backtesting

Minimum Track Record Length

The minimum number of return observations needed to conclude, with statistical confidence, that a strategy's true Sharpe Ratio is positive.

The Minimum Track Record Length (MinTRL), developed by Bailey and López de Prado (2012), answers: how many months (or years) of performance data are required to conclude that a strategy has genuine positive alpha, rather than being the result of luck?

MinTRL is an increasing function of:

  • The desired confidence level (how certain you want to be)
  • The returns' excess kurtosis (fat tails require more data)
  • The returns' negative skewness (asymmetric distributions require more data)

And a decreasing function of the observed Sharpe Ratio (stronger apparent performance requires less data to confirm).

Practical implications

  • Non-normal returns — fat tails and negative skew, common in strategies that sell optionality — require much longer track records to establish statistical significance than normal-return strategies at the same Sharpe.
  • A two-year track record is rarely sufficient to make a statistically valid claim of alpha in most trading strategy contexts, particularly with monthly data.
  • MinTRL represents the minimum necessary condition for trusting a track record — it is not sufficient on its own; multiple testing adjustments are still needed.

Related terms

Related articles