Skip to content
Descriptive only — these are measured per-36 rates, not projections.
NBA players · NBA · as of 2026-04-12

Stephon Castle

Historical measurements from this snapshot.

Corpus measurements: 2 seasons in the snapshot, 147 games, 4,143.1 minutes; snapshot 2026-04-12

Assists per 36 minutes, corpus seasons
7.2
Percentile rank 95 among 482 measured profiles
pct 95 of 482 measured
Field-goal percentage, corpus seasons
44.9%
Percentile rank 44 among 482 measured profiles
pct 44 of 482 measured
Three-point percentage, corpus seasons
30.5%
Percentile rank 19 among 482 measured profiles
pct 19 of 482 measured
Free-throw percentage, corpus seasons
72.9%
Percentile rank 29 among 482 measured profiles
pct 29 of 482 measured
Games played, corpus seasons
147
Percentile rank 49 among 482 measured profiles
pct 49 of 482 measured
Minutes played, corpus seasons
4143.1
Percentile rank 63 among 482 measured profiles
pct 63 of 482 measured
Points per 36 minutes, corpus seasons
20
Percentile rank 82 among 482 measured profiles
pct 82 of 482 measured
Rebounds per 36 minutes, corpus seasons
5.7
Percentile rank 50 among 482 measured profiles
pct 50 of 482 measured
Seasons in corpus
2
Percentile rank 18 among 482 measured profiles
pct 18 of 482 measured
Stored percentiles rank measured rows in this published pack; higher is only higher, not a quality judgement.

Also measured elsewhere

Joined onto 330 of 482 player cards; the leaderboard badges cover far fewer.

Team win rate with/without this player
SAS: active 0.7500 vs missed 0.7857 (delta -0.0357)
n_active=68, n_missed=14, ci95_offset=[-0.2383, 0.2383]
WITH/WITHOUT = ROSTER CONFOUND: this is NOT a causal player-impact estimate. Why a player missed games is entangled with injuries clustering, rest-game scheduling, blowout pulls, and a different replacement lineup -- none controlled for. DESCRIPTIVE_ONLY, not a predictive or advantage claim.
Read analysis
Form curve mover -- top_movers_risers #4
delta=+10.815
n_qual_games=146
NOT a live hot/cold streak or predictive signal (net-change over the whole span); NOT BPM/EPM/RAPM/DARKO (box-composite only, same limits as box_value_index.py); NOT age/opponent/role adjusted (raw box rates only); NOT player_id-deduped (accented names split across ids upstream -- carried through)
Read analysis