APBRmetrics

Yes, this would work. The challenge is how to implement the shrinkage to a value that is non-zero. The ridge regression algorithms in the R or Python packages by definition shrink towards zero. There are several ways to do priors; the methodology that I have worked with is to simply subtract out the...

It was remarkable how low Bam's game scored with Box Plus/Minus. In essence, this was because his true shooting percentage for the game was about the same as his teammates. In other words, BPM didn't see him producing a whole lot of marginal value per shot. Most of the other very high scoring games ...

The team level value should just be the difference in expected points for a possession starting with a live ball turnover versus a standard possession. Of course, the individual value (credit for the turnover) would be some factor less than the team value of a live ball turnover, since the player is...

I laid out a true BPM 2.0 projection system in this post:
viewtopic.php?p=37451#p37451

I just haven't gotten around to actually doing them ahead of the season.

I also have better priors now than those mentioned in that post

Yes.^^ I messed around with age adjustment and decided (again) not to bother. Also "yes" for WShr, perW, and eWin. Today's top-to-bottom spread is smallest yet! Down from 2.47 on Jan 9, to 1.91 now. The whole field has improved since then, by an avg of .60 in MAE: 1.04 KPel .89 DQin .88 C...

Hello, I have not done any positional scarcity research anytime recently. When using lineup data (on/off) to assess player value, the data is inherently assessing relative to position or at least to what player would replace you. Now, at the end of the positional spectrum, this leads to a fairly cle...

Thank you for putting these together as always, Mike!

I'm wondering if there is a minutes weighting vs. Possession weighting discrepancy in adding up the 7 games. I can't remember how that's done at the moment.

In other words... Do we weight by game or by minute or by possession?

The minutes weighted average BPM for a playoff game should be exactly equal to the minutes weighted average of those players regular season BPMs.

So if the average BPM was slightly lower in game two that indicates a few more minutes for slightly lower regular season BPM players.

Crow wrote: ↑Fri Apr 25, 2025 7:19 pm https://shiny.sradjoker.cc/NBA-Playoff-Probabilities/

Wow, that's a very interesting tool!

Kawhi last night in game 2 vs. the Nuggets was on the edge of this list.

Game score 37.4, BPM 24.5, 39 minutes played in a 3 point win. GmBPM of 19.9.

Here were my picks, realized I hadn't posted them here yet.

Thunder in five
Nuggets in seven
Lakers in seven
Warriors in six

Cavs in five
Pacers in five
Knicks in six
Celtics in four

We can just double the scoring for round 2, triple for CF, quadruple for the Finals. If you're behind in later rounds, you might pick the underdogs. It makes sense to me to guess each round as they arrive. Doing all in advance as you did, you leave default picks for the next round, if you don't cha...

“between the two methodologies - % of production* per game - predicting RAPM and/or +/- stint data“ This is an important distinction. Percent of production is a pretty self-evident approach, but it doesn't correlate directly to point differential on the court and thus wins (the ultimate goal). The o...

Another monster game microburst last night, including a new #1 for the season. Top 10 chronologically: gsbg date who min TS% Pt Rb As St Bk TO GmSc BPM 31.1 11.10 Jokic 38 .75 37 18 15 3 0 4 44.8 27.2 32.7 11.13 Giannis 44 .71 59 14 7 2 3 3 54.2 21.7 31.2 12.15 Luka 38 .83 45 11 13 3 2 3 48.7 25.3 ...