Adjust an observed Sharpe for the number of strategy variants tested, sample size, skew, and kurtosis of returns. Output is the probability the true Sharpe is positive after correcting for trial-selection bias and non-normality. Implements Bailey & López de Prado (2014).
The best Sharpe you reported across all variants, as the tearsheet states it. Converted to per-bar (÷ √bars per year) before the test.
How many strategy variants you tested on the same data. Honest counts beat optimistic ones.
Bars the Sharpe was computed from, at the same frequency as the factor below. e.g. 365 for a year of daily bars, 52,560 for 18 months of 15-minute bars.
Negative skew (crash risk) penalises Sharpe; 0 = symmetric.
Raw, not excess: 3 = Gaussian (excess kurtosis + 3). Crypto strategies typically 5–10.
The frequency of the bars behind T. The annualized Sharpe is divided by its square root so SR and T sit at one frequency; it does not cancel.
Across 20 zero-edge trials on 365 bars, the best Sharpe luck alone produces (0.0996 per bar).
Annualized Sharpe ÷ √(bars per year), the scale the test runs at.
1 − skew·SR + (kurt − 1)/4 · SR² at per-bar SR. Inflated by negative skew or fat tails.
DSR between 0.05 and 0.5: the observed Sharpe is roughly what selection alone would produce. Either the strategy has no real edge, the trial count is too high for the sample, or both. Cut the parameter search or extend the sample before deploying.
Holding SR, T, skew, and kurt constant, watch DSR collapse as the trial count climbs. The dot marks your current N.
The deflated Sharpe (DSR), on a scale of 0 to 1, over 775 days. Compared with the best of 385 versions by luck: 0.74. With one version it would be 0.96. No range shown: the method serves none for this number.
| Check | Measured | How |
|---|---|---|
| Sharpe discounted for versions tried | 0.74 | Discounts the Sharpe for the number of versions of the strategy's family Keel has seen tried. |
| Versions tried | 385 | 385 versions tried since Nov 16, 2026, about 4 separate; versions tried outside Keel: unknown. |
Keel's study of Momentum Strategy on Hyperliquid · computed Nov 16, 2026 · data Aug 15, 2024 to Sep 29, 2026
The Deflated Sharpe Ratio (DSR) is a probability: specifically, the probability that the true Sharpe of a strategy is greater than zero, conditional on the observed Sharpe, the number of variants tested, the sample size, and the higher moments of the return distribution. Formally:
Z = (SR − SR₀) · √(T − 1) / √(1 − skew·SR + (kurt − 1)/4 · SR²)
DSR = Φ(Z)Where SR₀ is the expected maximum Sharpe across N independent random strategies under the null of zero true edge:
SR₀ = √(2 ln N) − (γ + ln ln N) / √(2 ln N)
where γ ≈ 0.5772 (Euler–Mascheroni)What the deflator actually does. Subtracting SR₀ penalises the observed Sharpe for the selection bias introduced by running N trials. The variance term in the denominator inflates noise when returns have negative skew or fat tails, both of which make a given Sharpe less trustworthy. The √(T − 1) scaling tightens the estimate with longer samples; with a year of daily data you can distinguish less than you might think.
When N = 1 the deflator SR₀ collapses to zero and DSR reduces to the standard probabilistic Sharpe ratio (PSR): the probability the true Sharpe exceeds zero given a single, pre-specified strategy. For N > 1 the deflator does the work the t-test cannot.
The widget runs the full computation in your browser; nothing is uploaded. For the background on trial bias, why crypto Sharpes are particularly vulnerable to it, and how to count trials honestly, see the DSR explainer.
Keel is a strategy builder for Hyperliquid. Describe an idea, backtest it with fees, slippage and funding, and run it live on Hyperliquid’s crypto perps. Start free, and connect a wallet when you’re ready to trade.
First work out SR₀, the best Sharpe ratio you would expect from your number of variants if none had an edge. Then ask how likely your Sharpe is to beat it, given the number of bars and the skew and kurtosis of your returns: DSR = Φ((SR − SR₀) × √(T − 1) ÷ √(1 − skew × SR + (kurtosis − 1) ÷ 4 × SR²)), with both Sharpe ratios per bar. For an annualized Sharpe of 1.5 over 1,460 daily bars, the best of 30 variants, with skew -0.5 and kurtosis 5, SR₀ is 1.04 annualized and the deflated Sharpe ratio is 81.7%.
The deflated Sharpe ratio is a probability: the chance that the strategy's true Sharpe ratio is above zero after the variants you tried. The usual bar is 95.0%, the common confidence level. Near one half, your Sharpe ratio is about what picking the best of that many variants with no edge would give, which is no evidence of an edge at all.
It deflates the observed Sharpe ratio for four things at once: (1) the number of strategy variants you tested (more trials, higher expected max Sharpe under the null, even with zero edge); (2) sample size T: short samples produce noisy Sharpes; (3) skewness: negative skew (rare crashes) makes a given Sharpe less reliable; (4) excess kurtosis: fat tails make a given Sharpe less reliable. The output is the probability that the true Sharpe is greater than zero given everything you tried and the shape of your returns.
Because the expected max Sharpe across N independent random strategies grows with sqrt(2 ln N), even when no strategy has any real edge. If you tested 100 parameter variants and reported the best Sharpe, your number is contaminated by selection bias. DSR corrects for that by comparing your observed Sharpe to SR_0, the expected best Sharpe under the null of zero true edge. If you only ran one strategy, set N=1 and the trial-bias correction drops out (SR_0 = 0).
Count every parameter combination you evaluated on the same data, including informal ones. If you swept lookback ∈ {10, 20, 50, 100} × threshold ∈ {0.01, 0.02, 0.05} × asset ∈ {BTC, ETH, SOL}, that is 36 trials, not one. Add manual variants ("I also tried with a 7-day filter"); those count too. If anything, undercounting is the bigger risk; most practitioners underestimate N by 5-10x. When unsure, sensitivity-test by entering both your best-guess N and 10×N.
A regular Sharpe significance test asks "given a single sample, is the Sharpe distinguishable from zero?" DSR asks "given that I selected this Sharpe from N alternatives, is it distinguishable from what selection alone would have produced?" The two converge at N=1 (no selection); they diverge sharply as N grows. A Sharpe of 1.5 over four years of daily returns is about 3 sigma by t-test (the t-statistic is roughly the annual Sharpe times the square root of the years), but can be statistically noise once you account for testing 30 variants.
In Keel's Research, yes: its Check a strategy step "Sharpe discounted for versions tried" is a deflated Sharpe, counted over the versions of the strategy's family Keel has seen tried, as in Keel's study on this page. A single Keel backtest reports observed Sharpe, Sortino, max drawdown and other single-run metrics. To deflate a Sharpe from any backtest, enter it here with an honest count of the variants you tried, and this calculator gives the deflated probability.
Why selection bias inflates observed Sharpes, how the DSR deflator works, and what counts as a trial. Pairs directly with this widget.
Probability of backtest overfitting via combinatorial split. Sibling diagnostic: DSR penalises trial count, PBO measures how often best-in-sample loses out-of-sample.
Block-bootstrap CIs on Sharpe and max drawdown. Complementary rigor: DSR adjusts for selection; Monte Carlo adjusts for path luck.
DSR is a strong diagnostic, but it has limits. It does not catch look-ahead bias (using data unavailable at signal time), structural breaks (the regime that produced your Sharpe is gone), or non-stationarity (the return distribution itself shifted during the sample). It assumes your trial count is honest and that the N strategies were drawn from a single, exchangeable population. Treat a high DSR as necessary, not sufficient. Pair it with bootstrap CIs on the same returns and an out-of-sample period the model never saw.