Distributional draft model · 2007–2026 · out-of-fold
NBA Draft Board / career projections
Every prospect gets a predicted distribution of realized value, not a single rank — a two-part model: a propensity that he ever becomes an NBA player, and, if he does, his availability × on-court quality over four seasons (a Gaussian copula over the two). Never sticking is a floor below the worst player who plays, so a likely-bust sinks beneath a mediocre-but-real rotation piece. Each row shows the outcome curve, the quality and availability pieces, and — for closed classes — where he actually landed. Open any prospect for his scouting rubric and a career trajectory — rookie, four-year, and peak quality drawn as a banded arc. Out-of-fold, leakage-free.
value distribution (p10–p90) middle 50% (p25–p75) median actual outcome/ offensive / defensive impact (4yr)
Risk aversion · risk-neutral γ=0ranked by the certainty-equivalent CE(γ) of his value to the drafting club — four rookie-deal seasons at market, then what it keeps when it re-signs him; at γ=0 that is E[V] (risk-neutral) — the convex price already rewards upside