Calculator
Deflated Sharpe Ratio Calculator
Bailey & López de Prado deflated Sharpe — corrects observed Sharpe for multiple-testing across K trials. Returns deflated SR, PSR, and confidence interval.
- Inputs
- Form inputs / CSV
- Runtime
- Instant
- Privacy
- Client-side · no upload
- API key
- Not required
- Methodology
- Open →
Inputs
Notes
K trials = how many strategies you tried before picking this one. 100 is typical for grid search. Higher K demands a higher observed Sharpe to clear the multiple-testing haircut.
Deflated Sharpe
-38.17
Observed 2.00 minus expected-maximum-under-null 40.17. Positive after K trials means the edge survives multiple-testing.
Decision metrics
PSR (skill prob)
0.0%
P(true SR > deflated benchmark)
z-statistic
-36.93
non-normal corrected
Effective benchmark
40.17
Sharpe to beat
Reading the result
PSR > 95%: the edge clears the multiple-testing haircut — the strategy is statistically distinguishable from a lucky pick from K trials. 80–95%: marginal. Consider longer history or fewer trials. <80%: the apparent edge is consistent with noise after K-trial selection bias. See methodology.
How to use
Step-by-step
- 1
Enter your strategy's raw Sharpe ratio, sample size (number of return observations), skewness, and excess kurtosis.
- 2
Enter the total number of strategy variants tested — including the ones you discarded. This is critical: under-reporting trials inflates DSR.
- 3
Read DSR alongside raw Sharpe. The gap shows how much the multiple-testing penalty discounts your raw number.
- 4
Compare DSR to thresholds: ≥0 weak, ≥1 moderate, ≥1.65 95% confidence, ≥2 actionable for live capital.
- 5
If DSR is negative, the deflation correction overwhelms the raw Sharpe — test a longer sample, fewer variants, or a different signal.
Glossary references
Terms used by this tool
Questions people ask next
FAQ
What does 'deflated' mean here?
Standard Sharpe ratio assumes returns are i.i.d. Gaussian, which they aren't. Lopez de Prado's deflated Sharpe (2014) corrects for skew, kurtosis, and the multiple-testing penalty when you ran N strategy variants. The deflated value is what would be left after the curve-fitting comes out in the wash.
How many trials should I report?
All of them. The deflation factor depends on the total number of strategy variants tested, including the ones you discarded. Under-reporting trials shrinks the deflation and inflates the apparent Sharpe — that's exactly the bias the metric was designed to catch.
What's a meaningful deflated Sharpe threshold?
DSR ≥ 0 corresponds to ~50% confidence the true Sharpe is positive. DSR ≥ 1 is roughly 84%, ≥ 1.65 is 95%. Most academic papers use 1.65 (95%) as the publication threshold. For live capital allocation, requiring DSR ≥ 2 is reasonable when the trial count is high.
Can DSR be negative?
Yes. A negative DSR means the deflation correction is so large that the strategy's Sharpe is statistically indistinguishable from zero, or worse, from a draw of pure random noise. Negative DSR is common when the trial count is high (50+) and the raw Sharpe is modest.
Should I trust DSR if my trade log is short?
Be cautious below 100 observations. The deflation formula incorporates higher moments (skew, kurtosis) which need decent sample size to estimate. Below 50 observations, treat DSR as directional only. The tool reports the underlying moment estimates so you can sanity-check.
Related deep dive
All articles →Read further
Long-form context behind the tool output.
- Tutorial · Runnable·12 min
Did You Overfit? PBO and Deflated Sharpe
A practical tutorial on the two best-documented tests for backtest overfitting — PBO via CSCV and the Deflated Sharpe Ratio. Runnable Python + tool.
Read - Methodology · Opinion·8 min
The Sharpe Ratio Trap
Sharpe ignores tail risk, assumes Gaussian returns, and is trivially gameable. Four metrics to report alongside it: Sortino, Calmar, tail, deflated Sharpe.
Read - Comparison · Benchmark·12 min
Model Selection Framework for Finance Tasks
A task × latency × cost × context decision tree for finance LLM workloads. Ten concrete scenarios mapped to tier bands. Grounded in published pricing, not.
Read
Complementary tools
Users of this tool often explore
Backtest Overfitting Score
Upload a backtest trade log and compute Probability of Backtest Overfitting (PBO), Deflated Sharpe Ratio, and the odds your edge survives live trading.
Walk-Forward Validation Visualizer
Paste a strategy returns CSV, get per-window in-sample vs out-of-sample Sharpe and the IS→OOS drop. Rolling and anchored window modes. Browser-only.
Risk-Adjusted Returns Calculator
Paste a returns CSV. Sharpe, Sortino, Calmar, Omega, alpha, beta, tracking error, information ratio, max drawdown, and tail moments — plus.