APBRmetrics

This may be a good place for some future posts.

Detail is not necessarily trivia.

By 6 metrics, LeBron James varies by about 2 pts on each of offensive and defensive impact estimates and 4 altogether. The average is about +4 offensive, +1 defense and +5 overall for last season.

The low outlier was... LEBRON, the high outlier MAMBA, the closest to the simple blended average was 3 year RAPM.

The rank in league distribution was 7 to 19, with an average of just over 11th.

Any subjective evaluators or other metric followers want to comment?

LEBRON suggests he was a below average defender, all others have him as top 20% and one in top 10%.

There is something that could have discussion, further analysis.

Will it?

Metric histories show him 20-60% off his scored peaks last season.

Once again, I'd probably use a blend, simple average or intelligently weighted.

Relative to league average. LBJ is above average scorer and passer, below average offensive rebounder and just slightly negative on turnover rate (but likely better by A/TO). His at the rim frequency and accuracy are both high and retain much of his peak rates.

Team defensive rating with him on court has been better than league average only 4 times in career. Neutral last season, good positive 2 years ago, quite negative in the bubble.

#1 on 28 Year Lifetime RAPM, more on offense.

His on-ball time tied career low last season and was about 30% off his peak.

It appears his bad pass rate and the turnover adjusted efg% were both career highs, meaning getting better results despite more errors. Rim assist rate near a career high.

I wouldn’t look into the raw numbers quite as much, while the thing I made had a high number for him it was a tad lower on his best years than other things, but that’s pretty much dependent on how heavily ur pulling towards the box score prior, my original sigma had him pretty much with Curry, in that mid 2010s period even though it was RS only.

While it’s a bit more complicated than “creating bias in place of noise” since inherently you’re guiding it towards the right answers vs everyone to 0, you still have players different things will overshoot and undershoot but guys whose box score production underrates or overrated their legit impact are always going to be a bit harder done, imo lebron more accurately measured from things like TDRAPM I think

Doncic last season: highest on MAMBA, lowest on time-decayed RAPM. 3rd to 33rd best.

Average across 6 metrics, +5.45 on offense, +0.45 on defense. About +5.9 overall.

In last playoffs, worst rTS% of his career at slightly negative.

On a stat that may become public in near future, Durant is estimated to add the most points to his team based on scoring in playoffs (adjusted to recognize value of high usage volume) in modern era. Curry, Leonard, James, Jokic among those close.

Embid less than half that. Raw TS% falls 3.5%pts from regular season to playoffs. Last playoffs was barely 3rd best in 7 playoffs.

Gobert and Ayton surprisingly high and well above Embid. Lillard slightly behind Embid.

The biggest impact improvements against strong teams vs. weak are generally tough perimeter role players and stars are totally absent from the top 40.

Pau Gasol, C Anthony, D Howard, Siakam and Gobert are among the biggest name older players for nearly biggest drop-offs against strong teams compared to weak. Among the newer names, even higher: Jalen Green, SGA, Randle, S. Barnes and LaVine.

FWIW the mamba numbers might change a bit I have to rerun some things

Not because theres any glaring error but because my computer was lost lol, I still have the coeffecients for the priors and finished the Possession data and my code for the metric itself is on Collab. but honestly I thought it might be nice to redo everything just in case anyways

On DiVincenzo, BPM is highest at +3.4 last season, followed by MAMBA. 3 year RAPM is lowest at -1.8, because it has the most from earlier when he played worse. Darko and LeBron are closest to metric average (short of +1).

KAT, MAMBA and BPM are again relatively high with EPM. LeBron moves to lowest. Average is a bit above +2. I'm more accepting of the average here and in other cases than any one metric on its own.

By Darko projection, the worst rookie by end of last season was essentially Keyonte George. Jazz took him 16th. 59th on my draft board.


Alexander Sarr had a brother in NBA and that might have given him some credibility beyond other reasons but his brother was in bottom 10% by Darko.


Worst decline by Darko last season was Beverley. Gone. 2nd worst, C Paul. $11 million contract by Spurs.

D Rose, 5th biggest drop and knew it was time to go.

Tyus Jones, 11th biggest drop. But starting for Budenholzer and the Suns.

Vassell, 30th biggest drop here but almost everyone loves him and sees bigger and better ahead. I liked him before draft but not for awhile.

T Young, 34th biggest drop.

So much to see. Especially for teams, if interested.

Giannis, Siakam, Klay Thompson in bottom 10% on gross Darko change. Lillard and LeBron not far off. Adebayo and Durant in around bottom 15%.

Adebayo declining, who else noticed or believes that? Almost 27.5 years old, so not that surprising to be past peak.

Anybody else find some of these things notable?

On recent Darko improvement, beyond Jalen Williams and Holmgren, were Fox, KCP, Suggs, Kessler, Ellis and Brunson.

Darko improvement appears to be since start of 2022-23 season. Intended that way or not.




Jovic is also a big riser.

Why Dalen Terry rose a lot on Darko, beyond getting more minutes, was not immediately clear. BPM fell sharply. From end of season to end of season. But I now understand Darko isn't that way at this time. Probably an oversight. Not explicitly stated.

Average year to year change on Darko would be good to know. If any analyzed the data for that.

Is Scottie Pippen Jr. done improving? We'll find out. A prediction would take more work. Does his team have a prediction?

On 7 metrics, Cade Cunningham's impact is estimated between +1.2 and -1.5 and basically always has been.

That is not a good or average #1. Or a #2.

Could claim he is still a beginner at the role but he has had some time and isn't that far along on impact. Raw stats but raw stats need interpretation.

Last chance to break out coming up, in my view. Well maybe not the last chance but I'll say that over still patient.

Ryan Dunn pre-season 3pt shooting:

Virginia was a low volume, low success 3pt environment and low success offensive environment. It had a poor 3pt shooting PG and its 3pt shooters were not high profile.

Dunn was a defensive first player for a defense first Coach and focused on defense.

He took an extra year in high school and did not have a strong offensive record at all.
But he did have some 3pt success:
https://x.com/HoopKeesee/status/1800653271515713996

He isn't 22 yet.

Suns in preseason is a high volume 3pt environment with probably plenty of spacing and room to get shot off.

Dunn is 55% on corner 3s and 38% (barely above overall league average) from above the break. It is trivial volume but so far his 3pt success above average is mostly about success in right corner.

In regular season his defensive focus will probably be higher, usage lower and ability to get corner 3 should be pressured more, if defensive scouting advice is given and taken.

But he will be with high profile 3pt shooters, a PG who can shoot 3s and a Coach who likes / generates a high 3 pt volume and success environment and a broad one.

Someone could calculate likelihood of preseason results achieved with different “true 3pt shooting” assumptions.

Not sure if degree of contest data is available for preseason but I haven't found it yet. I'd guess it matters a lot. But the early shots had variety.


Some comments by Suns GM James Jones about his shooting here:

https://arizonasports.com/story/3549716 ... nba-draft/

He made a bet and looking good right now.

10-15 teams worked him out. Knicks had 2 chances and did other stuff. Thunder went hard for Dillon Jones. Lots could have had him and went other directions. Presumably mostly or entirely based on conventional scouting.

I debated whether to take the strong defense as shown by BPM and RAPM as "enough". I didn't in the end. Maybe next time. I sort of did this with Mogbo though.

One more thing to note: Dunn's offense overall in college last season was much worse against top 60 than the full schedule.

https://x.com/bballstrategy/status/1845771919548985577


If it real or not, look at the 10-20 best cases of a 3pt game coming out of nowhere. What are the most common elements or signs? Is plus height / length one of the more important ones? Coaching bad to good or 3pt not good or emphasized to good and emphasized? What?


Lots of people talking about this case. How many read this? Anybody write mire or better on this yet? Where? Link?

Adebayo declining, who else noticed or believes that? Almost 27.5 years old, so not that surprising to be past peak.

Bam improved most of his numbers from 2023; best Ast% and Blk% of his last 3 seasons, best DRb% and TO% of his career. Usg% near career high, and he's suddenly a credible 3-point threat (36%).
https://www.basketball-reference.com/pl ... s_advanced

He was mostly a F 5 years ago, has transitioned to full time C. Taking more shots, passing better.
His highest WS/48 and BPM were in shortened seasons. 2024 was first appearance on 1st team all-D.

For Adebayo, the decline is modest by Darko measures, in last 2 years and mainly on offense (including declining fta rate and at the rim rates / projections). Defense was at best in fall '22, fell but recovered late last season. EfG% and TS% last season were lowest since rookie. Shot versatility is one thing, overall results is another.