backtesting-frameworks
Construa sistemas robustos de backtesting para estratégias de negociação
也可从以下获取: wshobson
O desenvolvimento de estratégias de negociação exige backtesting rigoroso para evitar vieses dispendiosos. Esta skill fornece padrões de nível de produção para validação confiável de estratégias com tratamento adequado de look-ahead bias, survivorship bias e custos de transação.
下载技能 ZIP
在 Claude 中上传
前往 设置 → 功能 → 技能 → 上传技能
开启并开始使用
测试它
正在使用“backtesting-frameworks”。 Basic momentum strategy backtest with 20-day lookback
预期结果:
Strategy achieved 12.3% annual return with 15.2% volatility. Sharpe ratio: 0.81. Maximum drawdown: -18.4%. Win rate: 54.2%. Results include equity curve, drawdown chart, and monthly return heatmap.
正在使用“backtesting-frameworks”。 Walk-forward optimization for mean reversion parameters
预期结果:
Optimal lookback window: 14-21 days across 8 test periods. Out-of-sample Sharpe ratio: 0.65 (in-sample: 0.72). Parameter stability confirms strategy robustness. Combined equity curve shows consistent performance across market regimes.
安全审计
安全All 33 static analysis findings are false positives. The skill contains documentation and Python code examples in markdown format only. No executable code, network calls, or security risks detected. Markdown code block delimiters were incorrectly flagged as shell execution. Type annotations and common financial terms triggered false pattern matches.
质量评分
你能构建什么
Pesquisador quantitativo validando novos sinais de negociação
Um pesquisador quant desenvolvendo uma estratégia de negociação baseada em momentum precisa validar o desempenho em vários regimes de mercado enquanto evita overfitting e garante premissas realistas de custos.
Trader algorítmico construindo estratégias sistemáticas
Um trader algorítmico implementando estratégias sistemáticas requer infraestrutura de backtesting robusta com divisões adequadas de treino/validação/teste e otimização walk-forward para garantir robustez da estratégia.
Cientista de dados explorando aplicações financeiras
Um cientista de dados aplicando machine learning a dados financeiros precisa orientação sobre metodologia adequada de backtesting para evitar armadilhas comuns como look-ahead bias e survivorship bias.
试试这些提示
Help me set up a basic backtesting framework for a simple moving average crossover strategy. I have daily OHLCV data in a pandas DataFrame. Include transaction costs and calculate key performance metrics.
I need to implement walk-forward analysis for my mean reversion strategy. Use a 252-day training window and 63-day test window with anchored training. Optimize the lookback parameter and show the equity curve from combined test periods.
Run a Monte Carlo simulation on my strategy returns to assess robustness. I want bootstrap analysis of maximum drawdown distribution and probability of loss over 21, 63, and 252 day holding periods with 1000 simulations.
Build an event-driven backtester with custom execution logic for limit orders. Include realistic fill modeling based on order book dynamics, position tracking, and real-time PnL calculation. Support multiple assets with portfolio-level risk management.
最佳实践
- Always use point-in-time data to avoid look-ahead bias - ensure signals are generated using only information available at decision time
- Reserve out-of-sample data for final evaluation - never optimize parameters on your test set
- Include realistic transaction costs - model both explicit costs (commission) and implicit costs (slippage, market impact)
避免
- Using adjusted close prices without understanding the adjustments - can introduce look-ahead bias from future dividend or split information
- Optimizing too many parameters relative to available data - leads to overfitting and poor out-of-sample performance
- Ignoring survivorship bias by testing only on current constituents - must include delisted securities for accurate results