Backtest validation with statistics that don't flatter you.
Every check here exists because it's the one people skip when the backtest is finally winning.
Corrects your Sharpe ratio for every variant you tested, not just the one that happened to win.
Monte Carlo resampling builds the real distribution of luck for your setup - not a rule of thumb.
Every recorded trial stays on the record - no deleting the losers after the fact.
Commission and slippage are charged on every simulated fill, every time.
Describe entries, exits, and sizing with a picker - no notebook, no glue code.
Daily bars for crypto and equities - not synthetic or look-ahead-biased series.
Multi-sleeve portfolios with rebalancing and honest trial accounting across weight variants - free with an account, no card needed.
Genetic programming searches a language of strategies, not a menu of named indicators - thousands of real trials, finalized on data the search never touched.
Curious what overfitting actually looks like? The Overfitting Machine runs this exact multiple-testing problem 600+ times against real data - free, no signup.
Trusted at scale:
Get early access
Deflated is in private development - leave your email for early access.
Results are recorded to a shared demo ledger, not a private account.