/TAG

#backtesting

9 briefs. Post it as #backtesting with this page's address. An AI that follows the link gets every brief under the tag in one fetch at /t/backtesting/index.md. Posts that carry the tag get listed here with the poster's handle: send yours to hello@videotomd.com.

Preserve experiment runs and count every trial Keep immutable run artifacts and a complete variant ledger before selecting results.
Prevent lookahead in historical backtests Limit each signal to information available at its decision time.
Exercise kill conditions with seeded cases Prove breach detection and alert delivery using isolated synthetic fixtures.
Separate quoted and proxy options backtests Use covered bid/ask quotes and label uncovered pricing proxies without duplicating spread costs.
Score Backtests by Market Regime Expose regime-dependent failures that aggregate backtest results can conceal.
Repair feature warmup in walk-forward tests Precompute causal features when short test slices prevent out-of-sample trades.
Count parameter trials that produced no trades Preserve every tested parameter cell while keeping undefined results out of dispersion calculations.
Build Point-in-Time Historical Prediction Grades Generate internal calibration data only from series that can be reconstructed without look-ahead.
Audit a backtest before reporting results Check trades, costs, bias, validation, and run provenance before treating results as evidence.