Alex Caruso's 2025-26 on/off net rating vs. DataBaller's individual impact model

OKC was +17.8 per 100 with Caruso on court last season, but DataBaller's Player Impact v1 model ranks him 151st of 582 players at +1.07 per 48 minutes—illustrating the gap between lineup-level and individual metrics.

Published: September 21, 2026 · Figures as of September 21, 2026

Asked: What’s estimated impact of Alex Caruso in terms of impact on net rating ?

Caruso's on-court net rating: OKC was +17.8 per 100 with him on the floor, well above his own team's already elite mark

Reading this the way the metric is meant to be read: it's the team's scoring margin per 100 possessions while Caruso was playing, not an individual rating. Over last season (2025-26), across 56 games and 2,178 of his on-court possessions, Oklahoma City outscored opponents by 17.8 points per 100 possessions with Caruso on the floor. For context, OKC's full-season net rating (all minutes, every player) was +11.0, so the team ran about 6.8 points per 100 hotter than its own season average whenever he was in the game. League-wide, players clearing a games floor average -0.5 net rating per 100 with a spread of about 6.9, so Caruso's +17.8 sits roughly 2.7 standard deviations above a typical qualified player - among the best marks in the league on this measure.

The caveat that has to travel with that number: on-court net rating is a lineup signal keyed to Caruso, not a measurement of what he individually did. It's inflated by playing alongside Oklahoma City's other stars and against whatever OKC's bench-heavy opponents were running, and the registry is explicit that this isn't an adjusted, isolate-the-player rating (no RAPM/EPM here).

DataBaller's own individual-impact model, Player Impact v1, tries to strip some of that out with a minutes-weighted regression toward league-average - and it reads him far more modestly: +1.07 points of margin impact per 48 minutes versus a league-average player, ranked 151st of 582 rated players, with full confidence given his minutes sample. That model explicitly warns its individual values are compressed toward zero and shouldn't be used to rank or crown players - it's a lineup ingredient, not a verdict on Caruso.

So the honest summary: the team performs exceptionally well with Caruso on the floor (+17.8 net rating, a top-of-league on/off number), but the model built to separate his own contribution from his teammates' is far more measured (+1.07/48, middling among rated players) and neither number should be read as "how good Caruso is" on its own - each measures a different, narrower thing.

Confidence: High on the on-court number itself (large sample, exact stat); Medium on what it says about Caruso individually, since the box-score model built to isolate his own impact reads him as an ordinary rotation-level contributor rather than a driver of that margin.