APPLIED QUANTS — BACKTEST RESULT REVIEW

Experiment name / saved report / date:
Source: synthetic practice, historical simulation, or imported prior results?
Exact configuration, engine version and data provenance:
Requested dates versus dates actually represented:
Symbols requested versus symbols with usable data:
Records evaluated versus trades entered:
Filtered, no-trade, failed and missing records:

TRACE A TRADE
Signal and information available at that time:
Entry timestamp and modeled fill:
Quantity and sizing constraint:
Exit timestamp, price, rule and rule priority:
Gross result:
Commission / spread / slippage / borrow:
Net result:
Does the ledger reconcile to the account result?

INTERPRET
Return and its denominator:
Drawdown definition and account valuation frequency:
Number and distribution of trades:
Dependence on particular symbols or periods:
Benchmark and differences in its cost treatment:
Which fields were available at entry, and which describe later outcomes?
How many variations were tried on this same history?
What is still not modeled?

NEXT STEP
One conclusion supported by the evidence:
One conclusion the evidence does not support:
One follow-up experiment and its frozen assumptions:

Historical simulations do not guarantee future results.