APBRmetrics

His +/- per 100 the last 4 seasons is over +5, and his on-off is over +4. RAPM last 3 seasons over +3 (ranked 33rd in league).

I think the box score weaknesses you've cited are the same flaws that many models (BPM, DARKO, etc.) are overweighing with Dillon Brooks. His pbp data consistently shows positive impact, and his box score shows a dud. With as much data as we have on him, I'll take the pbp data.

I didn't have his RAPM to look at when I posted the other information, but that is an important part of the positive story.


Every case, if pursued with an open-mind, "should" look at everything and think about it and beyond. Considerations can be gathered by one, two or more.

Crow wrote: ↑Sat Dec 23, 2023 9:58 pm I didn't have his RAPM to look at when I posted the other information, but that is an important part of the positive story.


Every case, if pursued with an open-mind, "should" look at everything and think about it and beyond. Considerations can be gathered by one, two or more.

I hear you and agree. Curious: after infinite games, how much would you weigh box scores vs. some RAPM? I would way the box scores as zero.

Here's my thinking: box score weighing starts around 100% (game 1), gradually fades at first, then accelerates to something like 50% after a couple seasons, levels out at around 10% after several seasons, before asymptotically approaching zero. And RAPM follows the opposite curve, asymptotically approaching 100%.

I dunno after infinite games. It would make a difference in average errors if it was infinite games against the same 29 teams or more / infinite teams.

Within a single season, especially a partial season, I often go 2/3rds boxscore, 1/3rd RAPM.

The impact of a thug is almost by definition not sustainable. Poison, cancer, rot.

Mavs lineups:

https://x.com/bballstrategy/status/1742 ... 46749?s=20

Rockets lineup management:

Pair data is pretty good but 7 of 9 most used 5 mans are negative.

Most used is good and over 10 minutes per game but then only 1 other is over the equivalent of 2 minutes per game.

322 lineups used. IMO about 250 too many. Only 4 lineups over 1 minute / game and just 13 total over the super dink threshold of 0.5 min / gm.

Best big minute quad is VanVleet Brooks Smith Sengun. Need a new center or a wing or both?

Several other prominent players may have cases to keep, but they aren't heavily based on current team +/-. Eason is an exception.

Among starters, VanVleet has the clear best raw +/-s on court and on /off. Jalen Green the clear worst. The other 3 ok, but not strong.

Rockets try 66 new lineups in last 8 games. Go 2W-6L in that time.

Jaren Jackson over 25 pts / gm in last 10 games but also under 4 rebounds.

Rockets won last game to pull up to 3W - 9L in last 12. BRef gives about 2.5% to make playoffs outright and 10% chance at making play-in.

Not what they wanted / expected with aggressive moves last summer. No immediate help at deadline. Adams for next season. What other changes? VanVleet and Brooks, among others, would be included in trade talks for me. Didn't want either on those contracts.

Brooks' raw plus minus on and on / off is 9th on team for season. The on has been twice as bad this month. Darko projection trend has been straight down since early last season. Now in negative territory. Is the trend reversible? I am skeptical.

Spurs:

434 lineups.
1 short of 5 minutes per game.
5 more modestly over 1 min/gm but not 2 min/gm.
428 dinks.

Are Blazers closer to 5 or 10 moves away from playoffs?

Spurs?

Season is 80% over. Scoot Henderson at -6.4 BPM. There are some worse (including Hood-Schifino) but no lottery pick worse at end of season in recent times besides the Wizards pick Johnny Davis.

Nuggets are 8-2 in last 10. Lineup concentration on 5 most used is way up. But the starters performance has fallen completely off cliff and 3 of the 4 others go without one of Jokic or Murray and add in Braun, Watson, Jackson and / or Nnaji .

I'll want to study further but this appears to be a hidden crisis. Always had concerns about bench (though it appears some part of that rotation beyond the 5 most used must be working recently, first glance suggests good results in Porter and Holiday lineups) but did not know starters ground down to almost neutral recently.

There are few pairs actually negative recently and none involving 2 starters in big minutes. But the 2 most meh big minute pairs involve Gordon.

Still, the primary issue is the starting unit fall off. It appears to be partly in shooting efficiency, partly in rebounding and probably a lot in defense.