How is the deflated Sharpe different from the probabilistic Sharpe?

PSR tests if SR exceeds a fixed benchmark SR*. DSR sets the benchmark to the expected maximum Sharpe from testing N strategies: SR* = E[max(SR_1,...,SR_N)]. This accounts for the data mining problem.

What is the DSR benchmark for testing 100 strategies?

For 100 independent strategies with T=500 observations each: SR* ≈ sqrt(2*log(100)/(500-1)) * ((1-0.577)/2+1) ≈ 0.45. A strategy needs estimated SR > 0.45 just to have DSR > 0.5.

Deflated Sharpe Ratio Correction Calculator

Compute the deflated Sharpe ratio by adjusting for selection bias from testing multiple strategies. DSR = PSR(SR* = E[max Sharpe of N trials]), accounting for skewness, kurtosis, and data snooping.

Use the full interactive calculator below, or

Related Calculators

Probabilistic Sharpe

Finance Ultra-depth

Backtest Sharpe Bias

Finance Ultra-depth

Factor Zoo T-Stat

Finance Ultra-depth

Deflated Sharpe Ratio Correction Calculator

Related Calculators

Frequently Asked Questions

How is the deflated Sharpe different from the probabilistic Sharpe?

What is the DSR benchmark for testing 100 strategies?