Resources › Guides
GUIDES

Recording and backtesting: why to backtest on your own recording, not resampled bars

By TraderWe· Published Jul 8, 2026· Updated Aug 17, 2026· 💬 2
TraderWe records the market you actually saw, then backtests against that recording. Why it matters - Resampled bars hide the microstructure your strategy trades on. A 1-second recording keeps the order book ladder, trade flow and your universe rotation — so a backtest fill is the same fill you would have gotten live. How it works 1. Connect a broker — every subscribed symbol is recorded automatically (1s and 1m). 2. In Backtest, pick recorded days, choose your buy/sell strategies, press Run. 3. "Prepare engine" pre-converts recordings to a fast binary format, later runs are much faster. Reading results - The report includes an overfitting check: bootstrap p-value, confidence interval and Monte Carlo drawdown. A pretty equity curve with a weak p-value is a warning, not a result.

Related reading

Brokers & APIsrecording-based backtests: how many days before i trust it?Guides & TutorialsHow to record your own crypto market data for backtesting, and what to check firstGuides & TutorialsHow to backtest a day trading strategy on US stocks, with your own recorded dataGuides & TutorialsHow to backtest a crypto trading strategy: a complete walkthrough with free sample data
← All guides

Originally published by TraderWe on July 8, 2026. You may quote and link to this page. Republishing the full text without a link back to the original is not permitted.

Related reading

GUIDESPremarket trading explained: can you actually trade it?GUIDESWhat is SSR in stocks? Short sale restriction explainedGUIDESGap trading strategy: what 144,286 gaps say about fillsRESEARCHIndex up while most stocks fall: 24 of 180 sessionsSTRATEGY LIBRARYComposite indicators: building one condition from several partsMARKET MOVERSHarmony Jumps 130% as Layer 1 Blockchains Lead Today's Crypto Movers · Sep 19, 2026

2 replies

I
IndicatorSkeptic· Jul 2026 ago
Love that the report basically tells you when your beautiful equity curve is noise — though I'd argue most people will look at the curve, ignore the p-value, and call it edge anyway. My question: if the recording only exists for symbols I was subscribed to, doesn't my backtest universe quietly become 'stuff I already liked'? That's its own bias, just a friendlier one than resampled bars. Still, recording the actual ladder instead of pretending a bar close is a fill is the honest version. I'll stop pretending my fills were free.
ZenTrader_Ana· Aug 2026 ago· edited Aug 2026 ago
The overfitting check is the part I'd sit with longest. There's a very particular sting in seeing a curve you were already emotionally attached to come back with a weak p-value, I've caught myself re-running with slightly different dates just to make it look better, which tells you everything about who's actually driving. How do you all handle that moment when the numbers say your idea isn't real yet?
Sign in to reply →