Two systems with the same return are not equivalent
Two systems both returned 20% in a year: one on a calm path, the other with violent swings and repeated drawdowns. Return says they are equal, which is incomplete — the second demands more psychological endurance and carries a higher chance of being abandoned before the return arrives.
The Sharpe ratio separates them by dividing return by standard deviation: the higher the ratio, the calmer the return. It has an important limit, though: it penalises volatility upward and downward alike, so a system that occasionally jumps higher scores worse despite that not being harm — which is why it is read alongside drawdown rather than instead of it.
One record, several views
The reference record shared with the other performance pages: 100 trades, 40 winners averaging +2R and 60 losers averaging −1R, for an expectancy of +0.20R per trade.
Each measure here reads that same record from a different angle: one weighs return against variability, the other shows the path that produced the result. One outcome, several views — which is why they are read together rather than as substitutes.
The formula
Sharpe = (mean return − treasury-bill rate) ÷ standard deviation of returns
A worked example
Two systems returning 20% a year against a 4% treasury-bill rate:
- The first with 8% standard deviation → (20 − 4) ÷ 8 = 2.0
- The second with 32% → (20 − 4) ÷ 32 = 0.5
Identical return, entirely different experience. The second figure means four times the variability for the same result — and a far higher chance of being abandoned midway. See the equity curve to see the difference visually.
Common mistakes with this term
- Comparing Sharpe ratios computed over different periods or frequencies as if equivalent.
- Reading it without maximum drawdown, since it penalises upside variability as much as downside.