RealGM 2020 Top 100 Project: #1 (LeBron James)

[falcolombardi]

i am gonna comment a bit on the playoffs vs regular season debate from my point of view

djoker point of view is that playoff data is noisy compared to regular season...which i disagree with

if anythingh i believe regular season data to be MORE noisy than playoffs, the term we use so much -sample size- is more about the selection than the sheer quantity, good sampling is not just increasing the number of data points but selecting the more REPRESENTATIVE ones

sample sizes are kind of a reverse-exponential (is there a word for that?) thingh, a 20k people poll is not 10 times more reliable than a 2k people poll amd such. A 2k people poll with good methodology to pick the people has less noise than a 20k people poll with bad sampling

why do i say this?, cause playoffs are both, the more reliable part of a player career to know how good they really were AND the more important part of their careers too

i think reg season games are by all definition worth less than post season, they are less valuable in regards to competing for rings AND less valuable to evaluate the player (despite there being more of them)

think of all the things we use to evaluate players greatness/goodness/career value

mentality and ability to play in high stakes: this is playoffs by far

ability to beat strong teams: playoffs by far, at worst you get a 1st round vs an average team, in reg season half of your games will be against teams worse than this.

reg season play measures how well you play against both bad and good or trash and great teams and averages them out, less predictive on ability to win rings (the best team at cleaning up weak teams over the marathonic reg season may not be the best one vs strong teams, the ones you have to win to compete for championship)

health/availability: injuries happen in both, but players save their health for playoffs, load management games, back to backs etc, dissapear so teams play their best available teams nearly 100% of the time. no stting out stars in b2b's or pulling starters out and conceding early

Also players play more mins than in reg season increase their minutes, meaning the teams play closer to their 100% best, no energy saved

tactics/adaptation: teams (both, coach and olayers) have enough time to gameplan for each rival, as opposed to reg season where teams cannot dedicate themselves fully in a tactical sense to each rival
they also are not traveling constantly so there is more time.for practices too

team success (ring, finals runs, etc): self explanatory, teams dont hang best regular season banners

individual accolades (mvp, all nba, dpoy, all-D) the only thingh that is defined by reg season play

playoffs is when teams, coaches, players are all at their best and against the best of their rivals and when they are tested the most. is not only the most important games but the more.reliable ones to know how good they really are

we are talking a lot about sample sizes, the thingh is, sample sizes is as much about correct sampling than it is about the sample...well, size, regular season data is noisy in a different way

is both quality AND quantity. playoffs provides the former and over a full prime (1-3 seasons worth of playoffs games depending on the player) the quantity of playoff games is there too

that is why i think anythingh less than 50% of a player evaluation being his playoffs (to account for the difference in quality of play) play is erroneous. ideally more (imo).

I'm sorry but you're wrong about sample size.

Sample size definitely matters a lot. Variance in performance can be very very big and you need large sample to balance this effect out. Think of tossing a coin. There is 50/50 odds of a head or a tail but after 10 tosses you will often get 8 heads and 2 tails or 3 heads and 7 tails etc. With a 100 tosses you might get 47-53 ratio of a 59-41 ratio but the data is already more reliable. And with a 1000 tosses you will get very close to 500-500.

What you wrote about effort, tactics, no sitting out etc. can be used to discredit regular season data. Ainosterhaspie I think wrote about that in the other thread and I can actually agree. The regular season data has potentially questionable value.

But that still doesn't make the playoff data reliable. The sample is just so so so much smaller. Over a course of 10 seasons it becomes less noisy but the summary of your findings, for the years you chose, was an edge of +0.5 possessions per 100 in rORtg for Lebron's teams. It's ridiculous to take a gap of 0.5 possessions as meaningful over a sample of even 100-200 games. It's like saying a player who scores 30.5 ppg over 2 seasons is noticeably better than one who scores 30 ppg over 2 seasons. It just makes no sense. The two are the same for all intents and purposes.

And then the playoff data also has an issue of overrepresenting certain opponents. In 2017 and 2018 for instance, Lebron's Cavs faced only four teams: Pacers, Raptors, Celtics and Warriors. That's four out of a possible twenty nine opponents.

it comes down to where each one draws the line, imo a 150-250 range of playoff games is very reliable at that point (bill walton prime is literally evaluated with less games than that and nobody questions it) but it is open to disagreement of course

for player evaluation purposes i think reg season matters more for players who dont have many playoff games like lebron/jordan played

but even there i usually would like to consider their playoff data too (limited as it may be) to detect common trends. like for big underperformers i will take their reg season numbers with a bit of caution if they take too big of a dip in playoffs

now back to lebron vs jordan stuff. you are right that is a marginal difference...but that is the thingh. jordan and lebron, specially prime to prime, are for almost all intents and purposes comparable players with small differences between each other

jordan scoring gap or lebron playmaking gap or whoever you consider a better defender (lebron for me)are not huge unsurmountable gulfs between them

yet we still count them cause otherwise we would just have said "they are close enough in all factors that is pretty dumb to try to declare one is better" and we wouldnt even make lists like this and only would rank players by tiers

on opponent overrepresentation....sure, but is over representation of good teams compared to reg season mix of everythingh, is a minor problem imo

[Dr Positivity]

I was going to post this in the Greatest Peaks thread before the discussion switched away from Jordan vs. Lebron. This debate has kind of been done to death but to me the only argument for Lebron is longevity-based. I honestly don't think there is any reasonably strong case that Lebron was a better or even equal player to Jordan when both were at their best. Due to the two players' consistency we can compare very large playoff samples and thus minimize noise. Shorter samples of 5-6 years favor Jordan even more but If we take 10-year playoff primes there's still a clear gap in basic production:

Playoffs Per Game Totals

1986-1996 Jordan: 34.1 ppg, 6.5 rpg (1.7 o), 6.1 apg, 2.2 spg, 0.9 bpg on 57.7 %TS (+4.0 rTS) with 3.2 topg in 41.7 mpg
2009-2018 Lebron: 29.2 ppg, 9.1 rpg (1.6 o), 7.0 apg, 1.8 spg, 1.0 bpg on 59.1 %TS (+4.9 rTS) with 3.5 topg in 41.5 mpg

Playoffs Per 75 Possessions:

1986-1996 Jordan: 32.9 ppg, 6.3 rpg (1.6 o), 5.9 apg, 2.1 spg, 0.9 bpg on 57.7 %TS (+4.0 rTS) with 3.1 topg
2009-2018 Lebron: 28.4 ppg, 8.9 rpg (1.5 o), 6.8 apg, 1.7 spg, 1.0 bpg on 59.1 %TS (+4.9 rTS) with 3.4 topg

The 4.5+ ppg gap is glaring. That's just a huge difference in scoring volume evident from those numbers with all else on offense being relatively equal. The scoring volume differentiates their primes and puts them on separate tiers offensively IMO.

I'm not going to assume that Jordan is better on offense just cause he scores more points. Lebron is one of the best scorers I've ever seen, that he didn't score 30ppg every year is as reflective of his team orientated style of play as much as anything else, I'm not convinced 32ppg is always better than 28ppg or that it's better for the team to have someone who likes shooting as much as Jordan does. This is reflected in the passing numbers, while you could argue the triangle was not great for Jordan's assists much like Kobe, I would argue Lebron's talent as a passer is overall higher.

[Djoker]

i am gonna comment a bit on the playoffs vs regular season debate from my point of view

djoker point of view is that playoff data is noisy compared to regular season...which i disagree with

if anythingh i believe regular season data to be MORE noisy than playoffs, the term we use so much -sample size- is more about the selection than the sheer quantity, good sampling is not just increasing the number of data points but selecting the more REPRESENTATIVE ones

sample sizes are kind of a reverse-exponential (is there a word for that?) thingh, a 20k people poll is not 10 times more reliable than a 2k people poll amd such. A 2k people poll with good methodology to pick the people has less noise than a 20k people poll with bad sampling

why do i say this?, cause playoffs are both, the more reliable part of a player career to know how good they really were AND the more important part of their careers too

i think reg season games are by all definition worth less than post season, they are less valuable in regards to competing for rings AND less valuable to evaluate the player (despite there being more of them)

think of all the things we use to evaluate players greatness/goodness/career value

mentality and ability to play in high stakes: this is playoffs by far

ability to beat strong teams: playoffs by far, at worst you get a 1st round vs an average team, in reg season half of your games will be against teams worse than this.

reg season play measures how well you play against both bad and good or trash and great teams and averages them out, less predictive on ability to win rings (the best team at cleaning up weak teams over the marathonic reg season may not be the best one vs strong teams, the ones you have to win to compete for championship)

health/availability: injuries happen in both, but players save their health for playoffs, load management games, back to backs etc, dissapear so teams play their best available teams nearly 100% of the time. no stting out stars in b2b's or pulling starters out and conceding early

Also players play more mins than in reg season increase their minutes, meaning the teams play closer to their 100% best, no energy saved

tactics/adaptation: teams (both, coach and olayers) have enough time to gameplan for each rival, as opposed to reg season where teams cannot dedicate themselves fully in a tactical sense to each rival
they also are not traveling constantly so there is more time.for practices too

team success (ring, finals runs, etc): self explanatory, teams dont hang best regular season banners

individual accolades (mvp, all nba, dpoy, all-D) the only thingh that is defined by reg season play

playoffs is when teams, coaches, players are all at their best and against the best of their rivals and when they are tested the most. is not only the most important games but the more.reliable ones to know how good they really are

we are talking a lot about sample sizes, the thingh is, sample sizes is as much about correct sampling than it is about the sample...well, size, regular season data is noisy in a different way

is both quality AND quantity. playoffs provides the former and over a full prime (1-3 seasons worth of playoffs games depending on the player) the quantity of playoff games is there too

that is why i think anythingh less than 50% of a player evaluation being his playoffs (to account for the difference in quality of play) play is erroneous. ideally more (imo).

I'm sorry but you're wrong about sample size.

Sample size definitely matters a lot. Variance in performance can be very very big and you need large sample to balance this effect out. Think of tossing a coin. There is 50/50 odds of a head or a tail but after 10 tosses you will often get 8 heads and 2 tails or 3 heads and 7 tails etc. With a 100 tosses you might get 47-53 ratio of a 59-41 ratio but the data is already more reliable. And with a 1000 tosses you will get very close to 500-500.

What you wrote about effort, tactics, no sitting out etc. can be used to discredit regular season data. Ainosterhaspie I think wrote about that in the other thread and I can actually agree. The regular season data has potentially questionable value.

But that still doesn't make the playoff data reliable. The sample is just so so so much smaller. Over a course of 10 seasons it becomes less noisy but the summary of your findings, for the years you chose, was an edge of +0.5 possessions per 100 in rORtg for Lebron's teams. It's ridiculous to take a gap of 0.5 possessions as meaningful over a sample of even 100-200 games. It's like saying a player who scores 30.5 ppg over 2 seasons is noticeably better than one who scores 30 ppg over 2 seasons. It just makes no sense. The two are the same for all intents and purposes.

And then the playoff data also has an issue of overrepresenting certain opponents. In 2017 and 2018 for instance, Lebron's Cavs faced only four teams: Pacers, Raptors, Celtics and Warriors. That's four out of a possible twenty nine opponents.

it comes down to where each one draws the line, imo a 150-250 range of playoff games is very reliable at that point (bill walton prime is literally evaluated with less games than that and nobody questions it) but it is open to disagreement of course

for player evaluation purposes i think reg season matters more for players who dont have many playoff games like lebron/jordan played

but even there i usually would like to consider their playoff data too (limited as it may be) to detect common trends. like for big underperformers i will take their reg season numbers with a bit of caution if they take too big of a dip in playoffs

now back to lebron vs jordan stuff. you are right that is a marginal difference...but that is the thingh. jordan and lebron, specially prime to prime, are for almost all intents and purposes comparable players with small differences between each other

jordan scoring gap or lebron playmaking gap or whoever you consider a better defender (lebron for me)are not huge unsurmountable gulfs between them

yet we still count them cause otherwise we would just have said "they are close enough in all factors that is pretty dumb to try to declare one is better" and we wouldnt even make lists like this and only would rank players by tiers

on opponent overrepresentation....sure, but is over representation of good teams compared to reg season mix of everythingh, is a minor problem imo

Just because Walton's peak is judged on 100ish regular season games and 20 playoff games doesn't mean that's optimal. I don't have much faith in that data for one. There is no double standard here where I'm moving goalposts.

The other gaps like playmaking/defense etc. are marginal but the scoring gap isn't marginal at all. 4.5 points per 75 is a huge gap. Like I said it's approximately the difference between peak Shaq around 2000-2001 and peak Duncan around 2002-2003. I never heard anyone say those two are close in scoring.

Again if you're so insistent on team data and want to stick to your guns namely the +0.5 edge in rORtg then please post the rDRtg numbers as well. Let's see whose teams had better defenses too... Offense and defense don't exist in a vacuum anyways like VanWest pointed out.

[Djoker]

I was going to post this in the Greatest Peaks thread before the discussion switched away from Jordan vs. Lebron. This debate has kind of been done to death but to me the only argument for Lebron is longevity-based. I honestly don't think there is any reasonably strong case that Lebron was a better or even equal player to Jordan when both were at their best. Due to the two players' consistency we can compare very large playoff samples and thus minimize noise. Shorter samples of 5-6 years favor Jordan even more but If we take 10-year playoff primes there's still a clear gap in basic production:

Playoffs Per Game Totals

1986-1996 Jordan: 34.1 ppg, 6.5 rpg (1.7 o), 6.1 apg, 2.2 spg, 0.9 bpg on 57.7 %TS (+4.0 rTS) with 3.2 topg in 41.7 mpg
2009-2018 Lebron: 29.2 ppg, 9.1 rpg (1.6 o), 7.0 apg, 1.8 spg, 1.0 bpg on 59.1 %TS (+4.9 rTS) with 3.5 topg in 41.5 mpg

Playoffs Per 75 Possessions:

1986-1996 Jordan: 32.9 ppg, 6.3 rpg (1.6 o), 5.9 apg, 2.1 spg, 0.9 bpg on 57.7 %TS (+4.0 rTS) with 3.1 topg
2009-2018 Lebron: 28.4 ppg, 8.9 rpg (1.5 o), 6.8 apg, 1.7 spg, 1.0 bpg on 59.1 %TS (+4.9 rTS) with 3.4 topg

The 4.5+ ppg gap is glaring. That's just a huge difference in scoring volume evident from those numbers with all else on offense being relatively equal. The scoring volume differentiates their primes and puts them on separate tiers offensively IMO.

I'm not going to assume that Jordan is better on offense just cause he scores more points. Lebron is one of the best scorers I've ever seen, that he didn't score 30ppg every year is as reflective of his team orientated style of play as much as anything else, I'm not convinced 32ppg is always better than 28ppg or that it's better for the team to have someone who likes shooting as much as Jordan does. This is reflected in the passing numbers, while you could argue the triangle was not great for Jordan's assists much like Kobe, I would argue Lebron's talent as a passer is overall higher.

We are judging players individually. The guy who scores 4.5 points per 75 more with all else being more or less equal is clearly the better offensive player. The structure of the team is another issue and whether sharing the scoring load is more optimal. The answer is it depends but given that Jordan won 6 titles including several very dominant runs putting up 33+ points per 75 I'd venture a guess that it was quite optimal.

[Homer38]

Sometimes the numbers are very misleading in the regular season because in the last few weeks of the season, the regular season often means nothing and this is where sometimes the stats can drop .... Like just 2 examples. ... Last year the lakers were 4th in offense before the bubble but they dropped to 11th position in the 8 seeding games

In 2013 the heat were first most of the regular season but in the final 2-3 week of the season they rested LBJ and others when they were sure they were going to have the best NBA record and thunder. beat the heat just at the end

The heat had also dropped a lot in the rankings in the last few week in 2012.The playoffs, it is what matters and the offenses with LBJ have for the most part almost always been elite and been at their best.