APBRmetrics

Championship criteria

C1: 2 +5s (or near)?
C2: 3 or more +3s (or near)?
C3: 4-5 over +2?
C4: 6-8 over +1?

How does it look in this framework this season? I'll try with Darko, though the scale may not be the same as previously mainly done with BPM.


Celtics:

1 over +5 on Darko
2 over +3 and a 3rd close
5 over +2
5 over +1

So, fully meets 1 old criteria based on BPM, 2 close, 1 not.


Nuggets

1 over +5 on Darko
1 over +3 and a 2nd close
5 over +2
5 over +1

So, fully meets 1 old criteria based on BPM, 1 close, 2 not.

Thunder
0 over +5 on Darko
1 over +3 and none close
2 over +2
4 over +1

So, fully meets 0 old criteria based on BPM and none close.


Need to compare to others but so far not that impressed with the frontrunners.

To adjust the BPM values to an appropriate scale you should divide them by 1.2. This is due to the fact that BPM has a factor of 1.2 used to, on paper, handle the elastic band effect.

I fundamentally disagree that BPM, or any statistic, should include such a factor, but I have made those arguments before and won't seek to re-hash them.

But it is worth noting: BPM values are inflated by a factor of 1.2 when compared with a plus minus statistic such as DARKO.

Thanks for sharing that insight.

Timberwolves

0 over +5 on Darko
2 over +3 and 1 close
3 over +2
4 over +1

So, fully meets 0 old criteria based on BPM and 1 close.



Adjusting the criteria from BPM to Darko based on v-zero comment, I'd say Celtics measure up but Thunder and Timberwolves still don't and not close.


Clippers

1 over +5 on Darko and 2 somewhat close
3 well over +3
3 over +2
4 over +1

So, fully meets 1 old criteria based on BPM and 1 close.

Clippers not fully in tier 1 by this approach but better than the other Tier 2s checked so far.


Bucks

1 over +5 on Darko
2 over +3 and none close
4 over +2
4 over +1

So, fully meets 1 old criteria based on BPM and no others close. Tier 2.

Maybe I'll check others later, but I don't expect to find another Tier 1 and probably not anyone closer to Tier 1 than Clippers.

v-zero wrote: ↑Thu Mar 07, 2024 3:38 pm To adjust the BPM values to an appropriate scale you should divide them by 1.2. This is due to the fact that BPM has a factor of 1.2 used to, on paper, handle the elastic band effect.

I fundamentally disagree that BPM, or any statistic, should include such a factor, but I have made those arguments before and won't seek to re-hash them.

But it is worth noting: BPM values are inflated by a factor of 1.2 when compared with a plus minus statistic such as DARKO.

I suspect the factor is larger than 1.2 between BPM and DARKO, because DARKO is predictive and thus will have some additional regression. In other words it won't sum up to the team's total output.

True, but broadly speaking a statistic like DARKO will do a good job of summarising a team's current performance level in a given moment when taken in sum. The most shrinkage will appear when the level of uncertainty is highest, at the start of the season. At this point in the season DARKO is both predictive and descriptive.

More regression adjustment in Darko, maybe Clippers become legit Tier 1. Others get closer but probably not enough.

I think OKC have the talent, but I think they're about two years early.

"no champion since the 2013–2014 season has ever had their best player be worse than a top 8 player, their second best player be worse than a top 24 player and their third best player be worse than a top 39 player."

From https://medium.com/@johnchenmbb/the-nba ... 66675623d8


Forgot about this by round assessment in middle of page by same author

viewtopic.php?t=10002

Teams with 7th to 12th best player can have quite different round advancement. Not that likely to win it all, but out in 1st, 2nd or 3rd round depends heavily on rest of talent and matchups.

Did the Celtics have a top 8 player this season?
Tatum was #14-15 in BPM, WS/48 and PER this year. 10th in WS and VORP, thanks to big minutes.
Porzingis #4 in WS/48 but only @ 21 mpg. Hardly a factor in playoffs.

On the BPM list they had 4 in the top 25,
5 in the top 50,
6 in the top 65,
7 in the top 75.
Jaylen Brown is #80 and #8 on the team! So they had 1/10 of the best players at any level, or 3 times what an avg team would have.

Tatum, #2 on Darko.

But Celtics example does make an argument in favor of fabulous depth, in addition to having a top guy.

8 of the other 9 in top 10 on Darko did not make conference finals and only 1 other wasn't well short.

By the introduced 4 criteria on this page, the Mavs met none and were only close on 1. Not previously scored because did not expect it would be necessary.



Will anyone besides Celtics be strong on this criteria set next season?

I should score the new Sixers.


They are reasonably close on 3.