Free Tool · No signup

Minimum Backtest Length
Calculator.

Your backtest shows a Sharpe of 1.8. Is it real? Upload your MT4, MT5 or cTrader backtest HTML — we'll parse the trades, compute your actual Sharpe, skew and kurtosis, and use the Bailey & López de Prado (2014) formula to tell you whether your track record is long enough to be statistically distinguishable from luck.

🔒 100% client-side. Your report never leaves the browser.
▲
Drop backtest report here
or click to select · .htm / .html
MT4 Strategy Tester MT5 Strategy Tester cTrader Backtesting
§ 01 — Detected from your report

Backtest statistics.

§ 02 — Verdict

Is your backtest long enough?

▸ Verdict
—
—
Confidence level
Min trades needed
—
MinTRL (observations)
Trades in report
—
closed trades
Min years needed
—
at report's trade frequency
Years in report
—
actual backtest length
This is one of 19 tests.
MinTRL tells you if your track record is long enough. It doesn't tell you whether your backtest has been overfit, whether Monte Carlo simulations survive, or whether the drawdown structure hides tail risk. Edge Matrix runs all 19 institutional-grade tests in one pass.
Try Edge Matrix →

The math (Bailey & López de Prado, 2014)

The Minimum Track Record Length answers a specific question: given the observed Sharpe ratio and the higher-moment properties of the trade returns, how many observations are needed before we can reject the null hypothesis that the true Sharpe is at or below the benchmark SR*?

MinTRL = 1 + ( 1 − γ₃·SR̂ + ( ( γ₄ − 1 ) / 4 )·SR̂² ) · ( zα / ( SR̂ − SR* ) )² SR̂, SR* are per-period Sharpes; zα is the inverse normal CDF at the confidence level.

The key insight is the non-normality penalty. Skewness and kurtosis of your trade returns dramatically change how much data you need. A strategy with heavy left tails (negative skew, high kurtosis) needs far more track record to prove it's real, because a single bad tail event can wipe out apparent edge.

Citation: Bailey, D. H. & López de Prado, M. (2014). "The Deflated Sharpe Ratio: Correcting for Selection Bias, Backtest Overfitting and Non-Normality." Journal of Portfolio Management, 40(5), 94–107.