GOAT methodology for GOAT lists

Moderators: Clyde Frazier, Doctor MJ, trex_8063, penbeast0, PaulieWal

Ginoboleee
Sophomore
Posts: 210
And1: 75
Joined: Jun 19, 2022

Re: GOAT methodology for GOAT lists 

Post#41 » by Ginoboleee » Sun Jul 3, 2022 10:09 pm

Owly wrote:
f4p wrote:
Cavsfansince84 wrote: This is somewhat close to what I do though playoffs are a big factor as well. If a guy has 12 chances at winning a ring and always comes up way short and never raises his game that is something of a red flag to me. Because even though playoffs are a small sample at some point a guy has to work on his weaknesses that teams may exploit or gameplan for in the postseason then show the ability to perform when its most relevant. The guys who do that are the ones who most care about winning imo and often end up with the most rings relative to peers. The ones who lose then ask 'what do I need to do better' and make those improvements which also goes into how much they may need to rely on others at certain parts of a game. Maybe this is also why Russell didn't develop much offensively since what he was doing seemed to always work in terms of winning.


Honestly, when we're talking about the GOAT, I really don't understand considering almost anything other than the playoffs. Based on my analysis of NBA history, all of the titles have been handed out for the playoffs and none have been given for the regular season.
The playoffs are simply too important.

This isn't baseball. There you have decades of history where hardly any teams make the playoffs, and if you didn't play for the Yankees, you might never make the playoffs. Ted Williams has 5 career playoff games and 25 career at-bats. Your regular season performance to get your team to the postseason is a huge part of who you are as a player. And even once you reach the postseason, it's so random that it's hard to really grade anyone on what might only be 30 or 40 games.

And it's not football, where basically anyone who isn't a QB is hard enough to judge already from a legacy perspective, and even QB's can be affected by so many things like their offensive line or the defense they face that it might not be possible to even do a statistical playoff comparison, much less grade by who wins and who loses.

But the NBA? So many teams get to the playoffs that even guys who weren't in blessed circumstances, like Hakeem, played almost 2 years worth of playoff games (145). And some guys are closer to 3 years. That's a large sample to draw from, usually over many different years. The sport isn't as random as baseball and 7-game series make winning or losing less random than football (and even compared to series in baseball). And you typically don't have massive fluctuations in stats based on your opponents. A few points, a few percentages here or there? Sure. But you don't see the equivalent of a QB going from 3 TD/0 INT against a weak team to 1 TD/3 INT against a great team. The great players usually go down swinging with big time numbers. And great players have a huge impact on the game compared to anyone other than QB's or starting pitchers (for the 1 out of 4 or 5 games they are starting).

So you get plenty of chances to prove yourself, you usually get a chance every year, whether you win or lose might have a lot to do with your teammates but it won't be fluky and you almost always have a chance to shine statistically. And because basically everyone makes the playoffs, it's the only time that really matters. And in a sport where great players routinely take it easier in the regular season and build for the playoffs, it can truly be a different sport. A great regular season followed by a poor playoffs isn't just criticized by fans, the player himself is usually hugely disappointed. i guess i can see a little emphasis on the regular season because it's still a lot of games but it just usually doesn't mean much. winning MVP and losing in the first round is a disaster, not a feather in your cap.

now further down the list, i can see comparing lesser players with their regular seasons more accounted for because you are usually talking about people with flawed playoff resumes and the regular season can be a big part of what they did.

My issues with strong playoff weighting and/or things that make it hard to do fairly.

1) Proportionally it's still a much smaller sample.
2) It's an uneven sample (some will get long runs in their best years, others will get large ones in their larger samples in more pedestrian years).
3) Uneven schedules/matchups. Greater than normal variation in quality of opponents. Specific matchup much less varied. Related: increased planning may lead to increased defensive focus [unevenly so depending on teammates, opponent, matchup, coaches etc] limiting opportunity to produce, even where impact/advantage of increased attention continues to help the team.
4) slight disagreement with above - don't think for the majority of NBA history great players have tended to cruise through RS. Honestly as far as asset management, maximizing title odds etc, RS being insignificant [not always the prevailing thinking, I think] many on strong teams should have done so much, much more).
5) Playoff qualification not a given (GOAT candidate players have missed playoffs in prime).
6) Even playoff qualification is not always optimally aligned with team goodness - (an outlier but ... 5.57 SRS '72 Suns missed the playoffs).

Losing in the first round is a bad team outcome, but it doesn't mean that a player on the losing team had a bad year, is a bad player or had a bad series, just that the team they are on. The implication of "disaster" seems to be that the loss must mean the player did bad (or else relevance becomes a question mark - it being bad for the team is a given and the comment isn't about a team, it being frustrating for the player is a given ... it's hard for me to parse another meaning though I may have missed something). But let's just temporarily imagine a player can control everything on the court in his time on, regardless of the other 9 players ...Even just team level Embiid '19 +89 plus minus in ECSF, positive +/- in 6 of 7 games [double digits positive in 4]. His team lost and was outscored by 19. 34mpg would be a small ding on the player, but still. Leaving aside what you think of the specific player[Embiid]'s performance ... a guy could have been +100 and still had his team be outscored (difference between on and net is 108) ... and you don't need to be outscored to lose a series.

This is a bit rambly and could do with some editing - if it gets posted with this attached - I didn't do that. I feel like I'm missing some nuances somewhere.


Great post. Thank you!

I would (eventually like to, but not today) challenge points 1-3, both those are known problems which I am sure lots of folks have lots of potential solutions for. Or not.

Point 4 is interesting. You are right, that RS play probably meant back in the day when home court advantage was paramount, and overall media/culture hype with the playoffs was lower (playoffs were on tape delay - broadcast THE NEXT DAY - when I was a kid). But I still think the deeper truth is that the RS play has (naturally) always been at a lower intensity than the playoffs (though the gap might have been smaller in the past). In fact, in today's game, as the AAU-ification of the league steadily grows, it is entirely likely that the gap between RS and Playoffs has never been greater. It sure seemed that way to my eye this postseason just concluded - like there was a time warp back to real competitive/physical basketball! (Insert obligatory nostalgic 90s/00s reference here.)

But it is Point 5 and 6 that I respectfully disagree with almost entirely.
For teams to advance THROUGH the playoffs there are many factors that come into play - some are obvious - others are being highlighted in some of the posts in this thread.
But for team to qualify FOR the playoffs, well, that is an entirely different deal.
Simply, in the NBA, the circumstances have to be extreme and rare for a great player to not be able to get their team to qualify for postseason play.
I am confident that it has happened very rarely. And likely due to untimely injuries.
Sure there will be a rando bad year here and there, but Superstars MAKE the playoffs.

The most extreme situation I can remember is when KG played for years on end with a bunch of semi-professional teammates. Sure they always lost in the first round - that was a bit over-determined! But they MADE the playoffs once KG was legit.

There will be exceptions, but I am sure they are rare enough and/or obvious enough and/or weird enough to warrant a fairly confident rejection of Points 5 & 6.

(PS - I didn't think your post was rambling. But that is coming from the stream-of-consciousness Rambler In Chief, so, yeah.)
Life it is not just a series of calculations and a sum total of statistics, it's about experience, it's about participation, it is something more complex and more interesting than what is obvious.
Libeskind

Statistics are no substitute for judgment.
Clay
f4p
Sixth Man
Posts: 1,775
And1: 1,783
Joined: Sep 19, 2021
 

Re: GOAT methodology for GOAT lists 

Post#42 » by f4p » Mon Jul 4, 2022 3:17 am

Owly wrote:My issues with strong playoff weighting and/or things that make it hard to do fairly.

1) Proportionally it's still a much smaller sample.
2) It's an uneven sample (some will get long runs in their best years, others will get large ones in their larger samples in more pedestrian years).
3) Uneven schedules/matchups. Greater than normal variation in quality of opponents. Specific matchup much less varied. Related: increased planning may lead to increased defensive focus [unevenly so depending on teammates, opponent, matchup, coaches etc] limiting opportunity to produce, even where impact/advantage of increased attention continues to help the team.
4) slight disagreement with above - don't think for the majority of NBA history great players have tended to cruise through RS. Honestly as far as asset management, maximizing title odds etc, RS being insignificant [not always the prevailing thinking, I think] many on strong teams should have done so much, much more).
5) Playoff qualification not a given (GOAT candidate players have missed playoffs in prime).
6) Even playoff qualification is not always optimally aligned with team goodness - (an outlier but ... 5.57 SRS '72 Suns missed the playoffs).

Losing in the first round is a bad team outcome, but it doesn't mean that a player on the losing team had a bad year, is a bad player or had a bad series, just that the team they are on. The implication of "disaster" seems to be that the loss must mean the player did bad (or else relevance becomes a question mark - it being bad for the team is a given and the comment isn't about a team, it being frustrating for the player is a given ... it's hard for me to parse another meaning though I may have missed something). But let's just temporarily imagine a player can control everything on the court in his time on, regardless of the other 9 players ...Even just team level Embiid '19 +89 plus minus in ECSF, positive +/- in 6 of 7 games [double digits positive in 4]. His team lost and was outscored by 19. 34mpg would be a small ding on the player, but still. Leaving aside what you think of the specific player[Embiid]'s performance ... a guy could have been +100 and still had his team be outscored (difference between on and net is 108) ... and you don't need to be outscored to lose a series.

This is a bit rambly and could do with some editing - if it gets posted with this attached - I didn't do that. I feel like I'm missing some nuances somewhere.


1) Sure, it's proportionally smaller, but 20% of a sample of 1000 has much less variance than 20% of a sample of 10. Basketball already has 100 possessions a game to smooth out a lot of the game to game variation (unlike 4 at bats in baseball or 30 pass attempts in football) and then guys have 150-250 playoff games to smooth those variations out. That's a lot. If a player needs more than 200 games to show us their true basketball talent, then I guess that's just unfortunate because athletes only have finite careers and at some point we have to make a judgment based on what happened.

2/3) As for uneven samples/opponents, I would certainly advocate for as much granularity as possible. Especially for more modern series (even back to the 80's) that we've all had a good chance to see or read about or look up numbers from. I'm not saying just spit out a catch all number from 15 years and say that's that. I once even tried to adjust for what years a player played their playoff games by weighting people's regular season numbers by how many playoff games they played (so playing 10% of your playoff games in a year means the regular season stats get a 10% weight for your career), but it's an old spreadsheet at this point. Weirdly, the person who most benefited was Nate Archibald and the person who looked the worst after the adjustment was...Nate Archibald. His unadjusted dropoff in the playoffs was just truly astounding.

4) Players might not cruise in the regular season like 2nd half of career Lebron or like Kawhi, but i don't think the Showtime Lakers were killing themselves every night when they had an overwhelming talent advantage and were up 8 or 9 games on 2nd in the west. No matter what player is talking, they seem to all agree they took a step up in intensity and energy in the playoffs.

5) Yes, but it's pretty rare. I don't think Jordan, Russell, Wilt, Magic, Bird, or Duncan ever missed and after turning 20, neither did Shaq. Lebron 1 time after 20, Hakeem 1 time, Kobe 1 time, Kareem 2 I believe. That's the normal top 11 guys. After that, only KG with 3 and Steph with 5 are really significant off the top of my head.

6) I'm definitely not just saying see who wins. I'm saying what you are saying. That unlike some other sports, it seems much easier and quite common for great players to look great while losing. And we can account for that. Jordan might have lost to the Celtics and Pistons, but he put up GOAT stats while doing it so no one should discredit him for those losses. Luka might have lost to the clippers 2 years in a row in the 1st round and the Warriors this year, but his greatness was apparent with monster numbers and the other team having no idea how to stop him. But to me this is part of why we should focus on the playoffs. To make the case for Embiid in the series you reference, instead of spending that limited time talking about what he did on a Monday night in January.
Owly
Lead Assistant
Posts: 5,621
And1: 3,138
Joined: Mar 12, 2010

Re: GOAT methodology for GOAT lists 

Post#43 » by Owly » Mon Jul 4, 2022 3:44 pm

Ginoboleee wrote:
Owly wrote:
f4p wrote:
Honestly, when we're talking about the GOAT, I really don't understand considering almost anything other than the playoffs. Based on my analysis of NBA history, all of the titles have been handed out for the playoffs and none have been given for the regular season.
The playoffs are simply too important.

This isn't baseball. There you have decades of history where hardly any teams make the playoffs, and if you didn't play for the Yankees, you might never make the playoffs. Ted Williams has 5 career playoff games and 25 career at-bats. Your regular season performance to get your team to the postseason is a huge part of who you are as a player. And even once you reach the postseason, it's so random that it's hard to really grade anyone on what might only be 30 or 40 games.

And it's not football, where basically anyone who isn't a QB is hard enough to judge already from a legacy perspective, and even QB's can be affected by so many things like their offensive line or the defense they face that it might not be possible to even do a statistical playoff comparison, much less grade by who wins and who loses.

But the NBA? So many teams get to the playoffs that even guys who weren't in blessed circumstances, like Hakeem, played almost 2 years worth of playoff games (145). And some guys are closer to 3 years. That's a large sample to draw from, usually over many different years. The sport isn't as random as baseball and 7-game series make winning or losing less random than football (and even compared to series in baseball). And you typically don't have massive fluctuations in stats based on your opponents. A few points, a few percentages here or there? Sure. But you don't see the equivalent of a QB going from 3 TD/0 INT against a weak team to 1 TD/3 INT against a great team. The great players usually go down swinging with big time numbers. And great players have a huge impact on the game compared to anyone other than QB's or starting pitchers (for the 1 out of 4 or 5 games they are starting).

So you get plenty of chances to prove yourself, you usually get a chance every year, whether you win or lose might have a lot to do with your teammates but it won't be fluky and you almost always have a chance to shine statistically. And because basically everyone makes the playoffs, it's the only time that really matters. And in a sport where great players routinely take it easier in the regular season and build for the playoffs, it can truly be a different sport. A great regular season followed by a poor playoffs isn't just criticized by fans, the player himself is usually hugely disappointed. i guess i can see a little emphasis on the regular season because it's still a lot of games but it just usually doesn't mean much. winning MVP and losing in the first round is a disaster, not a feather in your cap.

now further down the list, i can see comparing lesser players with their regular seasons more accounted for because you are usually talking about people with flawed playoff resumes and the regular season can be a big part of what they did.

My issues with strong playoff weighting and/or things that make it hard to do fairly.

1) Proportionally it's still a much smaller sample.
2) It's an uneven sample (some will get long runs in their best years, others will get large ones in their larger samples in more pedestrian years).
3) Uneven schedules/matchups. Greater than normal variation in quality of opponents. Specific matchup much less varied. Related: increased planning may lead to increased defensive focus [unevenly so depending on teammates, opponent, matchup, coaches etc] limiting opportunity to produce, even where impact/advantage of increased attention continues to help the team.
4) slight disagreement with above - don't think for the majority of NBA history great players have tended to cruise through RS. Honestly as far as asset management, maximizing title odds etc, RS being insignificant [not always the prevailing thinking, I think] many on strong teams should have done so much, much more).
5) Playoff qualification not a given (GOAT candidate players have missed playoffs in prime).
6) Even playoff qualification is not always optimally aligned with team goodness - (an outlier but ... 5.57 SRS '72 Suns missed the playoffs).

Losing in the first round is a bad team outcome, but it doesn't mean that a player on the losing team had a bad year, is a bad player or had a bad series, just that the team they are on. The implication of "disaster" seems to be that the loss must mean the player did bad (or else relevance becomes a question mark - it being bad for the team is a given and the comment isn't about a team, it being frustrating for the player is a given ... it's hard for me to parse another meaning though I may have missed something). But let's just temporarily imagine a player can control everything on the court in his time on, regardless of the other 9 players ...Even just team level Embiid '19 +89 plus minus in ECSF, positive +/- in 6 of 7 games [double digits positive in 4]. His team lost and was outscored by 19. 34mpg would be a small ding on the player, but still. Leaving aside what you think of the specific player[Embiid]'s performance ... a guy could have been +100 and still had his team be outscored (difference between on and net is 108) ... and you don't need to be outscored to lose a series.

This is a bit rambly and could do with some editing - if it gets posted with this attached - I didn't do that. I feel like I'm missing some nuances somewhere.


Great post. Thank you!

I would (eventually like to, but not today) challenge points 1-3, both those are known problems which I am sure lots of folks have lots of potential solutions for. Or not.

Point 4 is interesting. You are right, that RS play probably meant back in the day when home court advantage was paramount, and overall media/culture hype with the playoffs was lower (playoffs were on tape delay - broadcast THE NEXT DAY - when I was a kid). But I still think the deeper truth is that the RS play has (naturally) always been at a lower intensity than the playoffs (though the gap might have been smaller in the past). In fact, in today's game, as the AAU-ification of the league steadily grows, it is entirely likely that the gap between RS and Playoffs has never been greater. It sure seemed that way to my eye this postseason just concluded - like there was a time warp back to real competitive/physical basketball! (Insert obligatory nostalgic 90s/00s reference here.)

But it is Point 5 and 6 that I respectfully disagree with almost entirely.
For teams to advance THROUGH the playoffs there are many factors that come into play - some are obvious - others are being highlighted in some of the posts in this thread.
But for team to qualify FOR the playoffs, well, that is an entirely different deal.
Simply, in the NBA, the circumstances have to be extreme and rare for a great player to not be able to get their team to qualify for postseason play.
I am confident that it has happened very rarely. And likely due to untimely injuries.
Sure there will be a rando bad year here and there, but Superstars MAKE the playoffs.

The most extreme situation I can remember is when KG played for years on end with a bunch of semi-professional teammates. Sure they always lost in the first round - that was a bit over-determined! But they MADE the playoffs once KG was legit.

There will be exceptions, but I am sure they are rare enough and/or obvious enough and/or weird enough to warrant a fairly confident rejection of Points 5 & 6.

(PS - I didn't think your post was rambling. But that is coming from the stream-of-consciousness Rambler In Chief, so, yeah.)

This doesn't address point 6 (about playoff qualification being murky bar for team quality.

It was to extend the examples further, some good teams miss the playoffs (SRS north of 2)
1971-72 Phoenix Suns
2013-14 Minnesota Timberwolves
2013-14 Phoenix Suns
2000-01 Houston Rockets
2014-15 Oklahoma City Thunder
1973-74 Golden State Warriors
2007-08 Golden State Warriors
2010-11 Houston Rockets
1970-71 Phoenix Suns
1996-97 Cleveland Cavaliers
1970-71 Boston Celtics
1978-79 Milwaukee Bucks

Some bad teams make it (SRS south of -2)

1950-51 NBA Indianapolis Olympians*
1985-86 NBA San Antonio Spurs*
1960-61 NBA Detroit Pistons*
2015-16 NBA Memphis Grizzlies*
1956-57 NBA Fort Wayne Pistons*
1970-71 ABA Texas Chaparrals*
2007-08 NBA Atlanta Hawks*
1984-85 NBA Cleveland Cavaliers*
1949-50 NBA Philadelphia Warriors*
1979-80 NBA Washington Bullets*
1973-74 ABA San Diego Conquistadors*
1952-53 NBA Indianapolis Olympians*
1957-58 NBA Detroit Pistons*
1984-85 NBA Phoenix Suns*
1983-84 NBA Washington Bullets*
1986-87 NBA Golden State Warriors*
1995-96 NBA Sacramento Kings*
1996-97 NBA Los Angeles Clippers*
1966-67 NBA New York Knicks*
1985-86 NBA Chicago Bulls*
2014-15 NBA Brooklyn Nets*
1970-71 ABA Denver Rockets*
1985-86 NBA Sacramento Kings*
1948-49 BAA St. Louis Bombers*
1966-67 NBA Chicago Bulls*
1962-63 NBA Detroit Pistons*
1959-60 NBA Detroit Pistons*
1972-73 ABA San Diego Conquistadors*
1963-64 NBA Philadelphia 76ers*
1947-48 BAA Boston Celtics*
1967-68 NBA Chicago Bulls*
1991-92 NBA Miami Heat*
1978-79 NBA New Jersey Nets*
1974-75 ABA Spirits of St. Louis*
1967-68 ABA Houston Mavericks*
1959-60 NBA Minneapolis Lakers*
1973-74 ABA Virginia Squires*
1974-75 ABA Memphis Sounds*
1987-88 NBA San Antonio Spurs*
1952-53 NBA Baltimore Bullets*
1972-73 ABA New York Nets*
1949-50 NBA Sheboygan Red Skins*

On 5 it wasn't asserted that it happened often. But as I said from some people's GOAT tier Jabbar misses two consecutive in prime (between different plausible peak years). Chamberlain misses one in prime and was set to miss a second before being traded. Garnett is in many top 10s here and misses 3 straight. Robertson who has been atop published GOAT lists (though now circa 30 years ago) missed 3 in his prime. Olajuwon missed in prime, Curry did, Bryant did, Barkley (twice), Pettit ... and that's without going to people who missed very large chunks of seasons, post prime years etc.

It doesn't have to be fatal to playoff weighting depending on what claim or structure is put but a very heavy playoff focus would seem to leave some great seasons uncounted.
f4p
Sixth Man
Posts: 1,775
And1: 1,783
Joined: Sep 19, 2021
 

Re: GOAT methodology for GOAT lists 

Post#44 » by f4p » Mon Jul 4, 2022 4:37 pm

Owly wrote:This doesn't address point 6 (about playoff qualification being murky bar for team quality.

It was to extend the examples further, some good teams miss the playoffs (SRS north of 2)
1971-72 Phoenix Suns
2013-14 Minnesota Timberwolves
2013-14 Phoenix Suns
2000-01 Houston Rockets
2014-15 Oklahoma City Thunder
1973-74 Golden State Warriors
2007-08 Golden State Warriors
2010-11 Houston Rockets
1970-71 Phoenix Suns
1996-97 Cleveland Cavaliers
1970-71 Boston Celtics
1978-79 Milwaukee Bucks

Some bad teams make it (SRS south of -2)

1950-51 NBA Indianapolis Olympians*
1985-86 NBA San Antonio Spurs*
1960-61 NBA Detroit Pistons*
2015-16 NBA Memphis Grizzlies*
1956-57 NBA Fort Wayne Pistons*
1970-71 ABA Texas Chaparrals*
2007-08 NBA Atlanta Hawks*
1984-85 NBA Cleveland Cavaliers*
1949-50 NBA Philadelphia Warriors*
1979-80 NBA Washington Bullets*
1973-74 ABA San Diego Conquistadors*
1952-53 NBA Indianapolis Olympians*
1957-58 NBA Detroit Pistons*
1984-85 NBA Phoenix Suns*
1983-84 NBA Washington Bullets*
1986-87 NBA Golden State Warriors*
1995-96 NBA Sacramento Kings*
1996-97 NBA Los Angeles Clippers*
1966-67 NBA New York Knicks*
1985-86 NBA Chicago Bulls*
2014-15 NBA Brooklyn Nets*
1970-71 ABA Denver Rockets*
1985-86 NBA Sacramento Kings*
1948-49 BAA St. Louis Bombers*
1966-67 NBA Chicago Bulls*
1962-63 NBA Detroit Pistons*
1959-60 NBA Detroit Pistons*
1972-73 ABA San Diego Conquistadors*
1963-64 NBA Philadelphia 76ers*
1947-48 BAA Boston Celtics*
1967-68 NBA Chicago Bulls*
1991-92 NBA Miami Heat*
1978-79 NBA New Jersey Nets*
1974-75 ABA Spirits of St. Louis*
1967-68 ABA Houston Mavericks*
1959-60 NBA Minneapolis Lakers*
1973-74 ABA Virginia Squires*
1974-75 ABA Memphis Sounds*
1987-88 NBA San Antonio Spurs*
1952-53 NBA Baltimore Bullets*
1972-73 ABA New York Nets*
1949-50 NBA Sheboygan Red Skins*

On 5 it wasn't asserted that it happened often. But as I said from some people's GOAT tier Jabbar misses two consecutive in prime (between different plausible peak years). Chamberlain misses one in prime and was set to miss a second before being traded. Garnett is in many top 10s here and misses 3 straight. Robertson who has been atop published GOAT lists (though now circa 30 years ago) missed 3 in his prime. Olajuwon missed in prime, Curry did, Bryant did, Barkley (twice), Pettit ... and that's without going to people who missed very large chunks of seasons, post prime years etc.

It doesn't have to be fatal to playoff weighting depending on what claim or structure is put but a very heavy playoff focus would seem to leave some great seasons uncounted.


Fair enough. I should add, comparing regular season to postseason numbers is probably my favorite way to use the playoffs. I can't be certain that AuPM or PER or even something simple like TS% are giving me a perfect look at a player, but I can be more sure that whatever is "right" or "wrong" for that player is equally right or wrong in the regular season and the playoffs, so it should be a good basis for comparison to see who stepped up and who didn't. Now some would rightfully say, is a 10 who turns into a 9 worse than a 7 who turns into an 8 in the playoffs? No, but I think some of these guys are so close (more like 9.5 vs 9.3), that I'm willing to use the differential. I feel more confident in saying Larry Bird looks worse in the playoffs than the regular season than in trying to figure out if Bird just barely falls behind Russell and just barely ahead of Magic in the regular season.
Owly
Lead Assistant
Posts: 5,621
And1: 3,138
Joined: Mar 12, 2010

Re: GOAT methodology for GOAT lists 

Post#45 » by Owly » Mon Jul 4, 2022 4:55 pm

f4p wrote:
Owly wrote:My issues with strong playoff weighting and/or things that make it hard to do fairly.

1) Proportionally it's still a much smaller sample.
2) It's an uneven sample (some will get long runs in their best years, others will get large ones in their larger samples in more pedestrian years).
3) Uneven schedules/matchups. Greater than normal variation in quality of opponents. Specific matchup much less varied. Related: increased planning may lead to increased defensive focus [unevenly so depending on teammates, opponent, matchup, coaches etc] limiting opportunity to produce, even where impact/advantage of increased attention continues to help the team.
4) slight disagreement with above - don't think for the majority of NBA history great players have tended to cruise through RS. Honestly as far as asset management, maximizing title odds etc, RS being insignificant [not always the prevailing thinking, I think] many on strong teams should have done so much, much more).
5) Playoff qualification not a given (GOAT candidate players have missed playoffs in prime).
6) Even playoff qualification is not always optimally aligned with team goodness - (an outlier but ... 5.57 SRS '72 Suns missed the playoffs).

Losing in the first round is a bad team outcome, but it doesn't mean that a player on the losing team had a bad year, is a bad player or had a bad series, just that the team they are on. The implication of "disaster" seems to be that the loss must mean the player did bad (or else relevance becomes a question mark - it being bad for the team is a given and the comment isn't about a team, it being frustrating for the player is a given ... it's hard for me to parse another meaning though I may have missed something). But let's just temporarily imagine a player can control everything on the court in his time on, regardless of the other 9 players ...Even just team level Embiid '19 +89 plus minus in ECSF, positive +/- in 6 of 7 games [double digits positive in 4]. His team lost and was outscored by 19. 34mpg would be a small ding on the player, but still. Leaving aside what you think of the specific player[Embiid]'s performance ... a guy could have been +100 and still had his team be outscored (difference between on and net is 108) ... and you don't need to be outscored to lose a series.

This is a bit rambly and could do with some editing - if it gets posted with this attached - I didn't do that. I feel like I'm missing some nuances somewhere.


1) Sure, it's proportionally smaller, but 20% of a sample of 1000 has much less variance than 20% of a sample of 10. Basketball already has 100 possessions a game to smooth out a lot of the game to game variation (unlike 4 at bats in baseball or 30 pass attempts in football) and then guys have 150-250 playoff games to smooth those variations out. That's a lot. If a player needs more than 200 games to show us their true basketball talent, then I guess that's just unfortunate because athletes only have finite careers and at some point we have to make a judgment based on what happened.

2/3) As for uneven samples/opponents, I would certainly advocate for as much granularity as possible. Especially for more modern series (even back to the 80's) that we've all had a good chance to see or read about or look up numbers from. I'm not saying just spit out a catch all number from 15 years and say that's that. I once even tried to adjust for what years a player played their playoff games by weighting people's regular season numbers by how many playoff games they played (so playing 10% of your playoff games in a year means the regular season stats get a 10% weight for your career), but it's an old spreadsheet at this point. Weirdly, the person who most benefited was Nate Archibald and the person who looked the worst after the adjustment was...Nate Archibald. His unadjusted dropoff in the playoffs was just truly astounding.

4) Players might not cruise in the regular season like 2nd half of career Lebron or like Kawhi, but i don't think the Showtime Lakers were killing themselves every night when they had an overwhelming talent advantage and were up 8 or 9 games on 2nd in the west. No matter what player is talking, they seem to all agree they took a step up in intensity and energy in the playoffs.

Is it possible some Showtime Laker teams eased off the gas. Maybe.
a) Teams can do this (more common with champs) without players doing it (minute distribution).
b) Maybe we overrate the early Showtime teams based on titles won off an easy route to the finals (greater chance of getting there, get there fresher).
c) Not sure about Riley letting his foot off the gas. Rputation is a hard taskmaster.
d) '87 Lakers won 65 games and posted 8.32 SRS with a 10 game lead in conference and 6 games on 2nd place overall Boston.
5) Yes, but it's pretty rare. I don't think Jordan, Russell, Wilt, Magic, Bird, or Duncan ever missed and after turning 20, neither did Shaq. Lebron 1 time after 20, Hakeem 1 time, Kobe 1 time, Kareem 2 I believe. That's the normal top 11 guys. After that, only KG with 3 and Steph with 5 are really significant off the top of my head.

6) I'm definitely not just saying see who wins. I'm saying what you are saying. That unlike some other sports, it seems much easier and quite common for great players to look great while losing. And we can account for that. Jordan might have lost to the Celtics and Pistons, but he put up GOAT stats while doing it so no one should discredit him for those losses. Luka might have lost to the clippers 2 years in a row in the 1st round and the Warriors this year, but his greatness was apparent with monster numbers and the other team having no idea how to stop him. But to me this is part of why we should focus on the playoffs. To make the case for Embiid in the series you reference, instead of spending that limited time talking about what he did on a Monday night in January.

So I'll say up front I think we disagree and we're going to continue to come to some very different conclusions. But you're putting your points sensibly, politely and I appreciate that.

That said some areas of pushback
1) Since you seem to be talking about comparisons at the GOAT-ish tier, my spreadsheet (think it's sorted by playoff Hollinger EWA over a very good (17.9 PER) baseline ... so already filtered/biased towards players with large playoff samples) going to the top 33 (up to Erving) ... through 2019 I had only 1 player whose playoff minutes made up 20% of total minutes. Robertson, Moses, Schayes are circa 8%; Howard, Garnett circa 9%. Pettit, Paul, Nowitzki circa 10%. The average of this group is 0.133741667 and as already stated this is a baseline list biased towards those with large playoff samples.
Sidenote: If one were to go to the top 100 EWA above good (and I would guess you personally would be looking at a different, less playoff tilted criteria at this point, based on what you said above) which is again already a biased list, we see players with as little as 3% of their career minutes in playoffs (King, Brand), average is circa 10%. Among all players it's circa 3%.
Then too look how some shooting stats fluctuate year to year. And shooting's the most important part of the game. If you're lucky we've got two RS's worth of data on these players (versus uneven defenses etc.).

2/3) All sounds good. It's hard to do. It's messy. It' hard to show you've done it fairly. Even after all that if you just happen to run up against your worst matchup a couple of times ... a heavy playoff tilt seems pretty rough.

4) Don't think "all" is right here. Can't source it but seem to recall Karl Malone objecting to the idea/implication of giving giving less in the RS. Struggle to see Cowens playing with the peddle off (or Garnett or ...). I don't know what the ratio on quotes would be nor that I'd always trust them either way (depends what they were asked, how they were asked, when they were asked - an "old timer" might want to say "We went hard every night" to promote their era, in the midst of a playoff run with a leading question a player might say "I've got a responsibility to find an extra gear to lead us ...")

5) Assume these are in prime (because for instance 0 on MJ), Wilt did once and would have a second time if not traded. From my last post in response to another poster
... Garnett is in many top 10s here and misses 3 straight. Robertson who has been atop published GOAT lists (though now circa 30 years ago) missed 3 in his prime. Olajuwon missed in prime, Curry did, Bryant did, Barkley (twice), Pettit ... and that's without going to people who missed very large chunks of seasons, post prime years etc.


"6") Regarding my paragraph re on-off etc. I'm not sure about how this gets to "this is ... why we should focus on the playoffs". But it does seem more sensible a take than I'd interpreted off "the winning MVP and losing in the first round is a disaster" line which (as stated above) seemed to start off talking about individual performance and then use a crude team level outcome against the individual - my guess is that my interpretation of this was not what you had in your head.
Owly
Lead Assistant
Posts: 5,621
And1: 3,138
Joined: Mar 12, 2010

Re: GOAT methodology for GOAT lists 

Post#46 » by Owly » Mon Jul 4, 2022 5:17 pm

f4p wrote:
Owly wrote:This doesn't address point 6 (about playoff qualification being murky bar for team quality.

It was to extend the examples further, some good teams miss the playoffs (SRS north of 2)
1971-72 Phoenix Suns
2013-14 Minnesota Timberwolves
2013-14 Phoenix Suns
2000-01 Houston Rockets
2014-15 Oklahoma City Thunder
1973-74 Golden State Warriors
2007-08 Golden State Warriors
2010-11 Houston Rockets
1970-71 Phoenix Suns
1996-97 Cleveland Cavaliers
1970-71 Boston Celtics
1978-79 Milwaukee Bucks

Some bad teams make it (SRS south of -2)

1950-51 NBA Indianapolis Olympians*
1985-86 NBA San Antonio Spurs*
1960-61 NBA Detroit Pistons*
2015-16 NBA Memphis Grizzlies*
1956-57 NBA Fort Wayne Pistons*
1970-71 ABA Texas Chaparrals*
2007-08 NBA Atlanta Hawks*
1984-85 NBA Cleveland Cavaliers*
1949-50 NBA Philadelphia Warriors*
1979-80 NBA Washington Bullets*
1973-74 ABA San Diego Conquistadors*
1952-53 NBA Indianapolis Olympians*
1957-58 NBA Detroit Pistons*
1984-85 NBA Phoenix Suns*
1983-84 NBA Washington Bullets*
1986-87 NBA Golden State Warriors*
1995-96 NBA Sacramento Kings*
1996-97 NBA Los Angeles Clippers*
1966-67 NBA New York Knicks*
1985-86 NBA Chicago Bulls*
2014-15 NBA Brooklyn Nets*
1970-71 ABA Denver Rockets*
1985-86 NBA Sacramento Kings*
1948-49 BAA St. Louis Bombers*
1966-67 NBA Chicago Bulls*
1962-63 NBA Detroit Pistons*
1959-60 NBA Detroit Pistons*
1972-73 ABA San Diego Conquistadors*
1963-64 NBA Philadelphia 76ers*
1947-48 BAA Boston Celtics*
1967-68 NBA Chicago Bulls*
1991-92 NBA Miami Heat*
1978-79 NBA New Jersey Nets*
1974-75 ABA Spirits of St. Louis*
1967-68 ABA Houston Mavericks*
1959-60 NBA Minneapolis Lakers*
1973-74 ABA Virginia Squires*
1974-75 ABA Memphis Sounds*
1987-88 NBA San Antonio Spurs*
1952-53 NBA Baltimore Bullets*
1972-73 ABA New York Nets*
1949-50 NBA Sheboygan Red Skins*

On 5 it wasn't asserted that it happened often. But as I said from some people's GOAT tier Jabbar misses two consecutive in prime (between different plausible peak years). Chamberlain misses one in prime and was set to miss a second before being traded. Garnett is in many top 10s here and misses 3 straight. Robertson who has been atop published GOAT lists (though now circa 30 years ago) missed 3 in his prime. Olajuwon missed in prime, Curry did, Bryant did, Barkley (twice), Pettit ... and that's without going to people who missed very large chunks of seasons, post prime years etc.

It doesn't have to be fatal to playoff weighting depending on what claim or structure is put but a very heavy playoff focus would seem to leave some great seasons uncounted.


Fair enough. I should add, comparing regular season to postseason numbers is probably my favorite way to use the playoffs. I can't be certain that AuPM or PER or even something simple like TS% are giving me a perfect look at a player, but I can be more sure that whatever is "right" or "wrong" for that player is equally right or wrong in the regular season and the playoffs, so it should be a good basis for comparison to see who stepped up and who didn't. Now some would rightfully say, is a 10 who turns into a 9 worse than a 7 who turns into an 8 in the playoffs? No, but I think some of these guys are so close (more like 9.5 vs 9.3), that I'm willing to use the differential. I feel more confident in saying Larry Bird looks worse in the playoffs than the regular season than in trying to figure out if Bird just barely falls behind Russell and just barely ahead of Magic in the regular season.

Again thanks for civility.

In brief I don't love risers and fallers (for reasons you outline, absolute standard matters more) ... unless it's an adjustment for a primarily RS based baseline (still absolute standard is what ultimately matters but if it's to tweak an RS baseline list then ... there's a substantial place for RS performance).
Doctor MJ
Senior Mod
Senior Mod
Posts: 53,285
And1: 22,288
Joined: Mar 10, 2005
Location: Cali
     

Re: GOAT methodology for GOAT lists 

Post#47 » by Doctor MJ » Mon Jul 4, 2022 5:37 pm

Re: why does the regular season even matter?

This is a great question that bothers me a great deal. :lol:

First thing I'll say: I'm not going to say you're wrong for caring only about how good someone can be in the playoffs. I think you need to be very careful about small sample size issues if you use such an approach and to not turn your nose up at the regular season to get a better sense of what these players are, but given the way the playoffs have become the place where legacies are truly set in stone, I can't fault such an approach as unreasonable.

For myself on a season-by-season level, it's more meaningful to me to start from the regular season and adjust based on the playoffs.

Aside from the small sample size issue, here's the big thing that bugs me about ignoring the regular season:

The regular season remains the vast majority of NBA basketball played. It's the life blood of the teams in terms of revenue coming in. It's key to the downtowns who have NBA teams and all the business that feeds off the NBA. It's the basketball we actually get to watch and discuss for most of the year. And it's the place where team culture gets built and the supporting players on the team learn to be their very best.

When a player doesn't take the regular season seriously, I see it as a problem, and not something to be brushed aside. While there are cases of this that are in the debatable range, when we look at Kawhi's attitude for at least a few seasons there, he was doing significant damage on all of these fronts, and I don't see any reason to ignore this in my assessments.

Now, in the end, the playoffs can take precedence over everything else. Let's take an extreme example. Say a player chooses to only play half the season, his team has to play in the play-in, but then they proceed to beat all comers, win the chip, with it being clear that player is the best player in the playoffs. He'll be my #1 in my Player of the Year assessment, and all that other stuff won't end up mattering for that.

But if that player's team ends up not getting very far in the playoffs, possibly because they're still getting used to each other, possibly for plenty of other reasons, then even if that player looks like the best player when I watch him in the playoffs, he may not rise very much on my season lists.

Again, I have no issue with others taking a different approach, but this is what makes most sense to me as someone who values the entirety of the season of play.

Now as I say all of this: This is a thread about GOAT, and on RealGM, "GOAT" typically means career. On a career level, stuff like this seems likely to not be very significant. Maybe there's an example I'm not thinking about here that others can bring up, but as much as I think about this on a season level, I'm not sure how much it's actually influencing my GOAT list.
Getting ready for the RealGM 100 on the PC Board

Come join the WNBA Board if you're a fan!
ChartFiction
Starter
Posts: 2,178
And1: 2,828
Joined: Mar 10, 2012

Re: GOAT methodology for GOAT lists 

Post#48 » by ChartFiction » Mon Jul 4, 2022 9:29 pm

I don't think it makes any sense to use relative greatness for judging peak so much as it makes sense to use relative greatness to judge who the fastest sprinter is. No one would argue Bolt is slower than someone who demonstrated a faster relative speed for their era. Now it's harder to discern in basketball. We don't have a singular metric and there are more changing conditions and rules. But on a fundamental level I disagree with the idea that peak should be ranked relative to the time period. It just doesn't make sense.

Conditions should be considered. Stats should be considered. Context should be considered.

But all to culminate to an assessment of skill and capability. Such that we can imagine putting that skill and capability into a simulation that does a large amount of iterations of realistic scenarios to some outcome. That is a mental model for judging peak in basketball that makes sense to me. I'm not going to project new skills onto players. I'm not going to scale players from old eras to new eras and imagine them better than they actually were. That does not make any sense to me. I can not look at someone who's clearly worse at basketball and argue that their peak is better just like I can't look at someone who's clearly slower than Bolt and argue that they're a faster sprinter. To do the former is only a cop out to the limitations of basketball analytics. It's not simple enough like a measurement of speed so we give up and perform comparisons to peers only. At which point if we're doing that, we should just partition the final rankings by time periods. But we don't, its hard to because basketball time periods are continuous not discrete. So instead to our own arbitrary will, after doing comparisons only among the same time period, we then shuffle the players among different times to reach some final ranking. Which honestly is just silly.

Now a general greatest career of all time ranking I think can be done in any way. You can include on/off court considerations or anything. For example, you can consider someone the greatest because they broke ground socially. But to say they had a higher peak because this year they did some particular thing no longer makes sense. Peak is an objective assessment of capability in the sport. That's not down to a single metric or any combination of metrics in basketball. But we can use our eyes, numbers, context and some analysis. I think that approach is more in line with how we would judge any sport. It's just obfuscated enough in basketball that people have started to do, in my opinion, fundamentally flawed approaches like scaling eras to make sense of it.
Cavsfansince84
RealGM
Posts: 14,940
And1: 11,443
Joined: Jun 13, 2017
   

Re: GOAT methodology for GOAT lists 

Post#49 » by Cavsfansince84 » Mon Jul 4, 2022 10:31 pm

ChartFiction wrote:I don't think it makes any sense to use relative greatness for judging peak so much as it makes sense to use relative greatness to judge who the fastest sprinter is. No one would argue Bolt is slower than someone who demonstrated a faster relative speed for their era. Now it's harder to discern in basketball. We don't have a singular metric and there are more changing conditions and rules. But on a fundamental level I disagree with the idea that peak should be ranked relative to the time period. It just doesn't make sense.

Conditions should be considered. Stats should be considered. Context should be considered.

But all to culminate to an assessment of skill and capability. Such that we can imagine putting that skill and capability into a simulation that does a large amount of iterations of realistic scenarios to some outcome. That is a mental model for judging peak in basketball that makes sense to me. I'm not going to project new skills onto players. I'm not going to scale players from old eras to new eras and imagine them better than they actually were. That does not make any sense to me. I can not look at someone who's clearly worse at basketball and argue that their peak is better just like I can't look at someone who's clearly slower than Bolt and argue that they're a faster sprinter. To do the former is only a cop out to the limitations of basketball analytics. It's not simple enough like a measurement of speed so we give up and perform comparisons to peers only. At which point if we're doing that, we should just partition the final rankings by time periods. But we don't, its hard to because basketball time periods are continuous not discrete. So instead to our own arbitrary will, after doing comparisons only among the same time period, we then shuffle the players among different times to reach some final ranking. Which honestly is just silly.

Now a general greatest career of all time ranking I think can be done in any way. You can include on/off court considerations or anything. For example, you can consider someone the greatest because they broke ground socially. But to say they had a higher peak because this year they did some particular thing no longer makes sense. Peak is an objective assessment of capability in the sport. That's not down to a single metric or any combination of metrics in basketball. But we can use our eyes, numbers, context and some analysis. I think that approach is more in line with how we would judge any sport. It's just obfuscated enough in basketball that people have started to do, in my opinion, fundamentally flawed approaches like scaling eras to make sense of it.


The issue with bb relative to sprinting isn't just that sprinting can be easily measured in terms of time it takes to run a certain length of ground. Its that each player is part of something much larger that is vastly different when you move from one decade/era to the next. Which includes teammates, opponents, number of teams, where players are drawn from, rules and even basic things like a 3 pt line. Metrics attempt to measure all of this stuff both relative to other players in the league and in more raw terms but there's no perfect way to do it. I think for most its a combination of dominance which comes from different ways of measuring it as well as things like post season success and accolades.
ChartFiction
Starter
Posts: 2,178
And1: 2,828
Joined: Mar 10, 2012

Re: GOAT methodology for GOAT lists 

Post#50 » by ChartFiction » Mon Jul 4, 2022 10:42 pm

Cavsfansince84 wrote:
ChartFiction wrote:I don't think it makes any sense to use relative greatness for judging peak so much as it makes sense to use relative greatness to judge who the fastest sprinter is. No one would argue Bolt is slower than someone who demonstrated a faster relative speed for their era. Now it's harder to discern in basketball. We don't have a singular metric and there are more changing conditions and rules. But on a fundamental level I disagree with the idea that peak should be ranked relative to the time period. It just doesn't make sense.

Conditions should be considered. Stats should be considered. Context should be considered.

But all to culminate to an assessment of skill and capability. Such that we can imagine putting that skill and capability into a simulation that does a large amount of iterations of realistic scenarios to some outcome. That is a mental model for judging peak in basketball that makes sense to me. I'm not going to project new skills onto players. I'm not going to scale players from old eras to new eras and imagine them better than they actually were. That does not make any sense to me. I can not look at someone who's clearly worse at basketball and argue that their peak is better just like I can't look at someone who's clearly slower than Bolt and argue that they're a faster sprinter. To do the former is only a cop out to the limitations of basketball analytics. It's not simple enough like a measurement of speed so we give up and perform comparisons to peers only. At which point if we're doing that, we should just partition the final rankings by time periods. But we don't, its hard to because basketball time periods are continuous not discrete. So instead to our own arbitrary will, after doing comparisons only among the same time period, we then shuffle the players among different times to reach some final ranking. Which honestly is just silly.

Now a general greatest career of all time ranking I think can be done in any way. You can include on/off court considerations or anything. For example, you can consider someone the greatest because they broke ground socially. But to say they had a higher peak because this year they did some particular thing no longer makes sense. Peak is an objective assessment of capability in the sport. That's not down to a single metric or any combination of metrics in basketball. But we can use our eyes, numbers, context and some analysis. I think that approach is more in line with how we would judge any sport. It's just obfuscated enough in basketball that people have started to do, in my opinion, fundamentally flawed approaches like scaling eras to make sense of it.


The issue with bb relative to sprinting isn't just that sprinting can be easily measured in terms of time it takes to run a certain length of ground. Its that each player is part of something much larger that is vastly different when you move from one decade/era to the next. Which includes teammates, opponents, number of teams, where players are drawn from, rules and even basic things like a 3 pt line. Metrics attempt to measure all of this stuff both relative to other players in the league and in more raw terms but there's no perfect way to do it. I think for most its a combination of dominance which comes from different ways of measuring it as well as things like post season success and accolades.


I agree it's harder. There are more changing conditions and rules in basketball. It's more difficult to discern. I've also acknowledged this within the post.

That does not change my argument though. I fundamentally disagree with the idea that we can have two identical players, one in one time period and one in another, and you tell me that the prior one has a higher peak.
Cavsfansince84
RealGM
Posts: 14,940
And1: 11,443
Joined: Jun 13, 2017
   

Re: GOAT methodology for GOAT lists 

Post#51 » by Cavsfansince84 » Mon Jul 4, 2022 10:47 pm

ChartFiction wrote:
I agree it's harder. There are more changing conditions and rules in basketball. It's more difficult to discern. I've also acknowledged this within the post.

That does not change my argument though. I fundamentally disagree with the idea that we can have two identical players, one in one time period and one in another, and you tell me that the prior one has a higher peak.


Which is fine but relatively speaking this is an impossible theoretical scenario yet at the same time could also easily happen. All you'd have to do is take Steph Curry and put him in the pre 3 pt line era to see it happen. There are other examples as well that I think would hold true.
ChartFiction
Starter
Posts: 2,178
And1: 2,828
Joined: Mar 10, 2012

Re: GOAT methodology for GOAT lists 

Post#52 » by ChartFiction » Mon Jul 4, 2022 11:04 pm

Cavsfansince84 wrote:
ChartFiction wrote:
I agree it's harder. There are more changing conditions and rules in basketball. It's more difficult to discern. I've also acknowledged this within the post.

That does not change my argument though. I fundamentally disagree with the idea that we can have two identical players, one in one time period and one in another, and you tell me that the prior one has a higher peak.


Which is fine but relatively speaking this is an impossible theoretical scenario yet at the same time could also easily happen. All you'd have to do is take Steph Curry and put him in the pre 3 pt line era to see it happen. There are other examples as well that I think would hold true.


Right and that problem exists no matter what. I think your approach to solving that problem is fundamentally wrong to a grave error. And given the disparity that we see in sports, at least ones that haven't had talent siphoned from it, from recent performance to dated ones, you're likely committing an egregious error by using a blind scale factor. That is to say if our goal is to measure peak the same way we would in a simply measured sport, which I think it is, the only reason we've deviated is the difficulty. But then we've started to get nonsensical output that now measures something completely different. I'm not even sure what it sensibly measures.
Cavsfansince84
RealGM
Posts: 14,940
And1: 11,443
Joined: Jun 13, 2017
   

Re: GOAT methodology for GOAT lists 

Post#53 » by Cavsfansince84 » Mon Jul 4, 2022 11:10 pm

ChartFiction wrote:
Right and that problem exists no matter what. I think your approach to solving that problem is fundamentally wrong to a grave error. And given the disparity that we see in sports, at least ones that haven't had talent siphoned from it, from recent performance to dated ones, you're likely committing an egregious error by using a blind scale factor. that is to say if our goal is to measure peak the same way we would in a simply measured sport. which I think it is. I think the only reason we've deviated is the difficulty. but then we've started to get nonsensical output that now measures something completely different. I'm not even sure what it sensibly measures.


I'm not entirely sure what approach you are referencing here or what approach you are actually endorsing either. I'm guessing you prefer one that greatly favors who played more recently but I'd be glad to hear exactly what you think is the best way to measure peaks.
ChartFiction
Starter
Posts: 2,178
And1: 2,828
Joined: Mar 10, 2012

Re: GOAT methodology for GOAT lists 

Post#54 » by ChartFiction » Mon Jul 4, 2022 11:44 pm

Cavsfansince84 wrote:
ChartFiction wrote:
Right and that problem exists no matter what. I think your approach to solving that problem is fundamentally wrong to a grave error. And given the disparity that we see in sports, at least ones that haven't had talent siphoned from it, from recent performance to dated ones, you're likely committing an egregious error by using a blind scale factor. that is to say if our goal is to measure peak the same way we would in a simply measured sport. which I think it is. I think the only reason we've deviated is the difficulty. but then we've started to get nonsensical output that now measures something completely different. I'm not even sure what it sensibly measures.


I'm not entirely sure what approach you are referencing here or what approach you are actually endorsing either. I'm guessing you prefer one that greatly favors who played more recently but I'd be glad to hear exactly what you think is the best way to measure peaks.


The dominance of era then shuffle around different eras approach to rankings. There's no coherent thought process to this approach that withstands any thorough questioning. It's fundamentally illogical at best and rife with error at worst. And you still have the problem of comparing players with different rules.

I find it more reasonable to rank peaks by engaging in a cognitive exercise of imagining players in realistic scenarios. Where you might find it unfair is that where a major break occurs, like the 3 point line, I would judge on what's current. And I wouldn't have any issues doing that because of the large difference in skills between the two sets of players prior/post. If you think that's unfair, I would find it much more reasonable to partition on that break and create two different rankings than to take the first approach mentioned. Which I find unreasonably unfair and erroneous and destroys any purpose of the ranking as far as I can understand it.
Cavsfansince84
RealGM
Posts: 14,940
And1: 11,443
Joined: Jun 13, 2017
   

Re: GOAT methodology for GOAT lists 

Post#55 » by Cavsfansince84 » Mon Jul 4, 2022 11:54 pm

ChartFiction wrote:
The dominance of era then shuffle around different eras approach to rankings. There's no coherent thought process to this approach that withstands any thorough questioning. It's fundamentally illogical at best and rife with error at worst. And you still have the problem of comparing players with different rules.

I find it more reasonable to rank peaks by engaging in a cognitive exercise of imagining players in realistic scenarios. Where you might find it unfair is that where a major break occurs, like the 3 point line, I would judge on what's current. And I wouldn't have any issues doing that because of the large difference in skills between the two sets of players prior/post. If you think that's unfair, I would find it much more reasonable to partition on that break and create two different rankings then take the first approach mentioned. Which I find unreasonably unfair and erroneous and destroys any purpose of the ranking as far as I can understand it.


Ok well keep in mind what you are referencing as my method is just the method I think is used by some. Its not necessarily my personal method that I use. I'm curious how your method actually works in practice though. As in who your top 3 peaks are and then how you came to that top 3 if you want to elaborate. Not so much so I can argue it with you but just so I can get a better idea of what you are endorsing.

Return to Player Comparisons