Check the measurement basis
Record the period, currency, benchmark, cash flows and whether the result is simulated or live. Separate gross returns from returns after costs. Annualized figures need the observation period and calculation convention alongside them.
Read the path as well as the endpoint
Two series can have the same endpoint and different interim losses. For a constructed example, 100 → 80 → 110 and 100 → 105 → 110 both end 10% above the start; their sampled maximum drawdowns are 20% and 0%. These are teaching sequences, not observed strategies.
Interpret risk-adjusted metrics
A Sharpe ratio depends on the return frequency, risk-free convention and variability estimate. Compare only compatible definitions. A single ratio cannot show position concentration, extreme losses or whether the rule was selected after many trials.
Ask what survives realistic costs
Inspect turnover, fees and slippage assumptions. Ask for a cost sensitivity analysis and a held-out period. Record conclusions as supported, conditional or unresolved, with the evidence for each. No score on this page determines whether to invest.
Sources & references
- William F. Sharpe — The Sharpe Ratio (1994) ↗
- Bailey et al. — The Probability of Backtest Overfitting (2015) ↗
References support specific definitions or methodological context. Examples and research checklists are explanatory; no performance result is claimed.