RealGM 2023 Top 100 Project - #18 (Dirk Nowitzki)

[One_and_Done]

Stockton basically became the starter in the 87 playoffs at age 25. I'd say we can judge him from there fairly, and all the meh playoffs I noted as examples happened after that point.

Take out 86 & 87. The Jazz were a 51.7 win team from 88 to 94, and were eliminated in the 1st round 3 times, by the 43 win Warriors in 89, by the Kevin Johnson Suns in 90, and the Sonics in 1993. They made it out of the 2nd round only twice in this 7 year stretch, before losing, including to the Blazers whose star Clyde Drexler has not even been whispered of yet. I assume the clamour for him will soon be defeaning, because somehow he defied the odds to defeat the combined might of the 2 all-time superstars who were apparently both better than him. Actually the Jazz were lucky to even make it out of the 1st round in 1992, they were trailing the Clippers (the Clippers!) 77 to 73 heading into the final quarter of the elimination game.

Imagine if 2 other stars like Curry/KD or Shaq/Kobe had a record like this is their prime? We'd rightly assume one or both of them weren't really superstars, given those results.

[OhayoKD]

Because your narrative is illogical and appears to be a backwards constructed attempt to justify the advanced metrics you favour. I don't agree that Stockton 'took a while to get going', or that he was a superior player in his mid 30s to previous years. The more sensible conclusion is that he just isn't as good as you think and the short run when the team improved in the mid to late 90s was due to other factors. There are so many guys we should be considering before Stockton or Pettit.

You’re completely misrepresenting my position. I’m saying that from Stockton’s second year as a starter through his age 37 or 38 season, the team was fantastic. They only had one season under 50 wins and they only had one loss to a team with an SRS under 6 that wasn’t one of Hakeem's championship Rockets team. They played great, not just in the late ‘90s, but over a 12 year span.

the reason we use SRS is because it "is a fair predictor" of championship prospects. Getting a high srs during expansion while losing to other high-srs non-champions isn't really notable.

[ijspeelman]

What I said there about Reggie basically applies to Stockton too RE: offense.

His assessment of these guys is annoying because there's a complete lack of analysis as to how they'd play today.

It's counterintuituve, but Reggie for example would be worse today. Why? Because when Reggie played teams didn't understand the value of 3s, which contributed to an environment where it was easier for him to get open 3s (especially the 3 years the line moved in).

Today teams sell out to stop open 3s. The way you take advantage of that is by being a deadly iso-player who blows by your defender when they get too close trying to deny the 3. Harden is a good example, with his shiftyness and handle letting him kill defenders for making the wrong decision. Miller was not much of a driver. He didn't have shake or a lightning first step or superior athleticism. In today's game he might take more 3s, but he wouldn't be able to exploit his 3pt ability to drive powerful offenses. He'd basically be a slightly lesser version of Klay. That's valuable, but it doesn't make him a star today. Ditto Stockton in relation to his 3pt shot.

He'd also be worse because he'd be taken advantage of so much more on D, where they'd bumhunt him on pick and rolls and score in him at will. In Reggie's day you had minimal movement and Reggie could chill and get some rest on a bad offensive player.

Does how they play today impact how they played in the era they played?

Its hard for me to force a translation from one era to another to evaluate players instead of taking their impact relative to the league. There is no answer in all this so you are free to evaluate players how you like, but I don't think attempting to "modernize" players in less than modern eras gets to the meat of which players are the greatest.

I think asking which player would be the best in today's league is a different question than who are the top 100 greatest players of all time.

[One_and_Done]

Impartiality is always partial.

While purporting to not punish old players, you're instead punishing modern ones for playing in a better league with better opponents. Go back far enough and there are G-leaguers who would be all-timers in those bygone eras. Stockton is not that bad; he'd be an all-star today too, but nothing more.

[rk2023]

Reggie for his time strikes me as more impressive than Stockton when stacking up primes (same story when it comes to modern translation - though I’m not putting any stock into that). As a very scalable / team-friendly offensive lead guy, his playoff scoring & ramp up in such is impressive. I don’t think he’s anything that insane on defense, but not a bad one by any means either. I would prefer him as a 1st option (and really anything after that) compared to Stockton.

I don’t really care to theorize that Reggie wouldn’t be as outlierish today when assessing his career though. If he played today, perhaps a coach would be more strict on having his feet behind the arc . Anyhow, like everybody else - it makes sense to assess Reggie for the 90s era he actually accrued value in.

[OhayoKD]

Anyhow, like everybody else - it makes sense to assess Reggie for the 90s era he actually accrued value in.

It might make sense to factor it in, but in no shape way or form is that "fairer" or more "sensible" than just absolute comparison without strings attached.

Fair is not a thing here beyond arbitrary filtering of context. There is directly comparing two players as they are, and there is indirectly comparing players by using two different foils.

Also while I'm at it, I'll point out that the "build-a-player" method advocated by some as a superior method of era-translation is more or less just the sports version of fan-fiction and completely underminines the point of absolute comparison(comparing the players as they are without setting an arbitrary line between context and what is attributable to the player in question)

[penbeast0]

Anyhow, like everybody else - it makes sense to assess Reggie for the 90s era he actually accrued value in.

It might make sense to factor it in, but in no shape way or form is that "fairer" or more "sensible" than just absolute comparison without strings attached.

Fair is not a thing here beyond arbitrary filtering of context. There is directly comparing two players as they are, and there is indirectly comparing players by using two different foils.

Also while I'm at it, I'll point out that the "build-a-player" method advocated by some as a superior method of era-translation is more or less just the sports version of fan-fiction and completely underminines the point of absolute comparison(comparing the players as they are without setting an arbitrary line between context and what is attributable to the player in question)

We use rebound rate (or approximations of it) to compare the rebounding of Wilt and Russell to modern bigs so that we can take into account the far greater amount of available rebounds due to pace and efficiency of the era. We should also use measures such as rTS% and TS Add as well as knowledge of the different contexts of different eras to compare scoring and efficiency across eras.

All these measures are imperfect as basketball is more than just numbers, but they are the best tools we have and although we may differ on some values, we do hopefully share an interest in making the best and most accurate comparison that is in our power to make.

[Doctor MJ]

BPM is literally determined by a regression on which stats lead to winning most. It's the definition of a stat that translates into impact/winning. .

So I would strongly disagree, but I think having the discussion about this is really important - and it's within the realm of possibility that something has been innovated that I'm not aware of.

Big picture, I see these domains Holistic Basketball Player Evaluation Metrics:

Box Score Weighted (Production)
Scoreboard Correlated (Impact)
Player Tracking Weighted (Record? Action? Track?)
Biographically Weighted (Stereotype? Measure? Scout?)
Hybrids

I'd say all of these domains have potential value, and the Hybrids possibly most of all...though I've long been known about the dangers of Hybrids to serious basketball analysts.

I'll also say that I'm super-high on continued progress with Player Tracking Weighted measures. I see a tremendous amount of potential there.

But right now we're mostly use Production, Impact, and Production-Impact Hybrid Holistics.

Anyway, speaking with this schema in mind - which I'd enjoy discussing further - the really relevant point to me with regards to our disagreement is this:

I'm not ever comfortable using Production data as substitute for Impact data.
I may effectively weight Production more heavily the less Impact data I have, but it doesn't fill the niche of Impact data, that void remains and shows its affect with greater uncertainty in my final assessment.

Why do I feel strongly about this? Honestly, because I've seen the scale of the problem in the case of someone like Steve Nash.

Impact stats have led me to conclude that Nash was the clear cut top offensive player in the league for many years...but he never leads the league Offensive Win Shares or BPM over on bkref. So obviously I think that means they underrate him, but the more important point to me is that they underrate him specifically because their are dimensions of impact that they simply don't have statistical access to...which means we have to bring that our adjustment ourselves, and we do so cognitive tools that are effectively heuristic approaches that emulate the other domains listed above.

And of course, the defensive problems are bigger. Measures like these think Magic Johnson was a better defender than Michael Cooper, because the things Cooper was tasked to do aren't what the box score tracks.

A guy like Nash is a limited defender, but he's not getting so few steals because he's doing nothing out there, but rather because it's not his job to be the one taking risks on defense. His job out there is to work within the team defense, and in Phoenix he really wasn't a clear-cut negative defensive impact guy. Nor should we think that because the Suns defenses were never elite that that means they couldn't have made a good defense with Nash on the court. During the brief stint that Kurt Thomas was the starting center for the Suns, the team had a legit top-third defense, and it's not like Thomas is known for being an amazing defender. He was just a lot better than Amar'e or Old Shaq.

This then points to what I'm looking for with Player Tracking. When we can get it to the point where it can effectively predict Impact by Tracking, we're going to have something incredibly powerful...but I really don't think anything all that close is possible based solely on the Box Score.

I feel like I used to be a lot more skeptical on BPM, but I've actually been very surprised over the years how well it correlates to actual impact data. I'm not saying it has zero blind spots whatsoever and elite aggressive passing is one of them, but on the whole I feel like it does surprisingly well. Obviously, having stuff like EPM and RAPTOR is best, and having say a good PI RAPM would be next best, but in general I think BPM does fill in the gaps pretty well. And in the case of Stockton where we already have 7 years of elite impact data for him at a ridiculously advanced age, I don't see anything wrong with taking it at face value when it says he was a consistent top 5 player.

Where we do have actual RAPM data, we come up with:
Stockton: 5.6 (off), -1.6 (def), +7.2 (overall)
Nash: 5.1 (off), 0.8 (def), +4.3 overall

IDK, I feel pretty comfortable with the conclusion that Stockton was at least comparable to Nash at peak and once you get away from the top three years for each, Stockton has a pretty comfortable lead each of the rest of the seasons from 4th through 18th.

"surprisingly well" is an interesting description. I'm well past a decade being disappointed at how inadequate box score stats are at capturing some really, really impactful things, but I suppose in the end, if we're simply disagreeing because of expectations, we can write a lot of the differences in our qualitative descriptions off to possibly being less dramatic than they first appear.

Re: "I don't see anything wrong with taking it at face value." Eh, well, I wouldn't have anywhere near that kind of confidence in it.

Re: "where we do have actual RAPM data, we come up with". A number of issues here:

1. There is not one standard RAPM, and there are numbers that don't look like what you're describing. Just googling right now I found this site Cryptbeam (wondering who it is), and by their PI RAPM beginning in 1998, Stockton's top year gives him a +5.25 number, while Nash's top number is +5.96.

This pertains to why I've pulled back from quoting the stat as much as I used to. If metrics by the same label can vary this much, it makes me be cautious with what I take to be a "clear cut" advantage.

2. I think it wise to be cautious around guys who have big +/- numbers in limited minutes. It doesn't mean the player "isn't that good" necessarily, but the Jazz chose to play Stockton in limited minutes for most of the minutes being used to determine RAPM, but not most of the minutes of his career as a starter.

We have 4 years from '93-94 to '96-97 where we can evaluate the On & On/Off of Malone & Stockton during Stockton's prime-level minutes (35-ish MPG), Malone had the advantage in 3 of those 4 seasons. Hence, I'd say that Stockton's apparent impact edge begins when he starts playing less...which should make us wonder if we're really talking about superior impact, or if what we're seeing is Stockton being allowed to play in more favorable situations than back before when he was playing minutes like a star.

3. When I see numbers where Nash doesn't stound out much compared to another guy by Offensive RAPM, it sometimes challenges realism to me. Using bkref's numbers:

In '03-04 the Suns had an ORtg of 101.4.
In '04-05 Steve Nash had an on-court ORtg of 120.3, well beyond his teammates, and way beyond anyone in the league the previous year.

That's Ruthian stuff, but Nash's PI RAPM numbers for the year come back meh because the #1 offense he was on in Dallas in a role with less decision making primacy didn't drop off enough without him.

I'm honestly not saying that that means there's something "wrong" with RAPM that should be corrected, I'm just saying that there's more to judging context than can be gleaned from linear regression.

Regardless of how important we should take RAPM when comparing these guys, there's not a dramatic jump or such an outlier team peak like this that we've seen from Stockton, and I think that's something we need to not think we can assume to be "covered appropriately" by an algorithm such as this.

[One_and_Done]

Iggy I have a serious question. Does your approach to ranking players actually require you to watch the games? I honestly can't tell. It feels like watching games is superfluous to your method, because someone could just run numbers through a computer to get the results. I'm with you on ranking the modern guys higher, but there's got to be more to it than a computer formula. This isn't a video game.

[rk2023]

Iggy I have a serious question. Does your approach to ranking players actually require you to watch the games? I honestly can't tell. It feels like watching games is superfluous to your method, because someone could just run numbers through a computer to get the results. I'm with you on ranking the modern guys higher, but there's got to be more to it than a computer formula. This isn't a video game.

Right message, but aren't you heavily into ORTG and /100 stats & so on more or less in a vacuum when getting your points across?

[MyUniBroDavis]

Iggy I have a serious question. Does your approach to ranking players actually require you to watch the games? I honestly can't tell. It feels like watching games is superfluous to your method, because someone could just run numbers through a computer to get the results. I'm with you on ranking the modern guys higher, but there's got to be more to it than a computer formula. This isn't a video game.

Extremely based take

Watching basketball is important, genuinely feel people forget this because they don’t trust their own eye test to discern things. A ton of data is an approximation of what goes on, on the court, without telling you the why or the how of things. The issue is more we can’t wTch every game than we can’t understand what’s going on in a game, but you gotta watch dat lol

If you can’t explain why something happens or how something happens it’s not really analysis, just link the data link then lol

[iggymcfrack]

Iggy I have a serious question. Does your approach to ranking players actually require you to watch the games? I honestly can't tell. It feels like watching games is superfluous to your method, because someone could just run numbers through a computer to get the results. I'm with you on ranking the modern guys higher, but there's got to be more to it than a computer formula. This isn't a video game.

I mean, I watch a ton of games and it obviously factors in sometimes, but is it “required”? No. I’ve seen very little film on West and Oscar for instance, but I feel like through stats and scouting reports and such I can come to a pretty decent opinion. In general, it’s fair to say that I trust the numbers much more than my eyes even on guys I’ve watched extensively.

Sometimes, something will really stand out like Kobe’s selfishness or the way that Jokic controls the game with his passing or Chris Paul’s defense and then I’ll tend to weight those factors a little more. Or in the case of Mikan I watched film to evaluate his defense because so little reliable information exists. But Dirk was my favorite player for over a decade and I can’t say there’s anything I picked up that would cause me to rate him any higher or lower than what the advanced stats would suggest.

If anything, it kind of goes the other way. Knowing KD’s RAPM failures means that when I watch him play, I tend to pay more attention to his weak passing and playmaking and am like “ha ha, there he is failing again just like the nunbers would suggest!” When I watch Derozan, I think “LOL, another **** mid-range shot”. So yeah, I’d say the way that I rate the players from their numbers causes me to watch the games differently more than the other way around.

[One_and_Done]

So basically you're telling me we could replace you with a computer formula and it would make no difference. You like watching games, but it's basically not relevant to your rankings so you could just as easily not watch them and your rankings would be the same.

[iggymcfrack]

So basically you're telling me we could replace you with a computer formula and it would make no difference. You like watching games, but it's basically not relevant to your rankings so you could just as easily not watch them and your rankings would be the same.

No, because there’s constantly new data coming and we each have to choose how to weight all the different pieces of data. That’s what’s fun about a project like this is learning things we didn’t learn before. The further back in time we go, the more incomplete our data is and the more we have to fit these little puzzle pieces together trying to figure out who was actually good at different things.

Like if we’re judging who was the best player in 2023, yeah, you can pretty much make a computer formula for that. The level of data is overwhelming and the computers see every minute of every game which a human couldn’t possibly hope to match. If you’re trying to choose how to judge players across eras, if you’re trying to figure out how valuable Jerry West’s defense was, if you’re trying to figure out how much KG’s impact was depressed by the front office in Minnesota, there’s tons of room for subjectivity. The fun of making rankings is about fitting together these pieces of incomplete information, not trying to judge an entire career through the snippets we see on our televisions.

[70sFan]

Iggy I have a serious question. Does your approach to ranking players actually require you to watch the games? I honestly can't tell. It feels like watching games is superfluous to your method, because someone could just run numbers through a computer to get the results. I'm with you on ranking the modern guys higher, but there's got to be more to it than a computer formula. This isn't a video game.

I think this is a critical question to all of the participants, because I don't think you can say anything credible about given player without trying to watch as much as possible. During my tracking project (that is not finished yet), I realized how little I know about the nuances that differentiate the best bigs in the league history. If someone never watched 1970s basketball game (or 2020s game, such "fans" also exist unfortunately), I have a hard time taking such fan's opinion too seriously.

I don't think it's just a question for Iggy, but for all of us who contribute to this project. I often see a lot of strong opinions about players most people likely never watched in a serious amount of time, my recent experience showed me that you can't have enough humility.

[70sFan]

Like if we’re judging who was the best player in 2023, yeah, you can pretty much make a computer formula for that. The level of data is overwhelming and the computers see every minute of every game which a human couldn’t possibly hope to match.

I strongly disagree with this. As far as I know, we (or teams) don't have any high level AI software for analyzing games that would surpass scouting ability of the best scouts in the world and that's the only way you could theoretically keep the analysis around similar level (not sure if it's possible though, basketball is a very complex game).

If you think that running a formula will give you a good understanding of who is the best player in the world, then I would only say that people's faith in statistics reached a level where we ended up in a strange place... and I'm saying this as a scientist.

The fun of making rankings is about fitting together these pieces of incomplete information, not trying to judge an entire career through the snippets we see on our televisions.

Well, in most cases we can watch way more than just the snippets. I remember when I joined the RealGM, there were so little information and footage about Artis Gilmore - probably less than 5 full games on YouTube. I decided to get a better idea of how he played and now I possess the biggest collection of his games online, probably outside of the NBA only. I did this as a teenager from Poland, so I don't think we have to limit ourselves to snippets from TV.

[SpreeS]

Stockton basically became the starter in the 87 playoffs at age 25. I'd say we can judge him from there fairly, and all the meh playoffs I noted as examples happened after that point.

Take out 86 & 87. The Jazz were a 51.7 win team from 88 to 94, and were eliminated in the 1st round 3 times, by the 43 win Warriors in 89, by the Kevin Johnson Suns in 90, and the Sonics in 1993. They made it out of the 2nd round only twice in this 7 year stretch, before losing, including to the Blazers whose star Clyde Drexler has not even been whispered of yet. I assume the clamour for him will soon be defeaning, because somehow he defied the odds to defeat the combined might of the 2 all-time superstars who were apparently both better than him. Actually the Jazz were lucky to even make it out of the 1st round in 1992, they were trailing the Clippers (the Clippers!) 77 to 73 heading into the final quarter of the elimination game.

Imagine if 2 other stars like Curry/KD or Shaq/Kobe had a record like this is their prime? We'd rightly assume one or both of them weren't really superstars, given those results.

If Stockton gets into TOP25, then Malone/Stockton as duo is the only one duo who won anything. And they played together whole career

Lebron/Wade 4 season 2 Chips
KAJ/Oscar 4 season 1 Chip
KAJ/Magic 9 season 5 Chips
Duncan/Robinson 6 seasons 2 Chips
O'Neal/Kobe 8 season 3 Chips
O'Neal/Wade 3 season 1 Chips
Wilt/West 5 season 1 Chip
Moses/Erving 4 season 1 Chip
Curry/Durant 3 season 2 Chips

Malone/Stockton 18 season 0 Chips w/o injuries

[OldSchoolNoBull]

Just a couple small nits to pick.....

Vote: Dirk Nowitzki

Again, between Dirk and Malone, this is fairly easily Dirk. It is true that Malone's box RS numbers look better:

34.4 per 100/13.9 per 100/.205 WS/48/5.1 BPM/+4.3 rtS
vs
31.9 per 100/11.6 per 100/.193 WS/48/4.5 BPM/+3.3 rtS(though Dirk's lousy final season drags that rTS down from +3.96, much closer to Malone)

But that's where his statistical advantages end. Dirk is superior in the playoffs:

Dirk: 32.8 per 100/12.9 per 100/.188 WS/48/5.9 BPM/57.7% TS
vs
Malone: 32.6 per 100/10.7 per 100/.140 WS/48/4.1 BPM/52.6% TS

Firstly, you noted Malone's DREBs vs Dirk's TOTAL Rebs. Malone's full-career TREBs per 100 in the playoffs is 13.9.

That was an honest mistake, my bad.

But secondly, I don't know that we should be comparing Malone's full-career playoffs stats here, at least without some disclaimers or mentioned details.

For example, you note Dirk is sort of being penalized in the rs samples by his final season, but do not do the same for Malone in the ps samples.

Dirk's teams didn't make the playoffs until '01, a season which could reasonably be called his prime (or at least extended prime). And they didn't make the playoffs any of his final three seasons. So basically NONE of his pre-prime, and relatively little of his post-prime is included in his playoff sample.
Malone, otoh, has ALL of his pre-prime and ALL of his post-prime included in his ps sample, because he never missed the playoffs. Additionally, the league avg TS% in Dirk's playoff sample years is just marginally better than in Malone's.


So if I, for instance, included only '88 thru '02 for Malone (a 15-year sample that includes all of his extended prime, plus at least one non-prime season [now similar to the 15 years sampled for Dirk]), I note Malone is about +0.04% rTS vs Dirk's +4.46% rTS.

It's still a very significant edge to Dirk; but where it previously looked >5%, now it's <4.5% (a small nudge, almost identical to the difference betweeen 3.3 and 3.96).

If I did similar with his per 100 numbers, I note that Malone ('88-'02) in the playoffs averaged 34.9 pts, 14.6 reb, and that his BPM and WS/48 are +5.0 and .156, respectively.

I could expand the Malone sample slightly (to include either '03 or '87), and that would damage his figures somewhat, but still the gap would not be what you've shown above.

Well, Dirk's last season, in terms of rTS, is a massive outlier. It's -9.1, where the next better one in his career is -2.3, two years earlier, so that one season is 6.8 points worse than the next best one for Dirk. Malone's worst rTS is his rookie year - -3.7, and his next worst is his sophomore year, -0.2. You can take different samples of Malone's career if you feel it makes it more fair, but the reason I only said that about Dirk is because I believe his last season is more of an outlier than any season of Malone's career. But reasonable people can disagree.

And here are their full RAPM sets(Malone's first three from Squared):

Dirk:

Code: Select all

1998-99   0.73   -0.92   -0.2
1999-00   1.9   -0.49   1.41
2000-01   1.8    2.4    4.2
2001-02   4.63   1.17   5.81
2002-03   4.93   2.41   7.35
2003-04   5.81   1.45   7.26
2004-05   3.73   2.17   5.9
2005-06   4.3    1.29   5.59
2006-07   4.88   0.88   5.76
2007-08   4.83   0.79   5.62
2008-09   4.31   0.89   5.2
2009-10   2.94   1.6   4.54
2010-11   4.19   2.67   6.86
2011-12   4.66   2.49   7.15
2012-13   4.5   2.21   6.71
2013-14   3.86   1.3   5.17
2014-15   3.53   -0.3   3.23
2015-16   1.78   1.41   3.18
2016-17   -0.62   0.22   -0.4
2017-18   -0.26   -0.08   -0.34
2018-19   -0.02   -1   -1.01

Malone:

Code: Select all

1987-88   3.44   -2.36   1.07 
1990-91   1.18   -2.01  -0.83
1995-96   2.86   -2.56   0.28
1996-97   4.28   0.14   4.42
1997-98   5.17   0.14   5.31
1998-99   5.02   -0.44   4.58
1999-00   5.54   -1.42   4.12
2000-01   2.5   -0.3   2.1
2001-02   1.93   -1.85   0.08
2002-03   2.97   -2.52   0.45
2003-04   -0.03   0.27   0.24

You tell me whose RAPM sets look better. Particularly on the defensive side.

It's important to note that those RAPM figures by Squared are from a VERY limited sample of those three seasons (reliability could be iffy). That in mind.....

Dirk, in his "extended prime" I previously established [prior post] as '01-'14, averages a +4.24 RAPM.

If we go ahead and use Squared's figures, despite sample size issue, focusing on his extended prime as previously established in prior post as '89-'02 for Malone, he averages a +3.56 RAPM for the eight seasons within that span that we have any data for.

If, otoh, we choose to use his rs AuPM for '94-'96 (rs only, but MUCH larger sample size [whole season] than Squared's data), we now have data for 9 seasons within his extended prime, and they average out to +4.46.

Or as a third option, we could use ONLY those [six] extended prime seasons we actually have full-season (including playoffs) RAPM, in which case Malone averages out to +4.07.

You make a fair point about Square's RAPM numbers being based on small sample sizes and then those RAPMs subsequently dragging down Malone's averages.

However, two of your alternate methodologies still end of with Dirk on top, and the one that doesn't, I'm a bit skeptical of because it seems like mixing two different metrics(RAPM and AuPM).

And I maintain that Malone never defeated teams in the playoffs as good as the 2006/2009 Spurs, 2006 Suns, 2011 Lakers, 2011 Thunder, and 2011 Heat.

idk, I think the Malone Jazz defeated a number of teams better than the '09 Spurs.
And I think someone like the '98 Lakers is at least debatable against one or two more of the teams you mention above.

I'd also note that Dirk's Mavs never had to face a team as good as the '97 Bulls, or arguably the ‘96 Sonics (at least in a year they were actually contending).

I probably am overrating the 2009 Spurs based on reputation, because it's still Duncan/Parker/Manu two years removed from a title, but yeah, based on the numbers(SRS, etc), they weren't as good as some of the teams Utah beat.

Actually, looking at the numbers, maybe my statement was a little hyperbolic. I'd forgotten that the 98 Lakers had a 6.88 SRS(higher than any of the teams I mentioned that Dirk beat), and the 91 Suns were 6.49 and the 96 Spurs 5.98. The 09 Spurs are 3.36 and the 11 Thunder are 3.81.

The gap isn't as big as I had thought, but I do still think Dirk has the edge in this category.

[OhayoKD]

Vote

1. Dirk

-> Led a team to a strong championship run
-> Great longevity
-> Succeeded with a wide variety of casts/co-stars
-> Playoff Riser with solid rs signals relative to the competition

(will elaborate on how his team-success compares to Malone's below)

Undecided on an alternate. Leaning Julius.

Nominate: Giannis

Spoiler:

Yeah not seeing it.

As for nomination...

For Jokic, I do think longevity is a bit of an issue. I said before the project began that while I don't put a lot of weight on longevity, less than ten years would give me pause. It's why, even as someone sympathetic to era-relativity, I only started championing Mikan a few threads ago rather than at the very beginning, because he only has, what, eight years(and two of them technically pre-date the NBA). Jokic is at eight years right now. He's at the beginning of a five-year contract right now. If he keeps playing at the level he's been for the duration of that contract and adds another ring or two, then I think he'll have a legit argument for Top 15 status.

For Giannis, it's not just the lack of longevity. I just think he's had some concerning playoff failures. The Bucks blew a 2-0

Yeah, I think this might be a good time to refresh everyone's memories of what each of these players "failures" looked like.

Keeping in mind Giannis has made 8 post-seasons to Jokic's 5

Giannis and the Bucks in the playoffs
2015: lose to the bulls as a role-player

2017: Giannis becomes a fringe superstar, team sees +3 srs improvement and plays a razor-close series(<1 ppg, 6 games) vs the +3.65 srs opponent(Giannis puts up strong offensive production)

2018: Giannis is a fringe MVP candidate, team mantains in the rs, and then playes an even closer series vs +3.2 srs Boston who nearly make the finals after beating the near +4.5 srs Sixers

2019: Giannis gets a not bad coach for the first time in his career and breaks out as a historically strong MVP winner as the Bucks jump by 8 points to post a historically remarkable +8 srs team(almost never happens in non-expansion periods) despite a cast that plays at .500 without him form 19-20(and marginally above from 21-23. That team improves to +13.75 in the playoffs on the back of a big defensive improvement. They are merely +7 in that oh so bad 6-game(1 ppg) loss to a coasting Toronto side which saw a cast capable of 60-win basketball add Kawhi Leonard, aka, clutch Durant, aka "resiliency king". In the conference finals Giannis's offensive production falters against one of the best defenses ever but he also puts up one of the best defensive performances ever to push a toronto side about as good as anyone Jokic has ever faced and far better than any team Jokic has ever beat to the brink(double-overtime and giannis fouling out prevented a 3-0 defecit).

2020: Giannis has one of the very best regular seasons ever(arguably better than any regular season from certain players who have already been voted in) and the Bucks post a +9.41 SRS(basically unheard of in non-expansion periods) with a team that plays average basketball without Antetokounmpo. Team collapses defensively in the bubble and are upset by the eventual finalists despite Giannis's offensive production improving from last year as their defense is torched by Miami. There is injury context with Giannis eventually missing a game and 3 quarters.

2021: Giannis coasts as merely a top 3 regular season player in the regular season and the Bucks post a +5.6 SRS(4th in the league) with a team that is a bit above .500 without him. The Bucks again get significantly better in the playoffs on the back of their defense and Giannis is good to great on both ends throughout as Giannis becomes one of the few players to win a championship...
-> without a 2nd superstar
-> without perennial all-star
-> without "help" that is significantly > .500 without him
-> without a strong playoff coach

The competition is fairly weak, but so was the support, and ultimately it's topped off with Giannis posting one of the greatest performances ever against a very good team on both ends of the floor

2022: Giannis is again, merely a top 3 regular season player, and the Bucks regress to +3(7th best) with the big-three missing a significant number of games. Bucks are(opponent-adjusted) more than +12 against the Bulls with Middleton and take a near-champion to 7 without a middleton in a not that close series(+8 point differential). Overall Bucks improve dramatically. again, on the back of their defense.

2023: Every contender is coasting and Giannis is again merely a top 3 regular season player as the Bucks post a 3rd best +3.61 SRS despite Middleton missing a bunch of games. Against Miami, Giannis misses almost half the series and is injured throughout. Consequently, the Bucks defense collapses as they lose to the eventual finalists(again)

8 postseasons total, 7 as a superstar, and the Bucks underperform twice and overperform 5 times despite a deeply flawed postseason coach, a cast who generally falls off in the playoffs(shooting especially). Both underperformances have injury context and when they lose, they are mostly losing to champions or finalists,

Now let's do Jokic:

Jokic and the Nuggets in the playoffs

2019 Jokic is a fringe MVP candidate and Nuggets see a 2.5 SRS improvement to post a strong +4.13(7th best). They win a razor-close series against the +1.8 Spurs(7 games, 1 ppg) and then lose a razor-close series(7 games, actually outscore by 1 ppg) against the +4.4 SRS Nuggets who proceed to get destroyed in a sweep against a losing-finalist. You may recall the champion that year was that Raptors side that just about survived Giannis.

2020 Jokic is again a fringe MVP candidate and the Nuggets regress to +2.5 thanks to injuries to Jokic's best teammates. In the playoffs they get lucky against the +2.5 Jazz winning in 7 despite getting outscored by 3-points a game. They then upset the +6.6 Clippers in a close series(7 games, <1ppg) before getting thumped by the eventual champs(5 games, 4 ppg). You may recall the Heat, without their leading scorer and with their defensive anchor hobbled, were the only team all playoffs to take the Lakers to a 6th game.

2021 MVP Jokic leads a +4.8 Nuggets side(6th best) despite a team that is outright bad without him. They proceed to win a razor-close series against the +1.8 Trailblazers(6 games, actually outscored) and are then obliterated in a sweep against the +5.5 eventual finalist Suns(15! ppg). Those suns would lose to...checks notes...Giannis's Bucks. Nuggets are bad without Jokic

2022 B2B MVP Jokic leads a +2.15 Nuggets team(injuries play a big-factor) and then is thumped in 5 by the +5.15 eventual champs(8 ppg).

2023 Should have been B2B2B MVP Jokic, with a team that is still bad without him in 13 games, leads the Nuggets to a +3 srs(6th best in the league). Against a relatively weak field(though everyone coasting undersells the competition) they are dominant in the postseason going 16-6 with a m.o.v of +8. This is an all-time dominant run, but it also coincides with dramatic cast elevation and unusually favorable injury context(like Milwaukee's 2021 Run). Nonetheless as a singular note it has a decent case against anything Giannis has done considering
-> team is bad without him(in the regular-season anyway)
-> unusually dominant
-> One-superstar(Murray is close)

5 postseasons total, I think it's fair to say the Nuggets overperformed in 2 and underperformed in 2. A weaker trackrecord than Giannis's Bucks despite
-> a better playoff coach
-> teammates generally elevating(Murray arguably outplayed Jokic in 2020)

The Nuggets are also flatly a far worse regular-season and postseason team getting destroyed when they face eventual finalists and champions which Milwaukee only really do if Giannis gets hurt. When the Nuggets faced a 2019 Raptors-calibre opponent, they were crushed despite Murray playing like a superstar. The Bucks have never suffered a defeat like the Nuggets did against the suns despite running into an eventual or defending finalist each of the last 5 playoffs.

Giannis's Bucks have also posted 2 regular-seasons where their srs nearly doubled any of the suns and one of those regular-seasons was followed by post-season improvement and a tough fight against the type of team the nuggets tend to get dominated by.

All considered, saying Giannis has "Playoff issues" and Jokic doesn't seems like you're applying a gigantic double-standard because Jokic id a one-way player while Giannis is a two-way one. Just like when we act like Jordan was "perfect" any-run he posts sub-2009 Lebron box-aggregates or when we act like Shaq is more "unstoppable" than two-way bigs because defense doesn't matter.

Excepting their championship years, Giannis has led far better regular season and playoff teams, and has also has a significant longetivity advantage, while elevating more often. And while Jokic's regular-season impact looks great(like Giannis)...

So, I'm very convinced by the Jokic > Giannis arguments. Time to change my nomination as well as my all-time list.

[/quote][/quote]
Since you've used playoff on/off as justification before...



And yet Giannis is the one with "issues" apparently

Do not plan on voting him soon as my criteria is era-relative and values longetvity but at this point I am nominating based on who I think has the most viable arguments and for Giannis

-> arguably top 15 resume with 2 mvps and a lone-superstar championship with an FMVP to boot to go with multiple all-time regular season teams
-> better longetivity than another top 15 resumes
-> arguably best player in the most talented version of the nba
-> all-time peak/prime, era-relative or absolute
-> one of the most versatile players ever and one of a handful who has carried a contender as his team's best defender, best playmaker, and best scorer.

The main candidates for induction here look to be Dirk and Malone so I'll offer something comparative.

Was planning to respond to this in the #17 thread, but Malone is a major candidate here so may as well bring this over

as for the stockton argument, i totally agree he absolutely cannot be in the top 30 or even top 35 if malone is going to be in the top 20.

stockton over dwade in the last project just seems like a terrible misapplication of "longevity" over peak results. stockton could still be playing today and i wouldn't put him over dwade. one guy can throw a team on his back and be an alpha on a championship for like 5 healthy playoffs (but his supporting cast wasn't good all 5 years) and could still be the second best player on other championships for a few other years while the other guy is just proving that slow and steady does not win the race in the nba.

To that, I want to point out that they did anchor some REALLY good teams; teams that win the title a lot of years in NBA history.

I'll come back to this, but even taking srs derived deviations at face-value(and tbf, by moonbeam's work, deviations are more predictive of titles than then the raw marks)...

Anyway, I bring that up to note that the '96 Jazz are ranked as the 35th-best team of All-Time there; and the '97 Jazz are the 64th-best all-time. But the '96 squad just happened to lose in 7 games to the 75th-best team of all-time; though even if they’d won, they had in their way the 2nd-best team of all-time; meanwhile the '97 squad had to face the 14th-best team ever.

Cool. By san's method, the 2006 Mavs lost to the 85th best team ever(well actually "post-shot clock" but whatever) in a razor close 6 game series(1 ppg, were a massive comeback away from going 3-0 up, last two games were lost by a combined 4 points) amid some officiating controversy. They also took the 7th best team ever to 7 in another really close series(m.o.v was 2) in 2014.

They also took the 37th and 53rd best teams ever to 6 in 2003 and 2005, but that wasn't so close by differential(5ish, and 6ish mov respectively).

None of Dirk's teams quite match the 96 jazz but both the 2006 and 2011 mavs come pretty close ranking 41st and 42nd respectively.

You might notice that the team getting the best of Dirks' Mavs is a...conference finalist who...didn't lose to eventual champions. By comparison, the 06, 14, and 2003 Mavericks all lost to the team that would win it all and all 3 can say they gave the best team in the league the best run for their money(by games or m.o.v). That last bit is not something you can say for the 97 Jazz as Reggie's Pacers played the bulls significantly closer by differential while also forcing a 7th game. That leaves Utah with just one year, 98, where they can say they proved they were 2nd-best.

And then we have the 2011 Mavericks who...were the best team in the league not only whacking everyone in their own conference, but also

-> beating the team that whacked everyone in their conference before proceeding to win back to back titles in spite of Wade's deteriorating knees
-> sweeping a two-time defending champion
-> crushing a better version of the OKC team that took that champion to 6 the year before

Their SRS(excepting 2013), may not be so gaudy, but as far as "getting close frequently" goes, I'd say the Thunder had the Sonics beat. And really, that's the main issue here...

Because your narrative is illogical and appears to be a backwards constructed attempt to justify the advanced metrics you favour. I don't agree that Stockton 'took a while to get going', or that he was a superior player in his mid 30s to previous years. The more sensible conclusion is that he just isn't as good as you think and the short run when the team improved in the mid to late 90s was due to other factors. There are so many guys we should be considering before Stockton or Pettit.

You’re completely misrepresenting my position. I’m saying that from Stockton’s second year as a starter through his age 37 or 38 season, the team was fantastic. They only had one season under 50 wins and they only had one loss to a team with an SRS under 6 that wasn’t one of Hakeem's championship Rockets team. They played great, not just in the late ‘90s, but over a 12 year span.

the reason we use SRS is because it "is a fair predictor" of championship prospects. Getting a high srs during expansion while losing to other high-srs non-champions isn't really notable.

Exactly two teams have hit +10 srs(regular-season) in lieu of expansion. Four have approached it(one of those being OKC...). I point that out not to denigrate the talent pool but rather to point out that the distribution of said talent in the league led to the SRS of top teams being inflated relative to their championship prospects, the thing we use srs as an indicator for.

This effect is somewhat muted with playoff-weighted analysis and standard deviations, but that proxy is still the starting point and we should mind when it is not actually lining up with the real thing.

Utah did not win. They also were not coming as close as frequently when they were not winning, even if we compare each team relative to other non-champions from their era.

That does not change whether you keep things strictly year to year...or broaden the scope to the "era".

I am open to arguments for Malone's lesser success being a result of support rather than what he offered, but saying the 96 Jazz were "good enough to win a title most years" is about as convincing to me as arguing that 2000 Portland was tougher than the near-dynasty Knicks and the Wilt-West Lakers.

Imo, this is just not a good use of data. And on that note...

Iggy I have a serious question. Does your approach to ranking players actually require you to watch the games? I honestly can't tell. It feels like watching games is superfluous to your method, because someone could just run numbers through a computer to get the results. I'm with you on ranking the modern guys higher, but there's got to be more to it than a computer formula. This isn't a video game.

I mean, I watch a ton of games and it obviously factors in sometimes, but is it “required”? No. I’ve seen very little film on West and Oscar for instance, but I feel like through stats and scouting reports and such I can come to a pretty decent opinion. In general, it’s fair to say that I trust the numbers much more than my eyes even on guys I’ve watched extensively.

Honestly, I'd agree "required" is strong, especially when discussing players we have little to no footage of. I'm always going to be skeptical of someone who does not bring up something in the way of evidence(defining it here as: observable to different independent parties, logically leads to the claim it is being used to support...) but is matter of factly sharing their impression of someone's game. Especially. if they jump 600 feet and follow that impression with a broader assertion they haven't actually connected the dots for("scoring>defense -> kg wasn't a first option -> Malone>Garnett!!!")

What is valuable or required really depends on what is already there and what is being argued. It is pretty important to watch the game(and maybe actually take some notes) when you assert that player's a rebound totals are overrated because x and people have already watched this players rebounding and are coming to a different conclusion.

It is not --as-- important when you are comparing a two-way big with a guard because all the eyetest can really do here is tell you there are distinct advantages for either(which no doubt plenty of people already have picked up on and can inform you of). It's also not so important when we have an abundance of data for player a in a variety of contexts and people are arguing that film shows they don't hold up so well in situation x y or z even though the data suggests they did.

Additionally if the "eyetest" amounts to "player a was a better man defender so he a better defender, boom" the question of whether you have watched or not watched really doesn't even matter as what the watcher seemingly discerned isn't all that useful for what they're concluding.

In that way, evidence has an advantage...

-> if it is actually evidence
-> and if it is being used properly(and consistently)

And here's where I think the feathers are really ruffled. Citing playoff on/off for one guy and ignoring when it's matched by another. Ignoring what larger samples say when they suggest different things than smaller ones. Pretending "accuracy" on a larger scale renders potential biases moot, or misconstruing even the broad-scale accuracy because you aren't tracking what's actually being measured.

Whatever subjective weighting is being applied in order to "better fit its stats to rapm", it is only capable of fitting what it has access to, and consequently is less accurate than stats which directly input rapm. And even if the makers were perfectly fitting weightings to what "winning" would support, they are still limited by what they actually have access to,

We can literally see this in the specifics of their formula. Stockton racked up lots of steals juicing his DBPM. Steals are, per 538, "hard to replace". But do they actually move the needle in terms of winning? Well, per Berkley,team-wide steal counts have little to no effect on how much an nba team scores or doesn't score and what effect it does have is tilted to offense. Yet, largely due to a lack of options, steals play a big-factor in a player's DBPM.

RAPM is very useful. It is also an approximation of what happens that mainly offers value over extended samples(where your eye should be towards replication of high-marks, not exactly how high the marks go). What actually happened is Stockton's primacy went down as the Jazz got better. As did his minutes. And as did his assists.

And here, looking at the film might be useful. Because an assist is not inherently very valuable. and I'd say many of Stockton's qualify as examples of this:

Spoiler:

Stockton in today's environment would almost certainly shoot a great deal more. The drive and dish, pass first guard that we were taught was the ideal PG growing up in the 70s just isn't a template taught today. Compare Stockton's per possession willingness to shoot with that of his peers rather than with today's guards to get a better feel for where he would fit on the spectrum of modern PG's willingness to shoot. And remember that he is an extremely efficient shooter with outstanding 3 point range even growing up then, maybe more of an outlier in that regard than Nash was relative to Nash's peers a decade later.

This assumes there isn't a point of diminishing returns.

The other thing is that Stockton, even relative to his peers, is a ridiculous outlier as a shot creator for others. This is even compared to guys like Magic or Nash, much less to everyone else.

You say "even", but I'd imagine it's an easier to be an outlier when Stockton played. My impression is also that Magic was a much better shot creator with the "quality" of his creations being much higher. Ditto with Nash. Stockton didn't really pry or penetrate which is a big limitation for a playmaker. I'd guess stockton's "quality of creation" was actually pretty low:
[url]

Watch on YouTube

[/url]
Looking at Stockton's first 10 assists, only 3 lead to an open shot. And only 2 have stockton taking more than 1 defender out of the play. 3 of these involve the player Stockton passes to dealing with multiple defenders to score, and all of these reads seem pretty simple by the standards of a modern helio. Stockton creates 1 open layup, and his other assist is a simple read to a shooter(and these are generally alot less valuable than say, layup assists which made up roughly half of Nash's dishes).

Would probably put his playmaking on par with someone like Pippen rather than all-time-great creators like Nash, Magic, or Jokic. Even comparing him to what would be a second-tier passer in today's NBA:
https://youtu.be/EmHJI0NRqmk?t=23
All 3 of Luka's assists involve Luka taking out multiple defenders and creating an open shot. All 3 have Luka passing through tighter windows with less time to make a decision.

Luka actively manipulates and anticipates openings. From what I've seen, Stockton generally just reacts to what shows up. Even in a comparison to his peers, someone like Isiah is alot more aggressive(and I think more developed handles contribute to this). Unless we consider rondo/draymond great playmakers, I don't think what stockton brings is a recipe for driving an offense.

I think NBA-Storyteller sums up the issue with stockton getting hyped based on his apg pretty well:
https://youtu.be/LsmVMIKWOSM?t=21
the rest of the video is an entertaining meme, but the malone analysis is on-point.

Regardless, if it is sufficient for you that BPM's creators tried to line-up their box-interpretation with winning, then perhaps you ought to consider a vote for Rodman. After all, I'm sure such efforts were also made by whoever created the formula for IBM. And while Stockton may look quite good, Rodman ended up looking league-best.

[OldSchoolNoBull]

What I said there about Reggie basically applies to Stockton too RE: offense.

His assessment of these guys is annoying because there's a complete lack of analysis as to how they'd play today.

It's counterintuituve, but Reggie for example would be worse today. Why? Because when Reggie played teams didn't understand the value of 3s, which contributed to an environment where it was easier for him to get open 3s (especially the 3 years the line moved in).

Today teams sell out to stop open 3s. The way you take advantage of that is by being a deadly iso-player who blows by your defender when they get too close trying to deny the 3. Harden is a good example, with his shiftyness and handle letting him kill defenders for making the wrong decision. Miller was not much of a driver. He didn't have shake or a lightning first step or superior athleticism. In today's game he might take more 3s, but he wouldn't be able to exploit his 3pt ability to drive powerful offenses. He'd basically be a slightly lesser version of Klay. That's valuable, but it doesn't make him a star today. Ditto Stockton in relation to his 3pt shot.

He'd also be worse because he'd be taken advantage of so much more on D, where they'd bumhunt him on pick and rolls and score in him at will. In Reggie's day you had minimal movement and Reggie could chill and get some rest on a bad offensive player.

Reggie for his time strikes me as more impressive than Stockton when stacking up primes (same story when it comes to modern translation - though I’m not putting any stock into that). As a very scalable / team-friendly offensive lead guy, his playoff scoring & ramp up in such is impressive. I don’t think he’s anything that insane on defense, but not a bad one by any means either. I would prefer him as a 1st option (and really anything after that) compared to Stockton.

I don’t really care to theorize that Reggie wouldn’t be as outlierish today when assessing his career though. If he played today, perhaps a coach would be more strict on having his feet behind the arc . Anyhow, like everybody else - it makes sense to assess Reggie for the 90s era he actually accrued value in.

I don't really agree that Reggie would be worse in the current league where the things he was best at - shooting and off-ball movement - are valued more than they ever have been.

Reggie was a career 39.5% and 39% 3P shooter in the RS and PO, respectively, having topped 40% 10 times in the RS and 9 times in the PO.

Reggie averaged +8.1 rTS in his 18 regular seasons.
Comparing Reggie's career PO TS to the RS league average over his career, it's +7.1.

As a point of comparison, Kevin Durant's numbers are +7.4 and +5.1.
(Yes, KD did it on higher volume - 37.1PP100 RS and 36.9PP100 PO, but Reggie's 27.5PP100 RS and 30.5PP100 PO are nothing to sneeze at.)

As another point of comparison, Ray Allen - a player with a very similar offensive skillset to Reggie who was still a pretty decent player as recently as the early 2010s - averaged +5.2 rTS over his 18 regular seasons and +5.5 over the RS league average in the playoffs on 28pp100 and 24.7pp100 respectively. That's a 118.3 career average TS Add vs Reggie's 191.7.

Obviously there would be more to comparing Reggie and Ray as players because there's more to the game than scoring, but it does bare mentioning that Reggie has a career .176 WS/48 and 3.5 BPM compared to Ray's .150 and 2.9, and in the playoffs it's Reggie's .180 and 5.0 vs Ray's .143 and 3.2.

Reggie achieved a scoring efficiency rarely seen over a career as long as his, and he didn't need the ball in his hands to do it, so I don't see why those skills wouldn't translate into the modern game that values those exact skills the most. OaD argues that Reggie wouldn't get as many open looks because defenses are better, but Reggie was one of the greatest off-ball movers ever, so I have a certain degree of confidence he'd find his way to where the good shots are.

If the assertion is he wouldn't be a #1 option today, I'd say I'm not entirely sure he was a traditional #1 option when he did play, as those Pacers teams were more of an ensemble piece, a la the 2004 Pistons.