Changelog
Guides are signed by their author and end with a list of sources.
Slippage is now charged on every fill, so a stop can fill worse than its level. The engine and its independent Python reference agree trade for trade on the new path.
Results now say whether they can be trusted: sample size, consistency across periods, drawdown, liquidations and losing streaks are judged, not just reported.
Sharpe ratio and longest losing streak added. Every page got a share card built from real backtest numbers.
Strategy gallery: eight well-known strategies run on six coins, with the losing results published too. Strategies are now explained in plain language above the results.
Visual overhaul to a single documented design standard, and the exit panel became an add-what-you-need list.
Eight timeframes (5m to 1w). The page now runs a backtest on arrival instead of waiting for a click.
Site launched: browser-side backtesting on seven years of Binance candles, no account.
Candle data refreshes every week by a scheduled job, and every run also pulls the latest closed candles directly from Binance, so results always include recent price action.