UI Workflow
Steps
- Load data in the Data Loader page.
- Create a strategy in the Strategy Editor, or skip strategy sizing entirely and use target-weight portfolio construction in the Backtest Runner.
- Run backtest in the Backtest Runner.
- Review results in Results Dashboard and Portfolio Analyzer.
- Deep-dive analytics in the Advanced Analytics page.
Advanced Analytics (New)
The 🔬 Advanced Analytics page provides:
Heatmaps
- Monthly returns heatmap — calendar grid (month × year) colour-coded by return.
- Correlation matrix — pairwise return correlations for multi-symbol backtests.
- Drawdown heatmap — worst monthly drawdown in calendar form.
Rolling Statistics
- Rolling Sharpe ratio with configurable window (30/60/90/252 days).
- Annualised rolling volatility with percentile bands.
- Rolling beta against an actual benchmark series loaded into the run.
- Rolling maximum drawdown (trailing window).
Compare Runs
Save multiple backtest results and compare them side-by-side with an equity-curve overlay and metrics table.
Parameter Sensitivity
Sweep two parameters across a grid and visualise the impact on return, Sharpe, drawdown, or win rate as a surface heatmap.
Export
- Download equity curve and trades as CSV.
- Download summary metrics as JSON.
- Download an institutional-style tearsheet in JSON or Markdown.
- Use browser print (Ctrl+P) for a quick PDF snapshot.
Dark Mode
Toggle 🌙 Dark Mode in the sidebar for a dark colour scheme.
Portfolio Construction Workflow
The Backtest Runner now supports a target-weight portfolio construction mode for multi-symbol research:
- Enter target weights as
SYMBOL=percentpairs, for exampleAAPL=60, MSFT=40. - Choose a rebalance schedule:
daily,weekly, ormonthly. - Add guardrails for drift, max position weight, max turnover per rebalance, cash floor, and a drawdown-triggered de-risking threshold.
- Review realized weights, drift, turnover, cash buffer, rebalance reasons, and constraint hits in the Portfolio Analyzer tab.
- Benchmark-relative analytics remain available when the benchmark symbol is present in the loaded market data.
Tips
- Start with sample data to validate logic quickly.
- Save configurations for reproducibility.
- Include benchmark bars in the loaded dataset to unlock real beta/alpha/information-ratio analytics.
- Use portfolio construction mode when you want reusable allocation policy + rebalance diagnostics instead of ad hoc strategy sizing.
- Export results or tearsheets for offline analysis.
- Use the Compare Runs tab to evaluate strategies against each other.