Backtest

Forward-quarter excess return (vs SPY) of the top 50 tickers by buysell_score, for three fund-weighting arms over the SAME fund population each period · QF_SKILL_WEIGHT=0.0 · Last refreshed never -- use the Dashboard's Backtest Refresh button

What this answers

qf is today's production fund-quality weighting. equal weights every in-universe fund 1.0 -- the null hypothesis: does Q_f-weighting the flow signal add anything over just averaging every in-universe manager equally? qf+skill is qf's formula plus a skill-hit-rate term, gated off by default (QF_SKILL_WEIGHT=0, in which case it's identical to qf by construction -- set that env var and re-run Backtest Refresh to actually test a nonzero value). Report this plainly: if equal beats qf, that is the finding, not a bug to chase, and QF_SKILL_WEIGHT should stay at 0 until an arm clearly and repeatably wins.

Pooled across all 0 periods

No backtest results yet -- click "Refresh Backtest" on the Dashboard.
Long US-equity 13F disclosures only · up to 45-day reporting lag · sells = reduce/avoid, not short. JSON: /api/signals · /api/funds · /api/status