Cloud mode: backtests & sweeps run locally only Single Run and Mega-Sweep need local Parquet data + megasweep state (11 GB, not in Supabase). View the sweep archive → Sweep explorer in the map.
Reading the result? Every sweep is measured against the 7-rung evaluation ladder - PBO < 0.2 (honest selection) and a clean Monte-Carlo risk profile decide whether a candidate may go live. Technical terms are hoverable everywhere here.
Evidence ladder →
Loading sweep status…
Loading…