Оптимизация параметров торговой стратегии на VectorBT

★ 7.1 · general

optimize is a Claude Code skill that generates a parameter optimization script for VectorBT-based trading strategies, iterating over all parameter combinations and recording total return, Sharpe ratio, max drawdown, and trade count for each. The output script is placed in `backtesting/{strategy_name}/`, fetches market data via OpenAlgo `client.history()` or directly from DuckDB, and relies exclusively on OpenAlgo ta for indicators — including Supertrend and Donchian — cleaning signals with `ta.exrem()`. Progress is shown via tqdm; results include top-10 combos ranked by both return and Sharpe, two dark-themed Plotly heatmaps, a printed comparison against the NIFTY benchmark, and a CSV export. Designed for optimizing ema-crossover, rsi, donchian, and supertrend strategies on NSE/NFO instruments, with Indian delivery fee defaults and futures lot-size awareness for NIFTY and BANKNIFTY.