Expected Returns

Universe ranked by Itô-corrected log-wealth growth g = E[r] − ½σ² for the active horizon. Calibrated P(up) · Wilson 95% CI when cohort N≥30 · cohort Brier from forecast ledger.

Top N

Methodology. Rank by g = E[r] − ½σ² (Itô-corrected log-wealth growth — D32 net-outcome mandate; Kelly 1956 / Itô 1944 / Cover-Thomas 2006). P(up) calibrated via cal-applier-2026-04-28-v1 when cohort N≥30; Wilson 95% CI (Brown-Cai-DasGupta 2001). Cohort Brier from forecast-ledger-supabase. Empty-state shows "Calibrating (N/30)" — never fabricates probUp or Brier. Active methodology hash: