Two traders finish the year up the same amount. One slept fine every night; the other nearly blew up twice getting there. The scoreboard says they’re equal. They are not — and risk-adjusted metrics exist to prove it.
Why raw return hides how much risk you ran
Total return is the number everyone quotes and the number that tells you the least. It answers “how much did you make?” while ignoring the only question that predicts your future: “how much risk did you run to make it?”
A 40% year built on wild equity swings and a couple of near-breach drawdowns is not obviously better than a 20% year that barely dipped. The 40% path is more likely to have ended in a blow-up under slightly worse luck — and for a prop-firm trader, a path that flirts with the drawdown limit is a path that eventually breaches it. Return tells you what happened; risk-adjusted metrics tell you whether it was repeatable or whether you just survived.
The whole point is to divide reward by the risk taken to get it, so two strategies can be compared on the same footing.
Sharpe and Sortino: return per unit of volatility
The Sharpe ratio measures return per unit of total volatility — how much reward you earned for each unit of “bumpiness” in your equity curve. Higher is better: more return, smoother ride. It’s the standard first-glance number for a reason.
Sharpe has one honest flaw: it penalizes upside volatility the same as downside. But a huge up-week isn’t a problem you want to be punished for — nobody complains about violent gains. The Sortino ratio fixes this by counting only downside deviation. It rewards you for the volatility that hurts and ignores the volatility that helps, which makes it a better fit for how traders actually experience risk.
- Sharpe — reward per unit of all volatility. Good baseline, slightly pessimistic on upside.
- Sortino — reward per unit of downside volatility only. Closer to felt risk.
Both are ratios, so their real value is comparative — one strategy’s Sharpe against another’s, not a single number in isolation.
MAR ratio: return relative to max drawdown
For prop-firm traders, the most viscerally relevant metric is the MAR ratio: annualized return divided by maximum drawdown. It answers the question your funded account actually cares about — how much did you make relative to the worst hole you dug?
MAR speaks the same language as your firm’s rules. Your drawdown limit is a hard wall; a strategy with a high MAR earned its return without approaching that wall, while a low-MAR strategy bought its return by repeatedly staring into the abyss. Two strategies with identical returns and very different max drawdowns are wildly different bets, and MAR is the number that separates them.
| Metric | Reward | Divided by | Best for |
|---|---|---|---|
| Sharpe | Return above risk-free | Total volatility | A general smoothness score |
| Sortino | Return above risk-free | Downside volatility | Judging felt, harmful risk |
| MAR | Annualized return | Max drawdown | Prop accounts with hard drawdown walls |
Higher is better for all three, and none of them is meaningful without a decent sample of trades behind it.
Reading these metrics for a trading account
Numbers without context are noise. A few honest guidelines:
- Direction over decimals. Whether your Sortino is rising or falling month over month matters more than its precise value. Trend beats snapshot.
- Sample size gates trust. A dazzling ratio over a dozen trades is a coin flip dressed as an edge. These metrics stabilize only over a real track record.
- Combine them. Strong Sharpe but weak MAR means your average ride is smooth but your worst moment was ugly — exactly the profile that breaches a drawdown limit. Read them together, not in isolation.
These sit alongside your core edge stats — expectancy and R-multiples — not in place of them. Expectancy tells you if you have an edge at all; risk-adjusted ratios tell you how efficiently you’re harvesting it. Get comfortable with expectancy and the R-multiple first, then layer these on top.
Using them to compare strategies fairly
The killer application is honest comparison. When you run two setups, or the same setup at two risk levels, raw P&L lies to you — the one that made more probably just risked more. Risk-adjusted ratios strip that out and put both on equal footing: return per unit of risk, so the comparison is apples to apples.
This is where a dedicated journal beats a spreadsheet. Computing rolling Sharpe, Sortino, and MAR per strategy by hand is tedious and error-prone; tools built for it — see how a purpose-built platform stacks up versus a generic journal — compute them continuously so you’re comparing live, not reconstructing after the fact. Shibiki takes this further by attaching a Wilson confidence interval to each strategy’s edge, so you can tell a genuinely superior setup from one that’s merely had a luckier sample — the difference between a real edge and a hot streak.
Improving the ratio by cutting risk, not chasing return
Here’s the liberating insight buried in every one of these formulas: they’re fractions. You improve a fraction by raising the numerator or lowering the denominator — and the denominator is almost always the easier lever.
Chasing a higher return to lift your Sharpe means taking more risk, which often inflates the very volatility in the denominator and leaves the ratio flat. But cutting your worst drawdowns — sizing down, sitting out marginal setups, enforcing a hard daily stop — shrinks the denominator directly and lifts every ratio at once, no new market prediction required. Improving risk-adjusted returns is usually a subtraction problem, not an addition one. The best trade you make this month may be the reckless one you didn’t take.
Related: Expectancy Calculator · Trading Expectancy Explained · R-Multiple Explained