Returns you can audit.
Every number on this page comes from the public Phase 37 walk-forward harness (shipped May 17, 2026) — sleeve-class quota top-5 over 78 rolling windows (180d train / 30d test, 8-year history, 2022 bear excluded). The same engine runs your live capital. No black boxes, no PDF tear-sheets you can't reproduce.
Pick any starting day on the curve.
Drag the slider to anchor the equity curve to a specific day in the past. Return and max drawdown recompute from that anchor.
How the curve breathes.
Returns alone don't tell you whether you can sleep through a drawdown. These two charts do.
How we validate every strategy.
Most "trading bots" show you the curve and ask you to trust them. We hand you the harness and ask you to break it.
Rolling 180-day train / 30-day test windows. Catalog is re-ranked by train-window Sharpe each step, then OOS returns are concatenated.
Strategy catalog is fixed at the start of the test. The bottom 5 sleeves by OOS Sharpe are dropped — hedge sleeves are exempt.
All trades modeled with taker fees and conservative slippage. No look-ahead, no infinite liquidity assumptions.
The exact code path that runs the backtest runs live execution on Binance, Bybit, Coinbase, Kraken, OKX, Gate.io and MEXC.
When vol-target or DD-brake de-risks the book below 100%, the idle fraction earns a conservative 5% APY from non-custodial stablecoin lending (Aave / sDAI / Ondo). Added on top of trading PnL — never replaces it. Full breakdown on the methodology page.
Crypto trading carries substantial risk of loss. Backtest results, including walk-forward out-of-sample results, are not a guarantee of future performance. Live results may differ materially due to execution slippage, exchange downtime, regime shifts and other factors. Only trade capital you can afford to lose.