In development
Forex Lab
A backtest engine that reports honestly
A strategy backtesting engine with parameter sweeps — and a results view that tells you when a good-looking number came from too few trades to trust.
The problem
Parameter sweeps always find a winner. Almost always it is overfitting: a great profit factor produced by fifteen trades, which proves nothing.
What we built
Sample size is a first-class output, not a footnote. A combination with a strong profit factor over too few trades is labelled unproven, so it never gets mistaken for a validated edge.
Capabilities
- Parameter sweeps with profit factor and drawdown
- Sample size surfaced next to every result
- Explicit unproven flag for low-trade-count winners
- Out-of-sample split support