GOAT relative team offenses?

[Doctor MJ]

Wait, 1956 Lakers played 10 playoff games? Is that a type or am I misremembering things?

Edit, i see, it’s just the bolder that are 10 games.

Yeah, I don’t think 3 games samples work for me. If you don’t actually win a series, then that means this sort of stat summary would imply both teams were all-time offenses and they just happened to play in the first round. Possible, but unlikely.


Sent from my iPhone using Tapatalk

[eminence]

Wait, 1956 Lakers played 10 playoff games? Is that a type or am I misremembering things?


Sent from my iPhone using Tapatalk

Nah, they’re in the not camp (3 games), I was being silly and poking a bit of fun at even a full playoff run being a not very great measure in these situations.

[70sFan]

Just a heads up...

Defense (min. 10 games):

1. 1964 Celtics: -12.4
2. 2004 Pistons: -11.8
3. 2004 Pistons: -11.8
4. 2022 Bucks: -10.8
5. 2000 Heat: -9.5

So good they made it twice?

Yeah, they were that good

Kidding, thanks for finding my mistake.

[Doctor MJ]

The 2015 Clippers being there is amazing with the obstacle of Deandre Jordan's FT shooting. Teams would hack him all the time and slow down their offense. They were the worst FT shooting team every year.

Do rememberthat they didn’t hack-a-De all season. It was largely a playoff thing.

Not saying the Clips don’t deserve credit, just saying that most of the data is based on games opponents don’t care enough to use (and possibly use up) that trchnique.


Sent from my iPhone using Tapatalk

[TrueLakerFan]

For those interested I have a list of year by year leaders with relative ratings for All-Season data (RS + PS) as one on of the columns where that data was available. The first sheet is actually focused on the top player by ORtg & DRtg going back to '96-97, and that's not mentioned so far, so let me package some stuff here to go along with the actual spreadsheet.

Among players who have led the league in ORtg while playing >50% of their team's minutes (so only the top guy each season), here are the ones with the biggest rORtg (based on bkref):

1. Steve Nash (Phx) '04-05 +13.9
2. Steph Curry (GS) '16-17 +13.2
3. Chris Paul (LAC) '14-15 +11.4
4. Steve Nash (Phx) '06-07 +11.3
(tie) Steph Curry (GS) '15-16 +11.3
(tie) Nikola Jokic (Den) '22-23 +11.3 (so far)
7. Dirk Nowitzki (Dal) '03-04 +11.0
8. James Harden (OKC) '11-12 +10.8
9. Steve Nash (Phx) '09-10 +10.6
10. Kawhi Leonard (LAC) '20-21 +10.5

For rDTG:

1. David Robinson (SAS) '98-99 -11.5
2. Jaren Jackson Sr. (SAS) '97-98 -10.4
3. Tim Duncan (SAS) '03-04 -10.1
4. Kendrick Perkins (BOS) '07-08 -9.9
(tie) Giannis Antetokounmpo (MIL) '19-20 -9.9
6. Alonzo Mourning (MIA) '96-97 -9.4
7. David Robinson (SAS) '00-01 -8.6
(tie) Kawhi Leonard (SAS) '15-16 -8.6
9. Kevin Garnett (BOS) '10-11 -8.5
10. Bruce Bowen (SAS) '06-07 -8.4

And the leaders in terms of leading in the most years by these standards:

ORtg
1. Steve Nash (PHX) 7 times
2. John Stockton (UTA) 3
(tie) Dirk Nowitzki (DAL) 3
(tie) Chris Paul (LAC) 3
5. Steph Curry (GS) 2
(tie) Kevin Durant (GS) 2

DRtg
1. Tim Duncan (SAS) 3 times
2. David Robinson (SAS) 2
(tie) Kevin Garnett (BOS) 2
(tie) Draymond Green (GS) 2
(tie) Giannis Antetokounmpo (MIL) 2

Goes to show defense wins championships

[penbeast0]

Per thinkingbasketball.net:

Could you write the best defences? I want to see if one of my theories will be confirmed

What theory?

1. 2004 Spurs: -8.8
2. 1964 Celtics: -8.7
3. 2008 Celtics: -8.6
4. 1993 Knicks: -8.3
5. 1994 Knicks: -8.1
6. 2020 Bucks: -7.7
7. 2004 Pistons: -7.5
8. 2016 Spurs: -7.4
9. 2014 Pacers: -7.3
10. 2005 Spurs: -7.3
11. 1999 Spurs: -7.2
12. 1965 Celtics: -7.2
13. 2011 Bulls: -6.9
14. 2011 Celtics: -6.9
15. 2007 Bulls: -6.7
16. 2006 Spurs: -6.6
17. 2007 Spurs: -6.5
18. 2009 Magic: -6.4
19. 2012 Celtics: -6.4
20. 1989 Jazz/2012 Bulls/1975 Bullets/1963 Celtics: -6.3

Surprised the Russell Celtics only show up 2.25 times.

[Ein Sof]

Per thinkingbasketball.net:

Could you write the best defences? I want to see if one of my theories will be confirmed

What theory?

1. 2004 Spurs: -8.8
2. 1964 Celtics: -8.7
3. 2008 Celtics: -8.6
4. 1993 Knicks: -8.3
5. 1994 Knicks: -8.1
6. 2020 Bucks: -7.7
7. 2004 Pistons: -7.5
8. 2016 Spurs: -7.4
9. 2014 Pacers: -7.3
10. 2005 Spurs: -7.3
11. 1999 Spurs: -7.2
12. 1965 Celtics: -7.2
13. 2011 Bulls: -6.9
14. 2011 Celtics: -6.9
15. 2007 Bulls: -6.7
16. 2006 Spurs: -6.6
17. 2007 Spurs: -6.5
18. 2009 Magic: -6.4
19. 2012 Celtics: -6.4
20. 1989 Jazz/2012 Bulls/1975 Bullets/1963 Celtics: -6.3

Interesting that 14/23 of the best defenses ever existed in a 15-year span, from 1999 to 2014. Seems like a weird coincidence.

Some of those seasons even had multiple GOAT defenses, apparently...

[Lost92Bricks]

Do rememberthat they didn’t hack-a-De all season. It was largely a playoff thing.

Not saying the Clips don’t deserve credit, just saying that most of the data is based on games opponents don’t care enough to use (and possibly use up) that trchnique.


Sent from my iPhone using Tapatalk

Not as much as in the playoffs and the 2016 season but they would still do it as soon as he had an easy layup. They were 3rd last in FT% and 4th in most attempts that year.

[csh 19792001]

Could you write the best defences? I want to see if one of my theories will be confirmed

What theory?

1. 2004 Spurs: -8.8
2. 1964 Celtics: -8.7
3. 2008 Celtics: -8.6
4. 1993 Knicks: -8.3
5. 1994 Knicks: -8.1
6. 2020 Bucks: -7.7
7. 2004 Pistons: -7.5
8. 2016 Spurs: -7.4
9. 2014 Pacers: -7.3
10. 2005 Spurs: -7.3
11. 1999 Spurs: -7.2
12. 1965 Celtics: -7.2
13. 2011 Bulls: -6.9
14. 2011 Celtics: -6.9
15. 2007 Bulls: -6.7
16. 2006 Spurs: -6.6
17. 2007 Spurs: -6.5
18. 2009 Magic: -6.4
19. 2012 Celtics: -6.4
20. 1989 Jazz/2012 Bulls/1975 Bullets/1963 Celtics: -6.3

Surprised the Russell Celtics only show up 2.25 times.

Yeah, me too. That really doesn't make any sense.

https://docs.google.com/spreadsheets/d/1DRXJdAr15iZmulqF0F_6SnxrB_b_PPFPM543ke30-qM/edit#gid=0

[70sFan]

What theory?

1. 2004 Spurs: -8.8
2. 1964 Celtics: -8.7
3. 2008 Celtics: -8.6
4. 1993 Knicks: -8.3
5. 1994 Knicks: -8.1
6. 2020 Bucks: -7.7
7. 2004 Pistons: -7.5
8. 2016 Spurs: -7.4
9. 2014 Pacers: -7.3
10. 2005 Spurs: -7.3
11. 1999 Spurs: -7.2
12. 1965 Celtics: -7.2
13. 2011 Bulls: -6.9
14. 2011 Celtics: -6.9
15. 2007 Bulls: -6.7
16. 2006 Spurs: -6.6
17. 2007 Spurs: -6.5
18. 2009 Magic: -6.4
19. 2012 Celtics: -6.4
20. 1989 Jazz/2012 Bulls/1975 Bullets/1963 Celtics: -6.3

Surprised the Russell Celtics only show up 2.25 times.

Yeah, me too. That really doesn't make any sense.

https://docs.google.com/spreadsheets/d/1DRXJdAr15iZmulqF0F_6SnxrB_b_PPFPM543ke30-qM/edit#gid=0

As I said, different pace estimations give you different ORtg and DRtg numbers. Here are Ben's defensive numbers for the 1960s Celtics:

1957: 86.9 DRtg, -2.5 rDRtg
1958: 85.6 DRtg, -3.1 rDRtg
1959: 87.6 DRtg, -3.5 rDRtg
1960: 86.6 DRtg, -3.7 rDRtg
1961: 83.5 DRtg, -5.5 rDRtg
1962: 85.0 DRtg, -6.1 rDRtg
1963: 88.3 DRtg, -6.3 rDRtg
1964: 83.8 DRtg, -8.7 rDRtg
1965: 84.7 DRtg, -7.2 rDRtg
1966: 88.2 DRtg, -4.5 rDRtg
1967: 90.4 DRtg, -4.3 rDRtg
1968: 92.0 DRtg, -2.8 rDRtg
1969: 88.7 DRtg, -4.6 rDRtg

One important thing I have to add - due to much smaller league, these relative numbers vastly underestimate Celtics advantage over the league average - by the sheer fact that Celtics played over 10% of all NBA games in most of these seasons. If we take 1964 (their best season), the league average by this estimate is 92.5 DRtg, but Celtics DRtg has a weight of 11% for the mean. If we exclude their success from calculations, we get 93.6 DRtg and that would give Celtics by far the best rDRtg ever at -9.8. It's not necessary to do that for modern teams, because the league is significantly bigger now and one team has a minor impact on the mean. In short, here is one example:

1961 Celtics with them in mean calculations: -5.5 rDRtg
1961 Celtics without them in mean calculations: -6.3 rDRtg
Difference: -0.8

1964 Celtics with them in mean calculations: -8.7 rDRtg
1964 Celtics without them in mean calculations: -9.8 rDRtg
Difference: -1.1

2008 Celtics with them in mean calculations: -8.6 rDRtg
2008 Celtics without them in mean calculations: -8.9 rDRtg
Difference: -0.3

It was statistically harder to be a huge outlier in a smaller league, which makes 1960s Celtics defense even more impressive.

[DCasey91]

Wait correct me if I’m wrong but smaller means more of a chance at greater discrepancy vs bigger meaning less so.

In other words in a smaller group you would more likely see a greater discrepancy than a sample data of a bigger group? Because a bigger group would have more consistent ratios to the norm of or a ball curve of whatever you are looking at.

That’s basic group sampling for such things as IQ, age, sex etc etc within a group of people. it would be the same when looking at numbers too. Or I’m completely wrong lol

[Doctor MJ]

For those interested I have a list of year by year leaders with relative ratings for All-Season data (RS + PS) as one on of the columns where that data was available. The first sheet is actually focused on the top player by ORtg & DRtg going back to '96-97, and that's not mentioned so far, so let me package some stuff here to go along with the actual spreadsheet.

Among players who have led the league in ORtg while playing >50% of their team's minutes (so only the top guy each season), here are the ones with the biggest rORtg (based on bkref):

1. Steve Nash (Phx) '04-05 +13.9
2. Steph Curry (GS) '16-17 +13.2
3. Chris Paul (LAC) '14-15 +11.4
4. Steve Nash (Phx) '06-07 +11.3
(tie) Steph Curry (GS) '15-16 +11.3
(tie) Nikola Jokic (Den) '22-23 +11.3 (so far)
7. Dirk Nowitzki (Dal) '03-04 +11.0
8. James Harden (OKC) '11-12 +10.8
9. Steve Nash (Phx) '09-10 +10.6
10. Kawhi Leonard (LAC) '20-21 +10.5

For rDTG:

1. David Robinson (SAS) '98-99 -11.5
2. Jaren Jackson Sr. (SAS) '97-98 -10.4
3. Tim Duncan (SAS) '03-04 -10.1
4. Kendrick Perkins (BOS) '07-08 -9.9
(tie) Giannis Antetokounmpo (MIL) '19-20 -9.9
6. Alonzo Mourning (MIA) '96-97 -9.4
7. David Robinson (SAS) '00-01 -8.6
(tie) Kawhi Leonard (SAS) '15-16 -8.6
9. Kevin Garnett (BOS) '10-11 -8.5
10. Bruce Bowen (SAS) '06-07 -8.4

And the leaders in terms of leading in the most years by these standards:

ORtg
1. Steve Nash (PHX) 7 times
2. John Stockton (UTA) 3
(tie) Dirk Nowitzki (DAL) 3
(tie) Chris Paul (LAC) 3
5. Steph Curry (GS) 2
(tie) Kevin Durant (GS) 2

DRtg
1. Tim Duncan (SAS) 3 times
2. David Robinson (SAS) 2
(tie) Kevin Garnett (BOS) 2
(tie) Draymond Green (GS) 2
(tie) Giannis Antetokounmpo (MIL) 2

Goes to show defense wins championships

I understand why you'd interpret this data that way, but there's a lot more going on than that.

[dygaction]

Nash is still the offensive GOAT and Giannis' defense is underrated... what's happening to Perkins?

[70sFan]

Wait correct me if I’m wrong but smaller means more of a chance at greater discrepancy vs bigger meaning less so.

In other words in a smaller group you would more likely see a greater discrepancy than a sample data of a bigger group? Because a bigger group would have more consistent ratios to the norm of or a ball curve of whatever you are looking at.

That’s basic group sampling for such things as IQ, age, sex etc etc within a group of people. it would be the same when looking at numbers too. Or I’m completely wrong lol

As I said, it's not true in pure mathematical sense, because one outlier result has a massive impact on mean in smaller groups, influencing the mean greatly. That's not the case in larger groups.

Besides, even if simply look at NBA history, outlier teams happened significantly more often in big league than small league.

[DCasey91]

But the proviso is smaller group means larger chance of discrepancy it maybe why Ben adjusted it who knows. Yes but what is the consensus of a large grouping in NBA. 20? 30? Then if the bigger groups have been the norm then it’s the basis to go off as the results should be more consistent. Outliers always happen because you need high/low in the bell curve distribution. And your right one result can skew the data if it’s small. That’s why results like that can be noisy at best.

There is more meat on the average gradient for obvious reasons as opposed to having a small sample group as there’s more of an expectation to have extreme data points because the maping can be all over the place with less.

It wouldn’t take long but (probably someone already has done it) an every year distribution of teams relative to whatever you are looking for. Then you can cross reference it to your hearts desire.

In mathematical terms there’s a log formula for cross comparing data X with data Y and gives you a ratio or % of how much of an outlier something actually is relative to the overall data sets for all years. It’s a more thorough approach imho.

Fwiw math is weird as we aren’t accounting for range of group size as in the NBA it can change from year to year.

Some predicted models are based something similar to this as well as general populous data because mapping a whole population for example takes way too long.

[NO-KG-AI]

I dont think listing relative offensive ratings is the best way to Guage the best offenses ever.

Offenses today are the best offenses ever--they shoot the most 3's, they operate with the most spacing, the teams contain the most and best shooters and the strategies are optimized to be as efficient as possible.

Meanwhile, defenses understand all this. They understand how to defend offenses, where to defend and what shots to allow yet offenses still are at an all-time high.

Very true.

This is pretty true, but I also think that almost everyone, if they have to choose, will put out a guy who can space the floor over a guy who can only defend, if they don't have enough guys to do both. I think the early to mid 2000's was the opposite. I still think the way the game is officiated is a lot easier to score... but it's for the better. I'm not at all advocating going back to the slog that it was.

[penbeast0]

Wait correct me if I’m wrong but smaller means more of a chance at greater discrepancy vs bigger meaning less so.

In other words in a smaller group you would more likely see a greater discrepancy than a sample data of a bigger group? Because a bigger group would have more consistent ratios to the norm of or a ball curve of whatever you are looking at.

That’s basic group sampling for such things as IQ, age, sex etc etc within a group of people. it would be the same when looking at numbers too. Or I’m completely wrong lol

As I said, it's not true in pure mathematical sense, because one outlier result has a massive impact on mean in smaller groups, influencing the mean greatly. That's not the case in larger groups.

Besides, even if simply look at NBA history, outlier teams happened significantly more often in big league than small league.

Assuming for me, small means 15 or less, when you say outlier teams happened more significantly, I assume you mean only W/L because the Russell Celtics were pretty extreme and consistent outliers in terms of Drtg.

I also think that a large part of that might be the league adapting to a much greater importance in the draft after the NBA/ABA merger. First team to get accused of tanking for draft status that I remember was Houston and it was shocking with howls of outrage (and only for a few games at the end of the season). Now it's widely accepted as a strategy by a good quarter of the league and for the whole season. That creates a lot of extra disparity.

[70sFan]

Wait correct me if I’m wrong but smaller means more of a chance at greater discrepancy vs bigger meaning less so.

In other words in a smaller group you would more likely see a greater discrepancy than a sample data of a bigger group? Because a bigger group would have more consistent ratios to the norm of or a ball curve of whatever you are looking at.

That’s basic group sampling for such things as IQ, age, sex etc etc within a group of people. it would be the same when looking at numbers too. Or I’m completely wrong lol

As I said, it's not true in pure mathematical sense, because one outlier result has a massive impact on mean in smaller groups, influencing the mean greatly. That's not the case in larger groups.

Besides, even if simply look at NBA history, outlier teams happened significantly more often in big league than small league.

Assuming for me, small means 15 or less, when you say outlier teams happened more significantly, I assume you mean only W/L because the Russell Celtics were pretty extreme and consistent outliers in terms of Drtg.

I also think that a large part of that might be the league adapting to a much greater importance in the draft after the NBA/ABA merger. First team to get accused of tanking for draft status that I remember was Houston and it was shocking with howls of outrage (and only for a few games at the end of the season). Now it's widely accepted as a strategy by a good quarter of the league and for the whole season. That creates a lot of extra disparity.

I meant by Net Rating, the bigger league is the higher difference is between the average and best/worst teams.