How Fidelity Lab tests a strategy: PIT-faithful walk-forward, explained

Written by the TradeHouse team.

Fidelity Lab runs every strategy through a point-in-time (PIT) faithful walk-forward test — meaning the backtest only ever sees data that would genuinely have been available at each historical moment, never data from the future. This is the same discipline TradeHouse applies to its own research: a backtest that leaks future information produces a fake edge.

On top of the walk-forward test, Fidelity Lab runs 11 contamination detectors that catch the most common ways a backtest silently cheats — look-ahead bias, survivorship bias, and data leaks among them — plus per-market rulers, because a rule that's fair for one market family (say, binary prediction markets) can be unfair for another (continuous-price equities).

The output is a single cost-netted verdict: does the strategy clear real trading costs (fees, spread, slippage) after all of that, or not. Fidelity Lab launched September 2026 on Free / Starter $49 / Pro $199 / Enterprise $499 per month tiers.

Try Fidelity Lab → · See TradeHouse pricing →

Data and research only. Not investment advice, not a recommendation to buy or sell, and not an offer of any security, swap or event contract. TradeHouse is not a registered investment adviser, broker-dealer or exchange. Any backtested, hypothetical or historical results are shown for illustration, have limitations (including hindsight and no trading costs or slippage unless stated), and do not guarantee future results. Prediction markets and options carry risk of loss.

DMCA / Copyright