Writing
Engineering and research notes on how Rearview is verified, how backtest engines differ under the hood, and how to separate real edges from selection bias.
Articles
- Checking a portfolio backtester against a clean-room second engine, vectorbt, bt and LEAN: How Rearview is tested before every release with hand-worked cases, differential testing against a separate Python engine, external anchors in three open-source backtesters and cut-date look-ahead replays, plus the design differences we documented across engines.
- Why AI coding agents p-hack backtests, and how a chained tries log catches it: What happens to the Deflated Sharpe Ratio when an agent runs 300 variations in an afternoon, why both DSR and PBO matter, and how a token-hashed, SHA-256-chained tries log tracks every try across an app and CLI.
- Five silent reasons the same trading strategy gives different results across backtesters: Signal-bar fills, pre-adjusted prices, survivorship and delisting haircuts, cash shortfalls, and tie-breaking and calendar conventions, plus a five-step protocol for reconciling two backtesters down to the first diverging fill.