Pretty popular J Vanderbilt is worst by luck adjusted RAPM.
Lakers and Pistons with 3 each in bottom 20.
LA-RAPM estimates no player in league worth +4 points. That suggests to me it is too tightly regressed.
Just 13 players worth 2 points? Just 70 total worth 1 point on total impact? Not believable.
If just 70 players are worth 1 point, the best 10 coaches should be paid $20 mil. / yr.
Million dollar analysts could be worth that or far more.
'22-23 Commentary
-
DarkStar48
- Posts: 116
- Joined: Sat Oct 09, 2021 1:24 am
Re: '22-23 Commentary
Yeah, Vanderbilt with the Jazz was actually having his worst statistical season overall on both ends since becoming a rotation player in the 2020-21 season. In addition, his early +/- data in LAL has not been favorable.
Out of the 3 Lakers in the LA-RAPM Bottom 20, Walker IV is unlikely to get regular playing minutes going forward. Of those, only Beasley has had both the box stats profile as well as +/- values with the new team to show promise so far.
RAPM uses ridge regression to produce a more stable, though biased, squared error estimator. However, 1-season play-by-play lineup data is still too small a sample to give precise estimates of player On/Off rating. That is why you have to use 3-year, 5-year or 6-year datasets to calculate results that are closer to reality.
Ultimately, single year RAPM is nonetheless helpful for not only evaluating the impact relative to others, but also the sign of that change.
Out of the 3 Lakers in the LA-RAPM Bottom 20, Walker IV is unlikely to get regular playing minutes going forward. Of those, only Beasley has had both the box stats profile as well as +/- values with the new team to show promise so far.
RAPM uses ridge regression to produce a more stable, though biased, squared error estimator. However, 1-season play-by-play lineup data is still too small a sample to give precise estimates of player On/Off rating. That is why you have to use 3-year, 5-year or 6-year datasets to calculate results that are closer to reality.
Ultimately, single year RAPM is nonetheless helpful for not only evaluating the impact relative to others, but also the sign of that change.
Re: '22-23 Commentary
I had fallen out of the habit of checking multi-year RAPM, partly because context can change a lot, partly because I was lead to believe that the 1 year average errors were now quite small compared to distant past RAPM (but never actually saw such data).
3 and 5 year RAPM have 16 and 17 players over +4. That is better imo than none but I am unsure if it is "enough". I'll accept it til I think about it further. 89 and 65 respectively over +2. That is better.
I'll probably return to checking 3 year more.
Sexton, Maledon, Knox and RJ Hampton in bottom 10 on both 3 yr RAPM and 3 yr LA-RAPM.
A Simons in bottom 10 of 3 yr LA-RAPM. Up to 301 on 1 yr RAPM and 371 on 1 yr LA-RAPM. Still a ridiculous contract by RAPM by folks that either didn't know or hardly cared.
All these names in bottom 11 on 5 yr LA-RAPM as well.
Thunder with 5 in bottom 50 here including current player Tre Mann. (4 more at least briefly on team in bottom 100.) 3 in top 19 and 5 in top 75 all departed (at least til Roberson possibly returns). Presti's draft pick Bazley, 985. My very strong and vocal preference at that draft lead up who was left on board in favor of Bazley Brandon Clarke, 88.
Consensus highly ranked draft prospects I didn't agree with / didn't like in bottom 50 here included K Johnson, P Wlliams and Reddish.
Rockets really went out of way to develop K Porter. 974 on 5 yr LA-RAPM. 398 on 1 yr RAPM, 498 on 1 yr LA-RAPM. Still extremely low. Jalen Green in bottom 100 on both 5 years. Near bottom 25 on both 1 years.
Tyler Herro, 943 on 5 yr RAPM. 130 on 1 yr RAPM, 109 on 1 yr LA-RAPM. Probably still an overpay. Herro and A Simons appear to be highest current pay in 5 year LA-RAPM bottom 100. Yay points can mislead. And on that subject, Jamal Crawford lead last over recent 5 years.
A neutral rating gets you about 450th place on 5 yr LA-RAPM. -1, 900th. If average error is at least one then ranking of 450 or 900 is not that reliably meaningfully different.
3 and 5 year RAPM have 16 and 17 players over +4. That is better imo than none but I am unsure if it is "enough". I'll accept it til I think about it further. 89 and 65 respectively over +2. That is better.
I'll probably return to checking 3 year more.
Sexton, Maledon, Knox and RJ Hampton in bottom 10 on both 3 yr RAPM and 3 yr LA-RAPM.
A Simons in bottom 10 of 3 yr LA-RAPM. Up to 301 on 1 yr RAPM and 371 on 1 yr LA-RAPM. Still a ridiculous contract by RAPM by folks that either didn't know or hardly cared.
All these names in bottom 11 on 5 yr LA-RAPM as well.
Thunder with 5 in bottom 50 here including current player Tre Mann. (4 more at least briefly on team in bottom 100.) 3 in top 19 and 5 in top 75 all departed (at least til Roberson possibly returns). Presti's draft pick Bazley, 985. My very strong and vocal preference at that draft lead up who was left on board in favor of Bazley Brandon Clarke, 88.
Consensus highly ranked draft prospects I didn't agree with / didn't like in bottom 50 here included K Johnson, P Wlliams and Reddish.
Rockets really went out of way to develop K Porter. 974 on 5 yr LA-RAPM. 398 on 1 yr RAPM, 498 on 1 yr LA-RAPM. Still extremely low. Jalen Green in bottom 100 on both 5 years. Near bottom 25 on both 1 years.
Tyler Herro, 943 on 5 yr RAPM. 130 on 1 yr RAPM, 109 on 1 yr LA-RAPM. Probably still an overpay. Herro and A Simons appear to be highest current pay in 5 year LA-RAPM bottom 100. Yay points can mislead. And on that subject, Jamal Crawford lead last over recent 5 years.
A neutral rating gets you about 450th place on 5 yr LA-RAPM. -1, 900th. If average error is at least one then ranking of 450 or 900 is not that reliably meaningfully different.
Re: '22-23 Commentary
By crude hoopsstats.com measure, Rockets' backcourt ranked 28th. Frontcourt 22nd.
Thunder backcourt 2nd, frontcourt 29th.
Mavs 1st backcourt (will it get better with Kyrie?)
Nuggets, Celtics, Bucks 1, 2, 3 on frontcourt. 8, 26, 16 on backcourt. Front courts and front court players certainly matter.
Pistons, 28th backcourt, 26th frontcourt.
Spurs, 25th backcourt, 30th frontcourt.
Lakers, 27th backcourt, 9th frontcourt.
Timberwolves, 22nd backcourt (essentially before trade), 3rd in frontcourt (even with Towns for a long time).
Celtics, only team in top 8 on both.
Thunder backcourt 2nd, frontcourt 29th.
Mavs 1st backcourt (will it get better with Kyrie?)
Nuggets, Celtics, Bucks 1, 2, 3 on frontcourt. 8, 26, 16 on backcourt. Front courts and front court players certainly matter.
Pistons, 28th backcourt, 26th frontcourt.
Spurs, 25th backcourt, 30th frontcourt.
Lakers, 27th backcourt, 9th frontcourt.
Timberwolves, 22nd backcourt (essentially before trade), 3rd in frontcourt (even with Towns for a long time).
Celtics, only team in top 8 on both.
Re: '22-23 Commentary
Jazz, 23rd backcourt, 7th frontcourt.
Re: '22-23 Commentary
Bucks with just 2 lineups used much more than the equivalent of 1 minute per game and both barely over 3 minutes per game. Last regular season had just 1, Last playoffs had 2 over 4 min / GM and a bunch over 1 minute but that concentration was pretty low.
Championship regular season has more concentration and playoffs way more (top lineup over 12 minutes per game). Why did they depart from that formula? Concentration level in next playoffs? Something to watch. Hear any discussion of that in Bucks media? If so, tell me / show me.
Championship regular season has more concentration and playoffs way more (top lineup over 12 minutes per game). Why did they depart from that formula? Concentration level in next playoffs? Something to watch. Hear any discussion of that in Bucks media? If so, tell me / show me.
Re: '22-23 Commentary
Spurs with just 2 lineups over 1 min / GM (ridiculous and common apparently) and none over about 3 min / gm. 2 best bigger lineups exercised by roster moves. Best remaining tested just 0.5 min / gm. Learning almost nothing from 5 man results.
10 most used pairs all bad to horrendous. McDermott - Collins just mildly bad. Will it be back next season? Then it is back to bad to horrendous for rest of 20 most used.
What was learned of positive value?
679 lineups and 181 pairs thrown out there.
1000 lineups for season unlikely but still possible, especially with roster moves. 900 has a pretty good prospect.
10 most used pairs all bad to horrendous. McDermott - Collins just mildly bad. Will it be back next season? Then it is back to bad to horrendous for rest of 20 most used.
What was learned of positive value?
679 lineups and 181 pairs thrown out there.
1000 lineups for season unlikely but still possible, especially with roster moves. 900 has a pretty good prospect.
Re: '22-23 Commentary
Compared to its 2918-19 team that went to Western Conference Finals, the current Blazers shoot better. They are worse at taking care of the ball and much lesser rebounding. Is rebounding a big enough part of the design? Not in the implementation so far. Not in Simons, Sharpe, Little, Winslow, Johnson, Reddish and especially not GRANT. Or Thybulle.
Re: '22-23 Commentary
Anthony Edwards, tied for lead on minutes, leads league on FGAs, 264th on efg%.
Top 18 in league on efg% (minute qualified) are 17 centers and a PF.
(But let's try an efficiency challenged perimeter player in the draft... say too many GMs.)
Top 18 in league on efg% (minute qualified) are 17 centers and a PF.
(But let's try an efficiency challenged perimeter player in the draft... say too many GMs.)
Re: '22-23 Commentary
Hawks have been strong with 3 biggest lineups, especially the first. An average of +8.8 pts / 48 min. But -3.8 on average for everything else negated the strength. New coach probably has to clean-up the dink game. A good part of that is increasing concentration on the best & biggest but it will probably have to go beyond that.
Some of the biggest problem pairs involve Murray and Okongwu but there also bad ones with Griffin, Bogdanovic and A. Holiday.
Some of the biggest problem pairs involve Murray and Okongwu but there also bad ones with Griffin, Bogdanovic and A. Holiday.
Re: '22-23 Commentary
Biggest problems for Jazz have been own turnovers and defensive rebounding. Trades could have other effects positive and negative at factor level and otherwise but may go some way toward improving both of these trouble spots.
Re: '22-23 Commentary
Lillard with career highs for season efg% and ts%. But will it be enough?
Re: '22-23 Commentary
Harden - Embid, +9 / 100p. One of the best but a handful in 20 most used for team were better, though surely often when 1 or 2 is out there as well unnamed. Barely +1 when pair wasn't not on court.
Re: '22-23 Commentary
Best and worst player development cases for Wizards in last 2-3 years?
Best is probably Porzingis, though how much credit goes to the player and how much to the player development program us an open / unanswerable from outside question.
Worst might be Kuzma, or Johnny Davis or Hachimura / Avdija or (others).
Best is probably Porzingis, though how much credit goes to the player and how much to the player development program us an open / unanswerable from outside question.
Worst might be Kuzma, or Johnny Davis or Hachimura / Avdija or (others).
Re: '22-23 Commentary
Pistons fall to 4-16 in last 20.
11 of 13 most used lineups negative. Biggest minute one horrendous. 90% of 20 most used pairs negative, most near -10 / 100p or worse.
Only 2 players much better than -1 on BPM are over 30.
11 of 13 most used lineups negative. Biggest minute one horrendous. 90% of 20 most used pairs negative, most near -10 / 100p or worse.
Only 2 players much better than -1 on BPM are over 30.