, btw, I think James is a pretty good choice at that point. Whether someone thinks it is 2012 or 2009. I personally think 2009 was better. In 2012 James just won the title and that seems to make people believe a player changed dramatically in terms of skills or mental approach, while it is far more likely that circumstances and variance are the reason for not winning in previous seasons.

I still have '76 Erving over LeBron.

ThaRegul8r wrote:I still have '76 Erving over LeBron.

76 Doc was unreal. Pure dominance.

Between these 3 for me

--------- RS PER, WS48, --------- PER, WS48 playoffs
KAJ 1971: 29.0, 0.33, -----------25.0, 0.27 (14 playoff games, title)
Wilt 1967: 26.5, 0.285------------25.3, 0.25 (15 playoff games, title)
Hakeem 1994: 25.3, 0.210----------27.7, 0.208 (23 playoff games, title)

Vote: Kareem 1971 (Not to mention put a team that never won anything and gave them a title) and was elite on both ends of the floor and had the best stats in the league by far.

colts18 wrote:
Biggest SRS dropoff in history:
1. 99 Bulls -15.82 (MJ/Pippen/Rodman)
2. 11 cavs -15.05 (LeBron)
3. 97 Spurs -13.91 (Drob injured)
4. 91 Nuggets -11.88 (English)
5. 83 Rockets -10.73 (Moses)

How do you list the Bulls lost 3 players and the Cavs only Lebron when the Cavs lost half it's team as well including Shaq, Big Z, etc. Not saying Lebron wasn't the most important piece but it goes without saying they lost more than just him. Not to mention the Spurs didn't only lose Robinson but Elliot was out as well.

I am voting for 09 Lebron. I have never seen a player carry a team so hard both in the regular season and in the playoffs. He was super clutch that year and his playoff run was legendary. He loss to the Magic but it was only because his team was even more pitiful than Dwight's team.
08 Lebron and 11 Lebron get honorable mentions.

EDIT: Changing my vote to Hakeem 94

Guess Lebron's gonna win?

I'm still going 03 Duncan for my vote. Basically played every role but perimeter defender. And as impressive as Lebron's 09 PS was he didn't bring it on one end (very nitpicky I know) and his crazy high impact numbers don't translate to balanced teams at all.

If Russell can get the no. 3 spot with largely one-way impact, Magic and Bird, should at least be in the discussion as well.

Look, it's not like they were net negatives on defense. And '87 Magic and '86 Bird are probably close the 1-2 in greatest offensive seasons ever with '67 Wilt in the discussion, of course. If '65 Russell gets voted in as the no. 1 best defensive season ever while his offensive impact was probably close to zero (I know that people like to talk about his passing, but if your team is the worst in the league in a particular category, you're probably not very good at that category), how can these two not be right with him when they both made at least some positive defensive contributions with their rebounding and smart team defense?