stochastly
Start free trial

Minimum track record length calculator

Estimate the observations needed for a Sharpe ratio to exceed a chosen benchmark at a stated confidence level. Inspect how the estimate changes with the Sharpe ratio and return kurtosis.

Calculator

The synthetic starting case uses monthly observations, an annual Sharpe ratio of 2, and an annual benchmark of 1. Skewness is -0.72 and raw kurtosis is 5.78.

Sharpe and observation frequency

An observation is one return interval. Calendar years here equal observations divided by this frequency.

Enter at least 2 to see PSR at the current track record length.

Return moments and confidence

Normal returns have raw kurtosis 3. Add 3 to excess kurtosis before entering it.

Sensitivity ranges
Open or copy scenario URL

Results

MinTRL, observations
59.90
Observations required, rounded up
60
Indicative calendar years
4.99
PSR(SR*) at current T
0.950151

At this Sharpe, 60 observations are required for PSR(SR*) >= 0.95.

Sensitivity to annual Sharpe

The line reports calendar years at each annual Sharpe value. A missing point means the observed Sharpe is at or below SR*.

Minimum track record length by annual Sharpe ratioThe curve loads when JavaScript runs. The table below contains equivalent values.
Annual Sharpe sensitivity data
Annual SharpeObservationsYears

Sensitivity to kurtosis

The observed annual Sharpe stays fixed while raw kurtosis varies. This isolates the contribution of the tail term in the estimator variance.

Minimum track record length by raw kurtosisThe curve loads when JavaScript runs. The table below contains equivalent values.
Raw kurtosis sensitivity data
Raw kurtosisObservationsYears

Method and assumptions

For each observation interval, SR is the mean excess return divided by its standard deviation. An annual Sharpe ratio is divided by the square root of observations per year before evaluating the formula. This conversion assumes independent return intervals.

MinTRL = 1 + (1 - skewness × SR + (kurtosis - 1) × SR² / 4) × [Φ⁻¹(confidence) / (SR - SR*)]²

This is Eq. 13 of Bailey and López de Prado (2012). The result is conditional on the observed Sharpe and entered moments. It does not predict future returns. Observations, trades and independent windows can have different counts; the formula treats the entered observations as independent.

At an annual Sharpe of 2 against 1, with monthly returns, skewness -0.72 and raw kurtosis 5.78, Appendix A.3 reports 59.895 observations, about 4.99 years. The calculator displays the unrounded estimate and the integer count needed to reach it.

The calculation runs in the browser. A scenario URL contains entered scalar values when you choose to share it. The form has no field for a return series.

Bailey and López de Prado, The Sharpe Ratio Efficient Frontier, Appendix A.3, p. 20; read 2026-10-03. Deflated Sharpe ratio calculator applies PSR to a multiple testing benchmark. Sharpe ratio and selection bias provides related context. All tools.