RealGM NBA Top 100 list -- #1

[magicmerl]

Although there are kinda 4 candidates for the #1 slot, I think that Wilt is just outclassed by both Russell and Jordan, and any Kareem arguement has to be based heavily on his longevity when I think that the peak/prime of Jordan/Russell is so much better.

Which really leaves us with 2 candidates, Jordan and Russell. I have several points to touch on, namely non-basketball aspects of the players, relative rankings at their respective positions, offense vs defense stats and differences in era.

#1 non-basketball related
One thing that Jordan did that is totally unprecedented before or since is how much he popularised the NBA. While Bird/Magic brought the NBA to the age of TV broadcasts, MJ in a sense was actually bigger than the NBA in that people were fands of him or his brand than they were of basketball. So he's certainly the greatest marketing force in NBA history. The question is, how much of a factor is that in the GOAT rankings? I personally completely discount this aspect of Jordan in terms of a list like this. Just like I mentally imagine that every modern NBA owner doesn't give a hoot about the luxury tax. If they get rid of a player to get under the tax, and make their team on the floor worse, that's 'bad' in a basketball sense.

#2 Relative rankings of respective positions
Jordan also stands out as special because of how much better he is than the next best shooting guard. In the 'normal' NBA the most precious and rare commodity is the big man simply because they are so rare. But across all of NBA history there are far more special big men, and they aren't as rare on an all-time basis. Put simply, the gap between Russell and Kareem is much narrower than the gap between Jordan and Kobe in my mind. And this counts as a credit to Jordan. In a similar way, I think that Duncan's incredible longevity and the inevitable comparisons to Kareem hurt Kareem in this type of discussion because he is no longer so unique in the NBA pantheon. His special uniqueness is just a little bit less unique than it was 5 years ago.

#3 Offense vs Defense stats
Part of the problem in a comparison like this is that the skills of offense are much more accurately captured than those of defense, particularly in the dark ages that Russell played in. So we can articulate how good Jordan was on offense in much more detail than we can about Russell's defense, because defense is just less well defined and understood in statistical terms in basketball. I think that as time goes on we are going to understand more and more of the nuances of great defense and this part of how people esteem Russell will continue to rise.

#4 Differences across era
A lot of people simply discount anything prior to the 'tv era', which also coincides nicely with the ABA merger and introduction of the 3pt line. We simply have much more video footage since that time to use to make player evaluations. Plus previous eras were less competitive because of the reduced player pool. 50's era basketball comes with a giant question mark because of the lack of black players. The lack of a 3pt line just made basketball quite different from the game it is today. I can see some merit to dismissing all 'historical' players on this basis. But if this is an arguement for stopping a player from being the #1 pick, then where do you stop? Does it also stop a player from being selected #2, or #10? Does it invalidate a player from *every* ranking? I think that ultimately this needs to be discounted, or at least presented in a more detailed manner than just a 'differences across era' arguement.

I vote for Bill Russell as #1.

[MisterWestside]

Please expound. What is it about each of those players that made them the exact type of player that Russell could thrive with defensively.

How about the fact that these players helped to take the load off of Russell by doing something with the basketball on the other end? Or (unrelated) that the coach dug in his heels and encouraged/implemented the style of play that Russell made famous?

I've watched enough sports franchises take something so breathtakingly simple and screw it up because of crappy management. Russell's the man here, and he's at the center of greatness; but thinking of a bunch of other ways in which the aura of Russell and the Celtics doesn't exist because of the lack of other pieces isn't repugnant to reason.

[therealbig3]

So first off, when looking to evaluate the dominance of an outlier it's not actually good to compare them to the mean because they distort the mean. This is particularly the case in a smaller group.

So for example, in '64, if you compare the Celtics to the median, they rate a +12.8, and if you compare them to the mean of the team's not including them it's still a +12. Basically then the Spurs have only 1 year even 2/3rds as dominant as that, even before you consider that a 12-13 edge relative to efficiencies south of 100 makes for a great percentage edge that it would in later eras with higher average efficiencies.

So yeah, just peak vs peak, there's absolutely no contest here. We can talk about degree of difficulty, we can talk about luck, but just the numbers, Russell's Celtics hang'em up there off the charts.

By standard deviation, the 64 Celtics were not 1.5x more dominant than the best Duncan team. The Celtics weren't even the most dominant defensive team by that measure.

Here is a B-R post on the players who played for the best defensive teams. This post was in 2010 which coincidentally is after Duncan's 13th season (Russell played 13).


4. Bill Russell -6.08
8. Tim Duncan -5.60

http://www.basketball-reference.com/blog/?p=7239

Okay, but the numbers I quote still stand.

I like using standard deviations and other statistical techniques too, but the numbers I'm quoting here are easy to understand. One team being 12 points beyond the median while another is 8 points beyond the median is a big difference no matter how you slice it.

Obviously, but that comes back to my original point...are we really going to give the Celtics (and thus Russell) that much more credit than the Spurs (and thus Duncan) when Russell played during a time when his coach was WAY ahead of his contemporaries and when offensive basketball in general was just not very good? I think it's fair to assume that it became much harder to deviate from the median the way Russell's Celtics did when league-wide offense started playing a lot better...I don't believe it's really possible to be much more than a +8 defense during the more modern era. The 08 Celtics were +8.9. The 93 Knicks were +8.1. The 94 Knicks were +8.5. The Spurs (not sure which year) were about +8.

I think this is something that people will just agree to disagree on. I have no problem with people voting for Russell, because I can understand it if you feel that he was just THAT much better defensively than anyone since. I don't feel that way...I feel he was a little luckier than some more modern defensive greats, because he was able to play in circumstances where his defense could shine through A LOT more. Personally, I don't want to reward him for that, or in other words, punish more modern players for that.

[Doctor MJ]

And my response to Russell's body-type: I'm not convinced that he was a better athlete than Robinson or Olajuwon or Garnett, and I know Duncan wasn't the same level of athlete that any of those guys were, but I've never been convinced that they were clearly superior defensive players than him despite that. I have them all in the same tier.

As for DRAPM over the years, compare him to Garnett when both were in their primes (02-07), as well as 08, which is Garnett's most famous year defensively:

02 Duncan: +2.3 (t-2nd)
02 Garnett: +1.2 (t-26th)

03 Duncan: +3.4 (t-1st)
03 Garnett: +2.3 (8th)

04 Duncan: +4.3 (2nd)
04 Garnett: +4.2 (3rd)

05 Duncan: +3.8 (3rd)
05 Garnett: +1.3 (t-57th)

06 Duncan: +3.6 (2nd)
06 Garnett: +1.9 (t-28th)

07 Duncan: +2.5 (t-18th)
07 Garnett: +4.3 (1st)

08 Duncan: +3.1 (t-5th)
08 Garnett: +5.2 (1st)

RAPM seems to support Duncan's candidacy as a GOAT-defender, as he was ranking as a top 1-3 defender every year of his prime we have the data available with the exception of 07, and seems to clearly outperform Kevin Garnett from 02-06. And Garnett doesn't do anything in 08 that Duncan didn't do prior to that.

Anyway, not really the place for a Duncan vs Garnett discussion, but my point is, nothing really supports that Duncan was a tier below ANYONE on defense, just a belief that because he wasn't as mobile as some other players, he couldn't possibly be on their level defensively. Which isn't the case imo. He's always struck me as a more intelligent and a more disciplined defender than Hakeem or Robinson, and his rim protection and superior post defense (due to strength) compared to Garnett compensates for his inferior but still excellent mobility, imo.

FWIW, I'm not treating RAPM (or any variant of +/-) as the be-all, end-all...like I said, I don't really rank Duncan over Garnett as a defensive player, I see them as more or less equal. And I see both of them as more or less equal defensively to Robinson and Olajuwon. And I see all 4 of them as more or less equal defensively to Russell.

Re: Russell vs Hakeem/Robinson/Garnett athletes. I agree that that's actually a debate. I can go into specific reasons why I rank Russell overall ahead of those guys, but you're right, they're all more impressive than Duncan in terms of their agility.

Is that enough to make them better defenders than Duncan? Not by itself no, but I would say you need a good reason why you'd knock them below Duncan to really make it debatable.


The Duncan vs Garnett defensive thing with RAPM is really simple actually:

Garnett in Boston
Duncan
Garnett in Minnesota

This is why in the numbers you showed Garnett when he gets to Boston immediately jumps past anything we see from Duncan. I could see something where you'd say "Well if Duncan got to focus on defense like Garnett did in Boston", but I actually think that goes against what's so cool about Duncan: He had a low energy game that let him play two-way without coasting on either end.

If your next thought is: Aha! So you admit Duncan's the better two-way player!

Well no, as I've mentioned, the overall RAPM favors Garnett soundly, it's just that he can also ramp up his one-side impact if you have others who can pick up the rest of the slack.

Even still in terms of Duncan being a GOAT defender in terms of RAPM numbers if the numbers actually did favor Duncan over Garnett, the best defender of the +/- era is clearly Dikembe Mutombo. For perspective, the best year we have from Duncan gives him a +6.78, whereas even in the limited span we have from Mutombo we have a +9.74 year.

Now, if you want to say "How do we know Russell was as good as Mutombo on defense?", well, that's one more avenue to take, but first let's acknowledge that you you thought you were seeing enough to say perhaps "Can we really expect numbers much better than what we've seen from Duncan from anyone?", and yeah, we really do have reason to think that Duncan's nowhere near that plausible peak human threshold.

[Doctor MJ]

Obviously, but that comes back to my original point...are we really going to give the Celtics (and thus Russell) that much more credit than the Spurs (and thus Duncan) when Russell played during a time when his coach was WAY ahead of his contemporaries and when offensive basketball in general was just not very good? I think it's fair to assume that it became much harder to deviate from the median the way Russell's Celtics did when league-wide offense started playing a lot better...I don't believe it's really possible to be much more than a +8 defense during the more modern era. The 08 Celtics were +8.9. The 93 Knicks were +8.1. The 94 Knicks were +8.5. The Spurs (not sure which year) were about +8.

I think this is something that people will just agree to disagree on. I have no problem with people voting for Russell, because I can understand it if you feel that he was just THAT much better defensively than anyone since. I don't feel that way...I feel he was a little luckier than some more modern defensive greats, because he was able to play in circumstances where his defense could shine through A LOT more. Personally, I don't want to reward him for that, or in other words, punish more modern players for that.

If you're willing to just concede the numbers and focus your attack on whether it's all that impressive given the era, cool. By no means do i expect people to call Russell GOAT just because he looks like Babe Ruth from some perspectives.

I'll try to some up my fascination like this:

Russell came in as the spearhead. It wasn't that Auerbach was ahead of his time so much as that Russell was, and Auerbach was smart enough to let Russell keep doing his thing in the pros.

That was an earthquake that began a period of tremendous growth in the NBA with brilliant new players and strategies, the likes of which no player has really seen since.

And through it all, to the very end of his career, with no coach other than himself at that point, his teams won, and won like no one else before or after.

It astonishes me.

[therealbig3]

And my response to Russell's body-type: I'm not convinced that he was a better athlete than Robinson or Olajuwon or Garnett, and I know Duncan wasn't the same level of athlete that any of those guys were, but I've never been convinced that they were clearly superior defensive players than him despite that. I have them all in the same tier.

As for DRAPM over the years, compare him to Garnett when both were in their primes (02-07), as well as 08, which is Garnett's most famous year defensively:

02 Duncan: +2.3 (t-2nd)
02 Garnett: +1.2 (t-26th)

03 Duncan: +3.4 (t-1st)
03 Garnett: +2.3 (8th)

04 Duncan: +4.3 (2nd)
04 Garnett: +4.2 (3rd)

05 Duncan: +3.8 (3rd)
05 Garnett: +1.3 (t-57th)

06 Duncan: +3.6 (2nd)
06 Garnett: +1.9 (t-28th)

07 Duncan: +2.5 (t-18th)
07 Garnett: +4.3 (1st)

08 Duncan: +3.1 (t-5th)
08 Garnett: +5.2 (1st)

RAPM seems to support Duncan's candidacy as a GOAT-defender, as he was ranking as a top 1-3 defender every year of his prime we have the data available with the exception of 07, and seems to clearly outperform Kevin Garnett from 02-06. And Garnett doesn't do anything in 08 that Duncan didn't do prior to that.

Anyway, not really the place for a Duncan vs Garnett discussion, but my point is, nothing really supports that Duncan was a tier below ANYONE on defense, just a belief that because he wasn't as mobile as some other players, he couldn't possibly be on their level defensively. Which isn't the case imo. He's always struck me as a more intelligent and a more disciplined defender than Hakeem or Robinson, and his rim protection and superior post defense (due to strength) compared to Garnett compensates for his inferior but still excellent mobility, imo.

FWIW, I'm not treating RAPM (or any variant of +/-) as the be-all, end-all...like I said, I don't really rank Duncan over Garnett as a defensive player, I see them as more or less equal. And I see both of them as more or less equal defensively to Robinson and Olajuwon. And I see all 4 of them as more or less equal defensively to Russell.

Re: Russell vs Hakeem/Robinson/Garnett athletes. I agree that that's actually a debate. I can go into specific reasons why I rank Russell overall ahead of those guys, but you're right, they're all more impressive than Duncan in terms of their agility.

Is that enough to make them better defenders than Duncan? Not by itself no, but I would say you need a good reason why you'd knock them below Duncan to really make it debatable.


The Duncan vs Garnett defensive thing with RAPM is really simple actually:

Garnett in Boston
Duncan
Garnett in Minnesota

This is why in the numbers you showed Garnett when he gets to Boston immediately jumps past anything we see from Duncan. I could see something where you'd say "Well if Duncan got to focus on defense like Garnett did in Boston", but I actually think that goes against what's so cool about Duncan: He had a low energy game that let him play two-way without coasting on either end.

If your next thought is: Aha! So you admit Duncan's the better two-way player!

Well no, as I've mentioned, the overall RAPM favors Garnett soundly, it's just that he can also ramp up his one-side impact if you have others who can pick up the rest of the slack.

Even still in terms of Duncan being a GOAT defender in terms of RAPM numbers if the numbers actually did favor Duncan over Garnett, the best defender of the +/- era is clearly Dikembe Mutombo. For perspective, the best year we have from Duncan gives him a +6.78, whereas even in the limited span we have from Mutombo we have a +9.74 year.

Now, if you want to say "How do we know Russell was as good as Mutombo on defense?", well, that's one more avenue to take, but first let's acknowledge that you you thought you were seeing enough to say perhaps "Can we really expect numbers much better than what we've seen from Duncan from anyone?", and yeah, we really do have reason to think that Duncan's nowhere near that plausible peak human threshold.

First of all...how exactly does Garnett jump past anything we saw from Duncan once he gets to Boston? Are you looking at the raw numbers? Because my understanding is that you can't compare raw RAPM across years like that. I also don't understand what you're saying by Garnett clearly trumping Duncan in terms of RAPM. I don't see that happening.

Second of all...at least with regards to DRAPM (not sure about APM or raw +/- or hybrid RAPM), Mutombo doesn't really rank ahead of Duncan.

02 Duncan: +2.3
02 Mutombo: +1.3

03 Duncan: +3.4
03 Mutombo: +1.3

04 Duncan: +4.3
04 Mutombo: +1.1

After 04, Mutombo became a 10-15 mpg role player to end his career and his defensive role was thus not even close to comparable to Duncan's. Furthermore, it's not really uncommon to see role players in unique situations have a surprisingly high RAPM...in one of these years, Jason Collins was far and away the leader in DRAPM, at +6.3.

[An Unbiased Fan]

Okay, but the numbers I quote still stand.

I like using standard deviations and other statistical techniques too, but the numbers I'm quoting here are easy to understand. One team being 12 points beyond the median while another is 8 points beyond the median is a big difference no matter how you slice it.

The deviation in a 8 team league is going to be much greater than in a 30 team league. In the context of league size, I'm more impressed by the Spurs defense.

In a 8 team league, each team has a 12.5% impact on offense/defense league numbers. In a 30 team league they only have 3.3% impact. Clearly, there will be MUCH more parity in a 30 team league, and less divergence. A couple bad offense can skew the whole metric in the 60's, but not so in the modern era where there are 28 other teams. its not a straight line correlation, but very significant.

I won't even get into the rule book differences, because the Spurs are more impressive in context of their eras.

[ceiling raiser]

Okay, but the numbers I quote still stand.

I like using standard deviations and other statistical techniques too, but the numbers I'm quoting here are easy to understand. One team being 12 points beyond the median while another is 8 points beyond the median is a big difference no matter how you slice it.

The deviation in a 8 team league is going to be much greater than in a 30 team league. In the context of league size, I'm more impressed by the Spurs defense.

In a 8 team league, each team has a 12.5% impact on offense/defense league numbers. In a 30 team league they only have 3.3% impact. Clearly, there will be MUCH more parity in a 30 team league, and less divergence. A couple bad offense can skew the whole metric in the 60's, but not so in the modern era where there are 28 other teams. its not a straight line correlation, but very significant.

I won't even get into the rule book differences, because the Spurs are more impressive in context of their eras.

By the same virtue, a single outlier defense can drastically alter the standard deviation as well. Goes both ways.

[Doctor MJ]

Please expound. What is it about each of those players that made them the exact type of player that Russell could thrive with defensively.

How about the fact that these players helped to take the load off of Russell by doing something with the basketball on the other end? Or (unrelated) that the coach dug in his heels and encouraged/implemented the style of play that Russell made famous?

I've watched enough sports franchises take something so breathtakingly simple and screw it up because of crappy management. Russell's the man here, and he's at the center of greatness; but thinking of a bunch of other ways in which the aura of Russell and the Celtics doesn't exist because of the lack of other pieces isn't repugnant to reason.

Help take the load off? I mean, c'mon, that's just normal supporting cast stuff. I've gone into how fortunate Jordan was too have rebounding specialists like Dennis Rodman. If you're going to allege that Russell was the one truly lucky along these lines, you need to give me something more than that they helped Russell "by doing something with the basketball".

Re: coach. As I've said before, it makes no sense to take credit away from a player for having a coach not be stupid. I'm not Red shouldn't get some credit here, but he's not out there playing, and he's not even the one who taught Russell how to play like this. Russell won two college titles this way for crying out loud.

Re: thinking of other possibilities. I'm all for it, but there's a distinction that needs to be made between thinking about what allowed things to come together, and what truly created the impact once things did come together. Russell could have been hit by a bus before he came to Boston, and then he wouldn't be the GOAT, but the fact that this could have happened doesn't change what Russell actually did out there.

[Doctor MJ]

First of all...how exactly does Garnett jump past anything we saw from Duncan once he gets to Boston? Are you looking at the raw numbers? Because my understanding is that you can't compare raw RAPM across years like that. I also don't understand what you're saying by Garnett clearly trumping Duncan in terms of RAPM. I don't see that happening.

Second of all...at least with regards to DRAPM (not sure about APM or raw +/- or hybrid RAPM), Mutombo doesn't really rank ahead of Duncan.

02 Duncan: +2.3
02 Mutombo: +1.3

03 Duncan: +3.4
03 Mutombo: +1.3

04 Duncan: +4.3
04 Mutombo: +1.1

After 04, Mutombo became a 10-15 mpg role player to end his career and his defensive role was thus not even close to comparable to Duncan's. Furthermore, it's not really uncommon to see role players in unique situations have a surprisingly high RAPM...in one of these years, Jason Collins was far and away the leader in DRAPM, at +6.3.

Aw raw numbers, sorry.

So, the data in my spreadsheet is all normalized for that stuff, so I'm not really worried about that any more, and I wasn't think of your numbers based on the flaws of the more raw data.

Since it sounds like you haven't heard from me what I did:

I normalized for standard deviation, and that gave me one set of numbers that I tend to label with "SD".
I then puffed them all back out using a standard deviation from a multi-year APM study from Ilardi, because APM doesn't have that study-by-study discrepancy that RAPM does. I call that the "Scaled" data.

So yeah in my speadsheet the data should be pretty well able to be compared as long as the NBA isn't going through massive changes each year:

https://docs.google.com/spreadsheet/ccc ... _web#gid=3

Also with regards to Mutombo, I'm including the RAPM data based on nba.com's recently released stats that acrossthecourt developed. It goes back so that we have an PI RAPM for Jordan's last Bull year. Pretty cool.

Re: Collins. Great point. I wouldn't be mentioning Mutombo like this if we weren't seeing the results repeat. Mutombo has 3 Scaled DRAPM years with ratings above 8.00. Garnett & Duncan don't have any. Between that and the fact we know Mutombo to be an established defensive superstar (4 DPOYs), I don't have any reason to really look at this skeptically.

[Doctor MJ]

Okay, but the numbers I quote still stand.

I like using standard deviations and other statistical techniques too, but the numbers I'm quoting here are easy to understand. One team being 12 points beyond the median while another is 8 points beyond the median is a big difference no matter how you slice it.

The deviation in a 8 team league is going to be much greater than in a 30 team league. In the context of league size, I'm more impressed by the Spurs defense.

In a 8 team league, each team has a 12.5% impact on offense/defense league numbers. In a 30 team league they only have 3.3% impact. Clearly, there will be MUCH more parity in a 30 team league, and less divergence. A couple bad offense can skew the whole metric in the 60's, but not so in the modern era where there are 28 other teams. its not a straight line correlation, but very significant.

I won't even get into the rule book differences, because the Spurs are more impressive in context of their eras.

By the same virtue, a single outlier defense can drastically alter the standard deviation as well. Goes both ways.

Right when you look at the data from '63-64, what you see clearly is that Boston's just on another planet than pretty much everyone else in term's of defense. Just go to the b-r page for it. If by using standard deviation you no longer see this clear distinction, then that's a clear sign that you're letting the analytics get in the way of your analysis.

[magicmerl]

I think people should be focusing more on era dominance and WHAT ACTUALLY HAPPENED than trying to fantasise about Russell's impact in today's era. Players are hard enough to evaluate anyway without thinking about some fantasy.

Not that I have an issue with a Jordan vote, I might change mine before the end of the thread, but I disagree with the what-if arguments about Russell.

Sent from my GT-I9300 using RealGM Forums mobile app

I agree.

If you're going to play the format shifting game, I wonder what sort of career MJ would have had if he'd had to play in the 60's, and had his legs taken out from underneath him every time he elevated for a dunk? He'd have been a perimeter shooter, or more likely, a small forward.

When people talk about 'era dependant', they really mean 'if that player played in the modern era'.

[MisterWestside]

Help take the load off? I mean, c'mon, that's just normal supporting cast stuff. I've gone into how fortunate Jordan was too have rebounding specialists like Dennis Rodman. If you're going to allege that Russell was the one truly lucky along these lines, you need to give me something more than that they helped Russell "by doing something with the basketball".

I don't have to. It's not even like I'm suggesting that Sam Jones and company were once-in-a-generation players that Russell was fitted with, by any means. Even if I called them all average Joes, their importance as complementary players adds up. They didn't get in the way of what the Celtics wanted to construct around Russell, and that still means something.

Re: coach. As I've said before, it makes no sense to take credit away from a player for having a coach not be stupid. I'm not Red shouldn't get some credit here, but he's not out there playing, and he's not even the one who taught Russell how to play like this. Russell won two college titles this way for crying out loud.

College isn't the pros, though. And again, Auerbach didn't have to go in that direction. Schayes. Petit. West. Robertson. Chamberlain. He saw and played against these guys, and knew what they could do on offense. And so did the media. Even if Auerbach wasn't the innovator, the fact that he helped the Cs stay true to the formula despite it all isn't trivial, either.

Re: thinking of other possibilities. I'm all for it, but there's a distinction that needs to be made between thinking about what allowed things to come together, and what truly created the impact once things did come together. Russell could have been hit by a bus before he came to Boston, and then he wouldn't be the GOAT, but the fact that this could have happened doesn't change what Russell actually did out there.

Nowhere in my post did I disagree with this, and I even specified that Russell was at the undisputed forefront of their success. The only thing I said was that things still fell into place for Russell to maximize said impact. Amd i didn't say anything about the era. Are any of us even talking about this man if they didn't? Or are we going to act as if the jawdropping rel. drtg numbers all came from the sound of one man swatting round, orange ball? And are we going to slot other ridiculously gifted defensive anchors in lower spots because we just assume that they enjoyed the exact same things otherwise? I would love for Garnett or Robinson to be "retro-fied", and then get dropped in the 50s and watch them (pardon the coarseness) wreck ****, because I think they were just as capable as Russell on defense.

[penbeast0]

So first off, when looking to evaluate the dominance of an outlier it's not actually good to compare them to the mean because they distort the mean. This is particularly the case in a smaller group.

So for example, in '64, if you compare the Celtics to the median, they rate a +12.8, and if you compare them to the mean of the team's not including them it's still a +12. Basically then the Spurs have only 1 year even 2/3rds as dominant as that, even before you consider that a 12-13 edge relative to efficiencies south of 100 makes for a great percentage edge that it would in later eras with higher average efficiencies.

So yeah, just peak vs peak, there's absolutely no contest here. We can talk about degree of difficulty, we can talk about luck, but just the numbers, Russell's Celtics hang'em up there off the charts.

By standard deviation, the 64 Celtics were not 1.5x more dominant than the best Duncan team. The Celtics weren't even the most dominant defensive team by that measure.

Here is a B-R post on the players who played for the best defensive teams. This post was in 2010 which coincidentally is after Duncan's 13th season (Russell played 13).


4. Bill Russell -6.08
8. Tim Duncan -5.60

http://www.basketball-reference.com/blog/?p=7239

Russell at #4 is a bit misleading. If you look at players number 1, 2, 3, and 5 . . . they were Russell's teammates. KC Jones (a great defender), Tom Heinsohn (contemporary rep was closer to Amare than Tim Duncan), Sam Jones (again, not a rep as a great defender), Satch Sanders (a good defensive specialist combo forward). What do they have in common? All played their full careers with Bill Russell. So, #4 if you don't look at the list . . . a bit misleading.

[Baller2014]

People who judge by play relative to era, when we know that era sucked, are punishing players for being born too late. It's also obviously selective. Nobody here has Mikan in their top 5 players (forget the thread limits in a sec, I mean on their own list), because they know it would be absurd to put a guy who would be a back-up in today's game in the top 5 players. People need to stop acting like Russell's Celtics dominating means so much. Russell had the best team and best organisation in a garbage era of basketball. If we transported any of the top 10 into his position it'd have been pretty similar. It was easier to win titles back then, not least of all because there were fewer teams and a handful of playoff games to play. So if you were the team with the most talent, it wasn't a surprise when you kept winning the title. Russell is a top 10 player, but not a top 5 one.

[Greatness]

Great discussion here, fpliii quickly becoming one of my favourite posters.

[An Unbiased Fan]

Okay, but the numbers I quote still stand.

I like using standard deviations and other statistical techniques too, but the numbers I'm quoting here are easy to understand. One team being 12 points beyond the median while another is 8 points beyond the median is a big difference no matter how you slice it.

The deviation in a 8 team league is going to be much greater than in a 30 team league. In the context of league size, I'm more impressed by the Spurs defense.

In a 8 team league, each team has a 12.5% impact on offense/defense league numbers. In a 30 team league they only have 3.3% impact. Clearly, there will be MUCH more parity in a 30 team league, and less divergence. A couple bad offense can skew the whole metric in the 60's, but not so in the modern era where there are 28 other teams. its not a straight line correlation, but very significant.

I won't even get into the rule book differences, because the Spurs are more impressive in context of their eras.

By the same virtue, a single outlier defense can drastically alter the standard deviation as well. Goes both ways.

Not sure I can agree here. The difference between a 12.5% impact on the league median, is massively different from a 3.3% impact.

Think of it this way. In a 2 team league, each team impacts the median 50%, in a 4 team league its 25%, in Russell's era it was 12.5%, in Duncan's it was 3.3%.

So even if they were EQUAL distribution of good/bad offenses & defenses, statistaclly, the 8 team league would have a greater divergence.

[O_6]

Just read everything, really good stuff so far. I'd love to be different and go against the grain, but I also came into this project thinking that MJ vs. Russell would be the only "real" candidates for my #1 spot.

The NBA (aka BAA) has been a league since 1947. The 3pt line was instituted in 1980. That means there have been 68 NBA seasons since 1947, 33 pre-3pt line and 35 post-3pt line. Those are the two distinct eras in the NBA's history imo. The 3pt line has changed the game so much and increased the value of outside shooting and dribble penetration, making it more of a guards game in the modern era.

Bill Russell is my GOAT pre-80 on the strength of his defensive dominance
Michael Jordan is my GOAT post-80 on the strength of his offensive dominance

The problem I have is trying to figure out which player was the GOAT across all eras.

I feel like Bill Russell was the GOAT when it comes to pure impact. He made more of a difference to his team in terms of Win Probability than any other player in NBA history imo. His defensive ability shifted the game. I also love how his athleticism was mentioned above. I find it offensive when people think Russell was a poor athlete who couldn't play in today's era.

IMO Russell was the GOAT defensive big athlete, better than Hakeem or Robinson even. He combined Robinson's straight-line explosion and coordination with Hakeem's pogo-stick jumping ability and even threw in Garnett's perimeter mobility to complete his utterly devastating defensive skillset.

But I think my real question about Russell being the GOAT has to do with that Hakeem/Robinson/Garnett/Duncan group. If Russell played in the post-80 3pt era, what kind of defensive impact was his ceiling? And would that defensive edge he had over the Hakeem/Robinson/Garnett/Duncan group be enough to overcome the offensive edge those other 4 historic defenders had over him?

Russell played 13 years in the league...

1st 13 years in league:
Hakeem ---- 24.0 DRB% ----- 5.6 BLK% --- 2.4 STL% ---- 81.7 DWS
Robinson --- 23.5 DRB% ----- 5.7 BLK% --- 2.1 STL% ---- 76.6 DWS
Duncan ----- 26.3 DRB% ----- 4.5 BLK% --- 1.1 STL% ---- 85.7 DWS
Garnett ----- 25.6 DRB% ---- 3.3 BLK% ---- 1.9 STL% ---- 65.8 DWS

All 4 of those guys were better offensive players than Russell. So for Russell to be clearly better than those guys, he would need a clear and massive defensive edge over those 4 guys OR have an offensive impact not too far from those 4. It needs to be one or the other.

I just don't see how he could be THAT much more valuable on defense. Even if Russell had a 30% DRB rate with greater court coverage than KG and better rim protection than Hakeem/Robinson/Duncan (asking for a lot), he'd still need to be a good offensive player on top of it to be as valuable overall as the other 4.

Since Russell's volume scoring skills pale in comparison to the other 4, he'd need to provide value with his finishing skills/screen setting/passing. Let's say he's a best of both worlds Tyson Chandler/Joakim Noah type of mix on offense. Providing Noah's passing with Chandler's PnR game. 12 PPG on .600TS% + 4 APG and great screen setting. That is a hell of an efficient offensive player but is that enough to trump consistent 20 PPG volume scorers who can also pass very well like the other 4 guys? And this is being really nice to Russell, who knows how his offensive game would translate to this era (.600 TS% + 4 APG may be too high).

So you have to assume that Russell was a way better defender than Hakeem/Robinson/Garnett/Duncan and you also have to assume that he could be an extremely efficient offensive player in today's era (since he doesn't have much offensive volume). I think it's possible both of those are true. Like I said above, Russell's athleticism is greatly underrated and I think he was a better athlete than the other 4 guys. But there is just so much ASSUMING and stretching of Russell's skills (especially his offensive skills) that need to be done to convince myself that Russell's game could remain at a GOAT level had he played post-1980.

With Michael Jordan, I simply don't have as many concerns. He dominated the post-80 3pt line era and he had a game that could seemingly fit in any era and thrive. His jumper is just too reliable and he is just way too athletic. He would have to adjust his game and use less dribble penetration pre-80, but he'd score in bunches and dominate across the board no matter what.

Although I believe Bill Russell provided the GOAT pure impact, Michael Jordan is my choice for NBA GOAT because he dominated his era to a similar degree as Russell while having a far less era-dependent type of impact.

[Doctor MJ]

Help take the load off? I mean, c'mon, that's just normal supporting cast stuff. I've gone into how fortunate Jordan was too have rebounding specialists like Dennis Rodman. If you're going to allege that Russell was the one truly lucky along these lines, you need to give me something more than that they helped Russell "by doing something with the basketball".

I don't have to. It's not even like I'm suggesting that Sam Jones and company were once-in-a-generation players that Russell was fitted with. Even if I called them average Joes, their importance as complementary players adds up. They didn't get in the way of what the Celtics wanted to construct around Russell, and that still means something.

You don't "have to"? Well no, I'm not your boss. I can't make demands of you. However if you're going to bring up X as something to seriously considered when talking about a player, one would presume X wasn't something true about pretty much every other player.

Russell had help. So did everyone else. So...?

Re: coach. As I've said before, it makes no sense to take credit away from a player for having a coach not be stupid. I'm not Red shouldn't get some credit here, but he's not out there playing, and he's not even the one who taught Russell how to play like this. Russell won two college titles this way for crying out loud.

College =/= Pros. And again, Auerbach didn't have to go in that direction. Schayes. Petit. West. Robertson. Chamberlain. He saw and played against these guys, and knew what they could do on offense. And so did the media. Even if Auerbach wasn't the innovator, the fact that he helped the Cs stay true to the formula despite it all isn't trivial, either.

I feel like you're just spouting cliches back at me. Here's how your point looks to me:

Friend: Hey that Eddie Van Halen is an amazing guitarist, you should get him in your band.
Bandleader: Hey Eddie Van Halen want to come to my band and play guitar?
<years later>
Doctor MJ: That Eddie Van Halen he was a great guitarist.
Mister Westside: Well yeah but that bad leader could have insisted he played the bagpipes instead.

I'm all for giving Red credit, but it's just not a zero sum thing. When you recognize a talent in someone, you deserve credit for the recognition, but you don't suck away talent credit from that person.

Re: thinking of other possibilities. I'm all for it, but there's a distinction that needs to be made between thinking about what allowed things to come together, and what truly created the impact once things did come together. Russell could have been hit by a bus before he came to Boston, and then he wouldn't be the GOAT, but the fact that this could have happened doesn't change what Russell actually did out there.

Nowhere in my post did I disagree with this, and I even specified that Russell was at the undisputed forefront of their success. The only thing I said was that things still fell into place for Russell to maximize said impact. Are any of us even talking about this man if they didn't? Or are we going to act as if the jawdropping rel. drtg numbers all came from the sound of one man swatting round, orange ball? And are we going to slot other ridiculously gifted defensive anchors in lower spots because we just assume that they enjoyed the same luck on the peripheral? I would love for Garnett or Robinson to be dropped in the 50s and watch them (pardon the coarseness) wreck ****, because I think they were just as capable as Russell on defense.

I think you'd be far better served to make your arguments for how Garnett & Robinson would do in that era based on the details of the game than playing these "what if" scenarios.

[ceiling raiser]

The deviation in a 8 team league is going to be much greater than in a 30 team league. In the context of league size, I'm more impressed by the Spurs defense.

In a 8 team league, each team has a 12.5% impact on offense/defense league numbers. In a 30 team league they only have 3.3% impact. Clearly, there will be MUCH more parity in a 30 team league, and less divergence. A couple bad offense can skew the whole metric in the 60's, but not so in the modern era where there are 28 other teams. its not a straight line correlation, but very significant.

I won't even get into the rule book differences, because the Spurs are more impressive in context of their eras.

By the same virtue, a single outlier defense can drastically alter the standard deviation as well. Goes both ways.

Not sure I can agree here. The difference between a 12.5% impact on the league median, is massively different from a 3.3% impact.

Think of it this way. In a 2 team league, each team impacts the median 50%, in a 4 team league its 25%, in Russell's era it was 12.5%, in Duncan's it was 3.3%.

So even if they were EQUAL distribution of good/bad offenses & defenses, statistaclly, the 8 team league would have a greater divergence.

Reread my post. You stated:

A couple bad offense can skew the whole metric in the 60's.

and I responded:

By the same virtue, a single outlier defense can drastically alter the standard deviation as well.

In red, you declared standard deviations as your metric of interest. So long as you have an outlier value x in sample 1 with mean mean_1 for which:

(x-mean_1)^2>(y-mean_2)^2+(z-mean_2)^2

for sample 2 (which includes y, z) with mean mean_2, if everything else is the same within the samples (or similar really, in case you want to nitpick), then yes, the outlier will impact standard deviation (again, your choice of metric) more than two values. Not something to agree/disagree about, unless when you mentioned "bad offenses", you meant they were so bad that they were outliers as well.