Sharpe Ratio, Max Drawdown, and Win Rate: How to Read Backtest Metrics
To read a backtest correctly, judge a strategy on three metrics together: the Sharpe ratio (risk-adjusted return), maximum drawdown (worst peak-to-trough loss), and win rate (share of profitable trades). No single number tells the whole story—win rate in particular is the easiest to misread.
Backtesting turns a trading idea into numbers, but numbers can flatter a weak strategy if you look at only one at a time. This guide defines each core metric in plain English and, more importantly, shows how to interpret them as a set so you can tell a durable edge from a fragile one.
What does the Sharpe ratio actually measure?
The Sharpe ratio measures return per unit of risk. It divides a strategy’s excess return (return above a risk-free rate) by the volatility of those returns. A higher Sharpe means you earned more reward for each unit of ups-and-downs endured—rewarding smooth, consistent gains over erratic ones.
Two strategies can post the same total return, but the one with a higher Sharpe got there with less turbulence. As rough context, values are often read like this:
| Sharpe ratio | Common interpretation |
|---|---|
| Below 1 | Modest risk-adjusted return |
| 1 to 2 | Generally considered solid |
| Above 2 | Strong, but check for overfitting |
Two cautions. First, Sharpe penalizes upside and downside volatility equally, so a strategy with occasional large gains can look “riskier” than it feels. Second, an unusually high backtested Sharpe often signals curve-fitting to historical noise rather than a real edge. Treat extreme numbers with skepticism and confirm them on out-of-sample data.
What is maximum drawdown (MDD) and why does it matter?
Maximum drawdown (MDD) is the largest peak-to-trough decline in your equity curve, expressed as a percentage. It answers a blunt, practical question: from its best point, how much did the account fall before recovering? MDD captures the worst-case pain a strategy would have put you through.
MDD matters because it maps to two things a spreadsheet can’t feel: survivability and psychology. A strategy that returns 40% but suffers a 60% drawdown may be untradeable in practice—you might get liquidated, hit a margin call, or simply abandon it at the bottom. Recovering from drawdowns is asymmetric:
- A 20% drawdown needs a 25% gain to break even
- A 50% drawdown needs a 100% gain to break even
- A 60% drawdown needs a 150% gain to break even
This is why experienced traders often prize a shallow MDD over a slightly higher return. Tools that cap risk per trade—like a trailing stop-loss—are frequently used to keep drawdowns inside a tolerable range. When you backtest a trading strategy, read MDD alongside how long the recovery took, not just its depth.
Why is win rate misleading on its own?
Win rate—the percentage of trades that closed profitable—is misleading alone because it ignores the size of wins and losses. A 90% win rate is worthless if the occasional 10% of losing trades are large enough to erase every gain. Win rate says how often you win, never how much.
Consider two strategies:
| Metric | Strategy A | Strategy B |
|---|---|---|
| Win rate | 80% | 40% |
| Average win | $50 | $400 |
| Average loss | $300 | $100 |
| Net result | Loses money | Makes money |
Strategy A wins four times as often yet bleeds money, because its rare losses dwarf its frequent small wins. Strategy B is wrong more often than right but is highly profitable. Trend-following systems commonly have win rates below 50%; mean-reversion and grid strategies often show high win rates paired with a painful tail risk when the market trends against them.
The fix is to pair win rate with two partners:
- Risk/reward ratio — average win size divided by average loss size
- Profit factor — gross profit divided by gross loss (above 1.0 means net profitable)
Together these turn “how often” into “does it actually make money.”
How do you read Sharpe, MDD, and win rate together?
Read the three as a system: win rate and risk/reward tell you whether the edge is profitable, Sharpe tells you how smoothly it earns, and MDD tells you whether you could survive the rough patches. A strategy needs to pass all three—not maximize any one.
A practical checklist when reviewing a backtest:
- Is it profitable? Confirm a positive expectancy via profit factor or net return—never win rate alone.
- Is the ride tolerable? Check MDD against your personal risk limit and how long recoveries lasted.
- Is the return efficient? Use Sharpe to compare against alternatives with similar drawdowns.
- Is it real? Validate on out-of-sample data and enough trades that results aren’t luck.
- Does it fit your venue? Costs, fees, and slippage differ across perpetual futures and prediction markets.
Beware the trap of optimizing one metric. Push win rate too high and you often inherit fat-tailed losses; chase Sharpe and you may overfit; ignore MDD and you build something you’ll never stick with in a real drawdown. Balance beats extremes.
How does Hey-Traders help you read these metrics?
Hey-Traders is a natural-language quant platform: you describe a strategy in plain English, AI converts it into executable code, and it backtests on historical data with professional metrics—Sharpe ratio, maximum drawdown, win rate, and a full equity curve—no coding required. You see all three side by side instead of fixating on one.
From there you can generate live signals and, once you enable trading permissions, automate execution on supported venues such as Binance, Bybit, OKX, Hyperliquid, and Polymarket, subject to venue support. Because Hey-Traders lets you iterate on plain-English descriptions, it’s straightforward to add risk controls—stops, take-profits, or position sizing—and re-run the backtest to see how each change moves your MDD and Sharpe. Learn more about the approach in natural-language trading and how algorithmic trading works, or review supported order types.
A brief, honest note: trading involves risk of loss, leverage amplifies that risk, and backtested results do not guarantee future performance. Metrics describe the past; they narrow uncertainty, they don’t remove it.
Frequently Asked Questions
Is a high win rate good or bad?
Neither on its own. A high win rate is only good if losses stay small relative to wins. Many profitable trend-following systems win less than half their trades, while some high-win-rate strategies lose money because rare losses are oversized. Always pair win rate with risk/reward or profit factor.
What is a good Sharpe ratio for a trading strategy?
As a rough guide, a Sharpe between 1 and 2 is often considered solid, and above 2 is strong. Be skeptical of very high backtested values, which frequently indicate overfitting rather than a durable edge. Confirm on out-of-sample data.
What maximum drawdown is acceptable?
There’s no universal number—it depends on your risk tolerance and capital. The key is that MDD stays within a range you could psychologically and financially survive, since a 50% drawdown requires a 100% gain just to break even.
Why can’t I judge a strategy by return alone?
Total return ignores the risk taken to earn it. Two strategies with identical returns can have very different drawdowns and volatility. Reading Sharpe, MDD, and win rate together reveals whether the return was efficient, survivable, and repeatable.
Does Hey-Traders calculate these metrics automatically?
Yes. When you describe and backtest a strategy on Hey-Traders, it reports Sharpe ratio, maximum drawdown, win rate, and an equity curve automatically, so you can evaluate all three together without writing code.
Ready to see these metrics on your own idea? Describe a strategy in plain English on Hey-Traders, run the backtest, and read Sharpe, MDD, and win rate side by side before you risk a dollar.