Metrics That Matter (And Ones That Don't)

Win rate is vanity. Profit factor is sanity. Sharpe ratio is reality.

Every trading backtest spits out dozens of metrics. Some are essential for understanding whether a strategy works. Others are distractions that make losing strategies look good. Knowing the difference is critical for honest validation.

Win Rate: Overrated

Win rate, the percentage of trades that are profitable, is the most intuitive metric and the most overrated.

A high win rate feels good psychologically. Winning 80% of trades sounds impressive. But win rate alone tells you nothing about profitability. A strategy that wins 90% of trades but loses 10 times as much on losers as it makes on winners will destroy your account.

Conversely, many highly profitable strategies have low win rates. Trend-following strategies often win only 30-40% of trades but make so much on the winners that overall performance is strong.

Win rate matters for execution psychology and it matters when combined with average win and loss sizes. On its own, it's nearly useless for evaluation.

Profit Factor: Essential

Profit factor is the ratio of gross profit to gross loss. If your strategy made $50,000 in winning trades and lost $30,000 in losing trades, your profit factor is 50,000/30,000 = 1.67.

Profit factor greater than 1.0 means the strategy is profitable. Below 1.0 means it's losing money.

The beauty of profit factor is that it captures the full picture: win rate, average win, and average loss all rolled into one number. A strategy with 40% win rate but 2.5 average win/loss ratio will have a healthy profit factor. A strategy with 80% win rate but 0.2 average win/loss ratio will have a terrible profit factor.

For live trading, we target profit factors above 1.5 for signal strategies. Between 1.0 and 1.5 is marginal. Below 1.0 is obviously unprofitable.

Profit factor can be manipulated by excluding losing trades or cherry-picking time periods, so verify it's calculated on the full trade history with realistic assumptions.

Sharpe Ratio: Gold Standard

The Sharpe ratio measures risk-adjusted returns. It answers the question: how much return do you get per unit of risk?

Calculated as (average return - risk-free rate) / standard deviation of returns, the Sharpe ratio penalizes strategies with volatile returns. Two strategies with identical total returns can have very different Sharpe ratios if one achieved those returns steadily while the other whipsawed wildly.

Why does this matter? Because volatile strategies require larger drawdowns to achieve their returns. A strategy with a Sharpe ratio of 0.5 might need to endure 40% drawdowns to capture its long-term edge. A strategy with a Sharpe ratio of 2.0 might only face 10% drawdowns for similar returns.

Higher Sharpe ratios also indicate more reliable edges. Random noise produces low Sharpe ratios. Consistent alpha produces high ones.

In practice, annual Sharpe ratios above 1.0 are decent, above 2.0 are excellent, and above 3.0 are exceptional and should be viewed skeptically (possibly overfit or misreported).

Maximum Drawdown: Survival Metric

Maximum drawdown measures the largest peak-to-trough decline in your equity curve. If your account went from $100,000 to $70,000 before recovering, your maximum drawdown was 30%.

This metric matters for survival. No strategy works if you blow up before it has time to play out. Large drawdowns also cause psychological damage that leads to bad decisions like abandoning strategies at the worst time.

The relationship between expected return and maximum drawdown determines whether a strategy is tradeable. A 50% maximum drawdown for 100% annual return might be acceptable. A 50% maximum drawdown for 20% annual return is terrible.

We evaluate maximum drawdown in context: how long did recovery take? Did the strategy experience multiple similar drawdowns or was it a one-time event? What market conditions caused the drawdown?

Expectancy: The Mathematical Edge

Expectancy calculates the average profit per trade.

The formula: (Win Rate × Average Win) - (Loss Rate × Average Loss)

A strategy with 60% win rate, $100 average win, and $80 average loss has expectancy of (0.60 × $100) - (0.40 × $80) = $60 - $32 = $28 per trade.

Expectancy tells you what to expect from a large number of trades. Positive expectancy means you make money over time. Negative expectancy means you lose. Zero expectancy means you're paying transaction costs to break even.

Simple as it is, expectancy grounds your evaluation in mathematics rather than feelings. Every trade either adds to or subtracts from your statistical edge. Knowing your expectancy keeps you focused on process over outcome.

How Metrics Can Mislead

Every metric can be gamed or misrepresented.

Survivorship bias in metrics: Reporting only strategies that survived to the present moment inflates all metrics because failed strategies aren't included.

Optimized time periods: Choosing start and end dates that maximize performance makes any metric look better than reality.

Ignoring transaction costs: A strategy with 1.8 profit factor before costs might have 0.9 profit factor after costs. Always calculate metrics net of realistic transaction costs.

Ignoring slippage: Backtests often assume perfect execution. Real slippage degrades every metric.

Annualized metrics on short periods: Annualizing a 2-month return makes the numbers look bigger but doesn't give meaningful annualized projections.

Always understand how metrics were calculated and what assumptions went into them. Request full trade history and recalculate yourself if anything seems suspicious.

Our Metric Priorities

At TargetHit, we evaluate strategies using this hierarchy:

First filter: Profit factor above 1.5 with 1,000+ trade sample size. This eliminates unprofitable strategies and those without statistical validity.

Second filter: Sharpe ratio above 1.0 annually. This ensures risk-adjusted performance is meaningful.

Third filter: Maximum drawdown below 25%. This ensures the strategy is survivable.

Fourth filter: Consistent performance across walk-forward windows. This ensures robustness rather than overfitting.

If a strategy passes all four filters, it's a candidate for further testing and potential deployment.

Key Takeaways

Win rate alone is meaningless without average win/loss context. Profit factor captures the full picture of strategy profitability. Sharpe ratio measures risk-adjusted returns and is the gold standard for comparison. Maximum drawdown determines survivability and psychological sustainability. Expectancy quantifies your mathematical edge per trade. Every metric can be manipulated, so understand the calculations and verify assumptions.

Next, we'll cover stress testing with Monte Carlo simulation to see whether your strategy can survive black swan events.