the measure
A sharpe ratio is a return divided by how much that return jumped around, so a steady 10% scores higher than a wild 10%. The convention is that above 1 is good. We have 2,067 of them, measured the same way, and only 9.7% clear 1.
the short answer
Above 1 is the number the industry uses, and it is a reasonable one. What nobody tells you is how uncommon it is outside a brochure. We ran 15 trading rules across 144 markets and kept every result, including the bad ones. That gives 2,067 sharpe ratios measured identically. The median is -0.24. 41.8% are above zero at all. 9.7% clear 1. 1.0% clear 2.
So a strategy showing 1.2 is in the top tenth of everything we have measured, and one showing 3 is at the edge of the whole distribution. Our highest is 3.21 and our lowest is -4.21.
| where it sits | sharpe ratio | what that means |
|---|---|---|
| bottom quarter | -0.90 | losing, and volatile about it |
| the middle | -0.24 | the typical result here |
| top quarter | 0.42 | positive but modest |
| top tenth | 0.99 | just short of the conventional bar |
| top hundredth | 2.00 | rare, and worth a second look for mistakes |
how it works
Take the average daily return of a strategy. Divide it by the standard deviation of those daily returns, which is a measure of how much they scatter. Multiply by the square root of the number of periods in a year to put it in annual terms. That is the whole calculation.
The denominator is the interesting part. It treats an upward surprise and a downward one as equally bad, which is why a strategy that makes its money in a few enormous days scores worse than one that grinds. Whether that is the right way round depends on what you are trying to survive. If the thing that would make you quit is a sickening fall, then drawdown is the more honest measure, and we publish both for every market.
the catch
Run enough rules on enough markets and something will score well on any of them. That is not a claim, it is arithmetic, and our own table is the demonstration: 15 rules on 144 markets gives 2,160 results, so a one-in-a-hundred sharpe ratio should show up about twenty times by chance alone, and 1.0% of ours are above 2.
The defence is to decide the rule first and test it on data it has never seen. We cut every market's history at 70% and judge on the rest. When a strategy is sold to you on a backtest, the question worth asking is not what its sharpe ratio was. It is which half of the data the rule was chosen on, and what happens when you pick the winner.
questions
Above 1 is the usual convention. In 2,067 strategy and market tests we ran, 9.7% cleared 1 and 1.0% cleared 2, with a median of -0.24. Treat anything above 2 on a single market and a single window as more likely to be luck or an error than an edge.
It means the strategy lost money over the period and was volatile while doing it. 58.2% of our out of sample results were negative.
Average return divided by the standard deviation of those returns, multiplied by the square root of the number of periods in a year. We compute it on daily returns and annualise by the square root of 365.
Sharpe measures return against how much the returns scattered, treating upside and downside scatter the same. Drawdown measures only the worst fall from a peak, which is closer to what makes people abandon a strategy.
the evidence
144 markets, 15 rules, 2,160 tests, every one published with both halves and the dates. Read them at the market pages. Figures on this page were computed 2026-11-17.
more