Post#23 » by Chinook » Thu May 15, 2025 12:08 am
Plus-minus is a box-score stat, just as much as FG% is. RAPM is weird, because in many ways, it's the most advanced stat there is. But it also doesn't have the same derivative and weighted nature that more traditional advanced stats do.
RAPM is rarely wrong. Its problem is that it doesn't have the use a lot of people want it to have. It is basically a career-long stat rather than a season or especially a game stat. If a guy is having a breakout year, you likely wouldn't be able to tell based on their RAPM, even once it gets updated. To try to account for that inertia, folks have made "improved" versions of RAPM like RPM that try to work box-score stats into the formula to be more immediately useful. But that's dumb, because RAPM's momentum is what makes it's so good in the first place. The stat is good, but people misuse it a lot.
(I'm not commenting on Keldon here. I'm just talking about the stat in general. I think all of the Spurs' advanced stats will look better if they could win more games. That seems like I have causality backwards, but bad players on good teams will often have better stats than players of equal caliber on bad teams. RAPM in particular tries to account for this, but it requires a LOT of data to make it work. Hence why it's so plodding.)