Why Strategy Metrics Matter
Net profit is the first number everyone looks at and the least informative one on its own. Two strategies can make the same profit while one of them fell 40% along the way. Metrics tell you how the profit was made: how deep the losses went, how consistent the returns were, and whether a few lucky trades carried the result.
Every metric here describes a historical simulation. None of them forecasts what a strategy will do next.
Where You'll See These Metrics
TradingView's Strategy Tester shows a performance summary for a Pine strategy, with measures such as net profit, maximum drawdown, percent profitable, profit factor, and the Sharpe and Sortino ratios.
Roboquant backtests report net P&L, total return and monthly returns; maximum drawdown, Sharpe and Sortino; win rate, profit factor, average trade and largest win and loss; and, where available, MAE and MFE, alpha, beta and probabilistic Sharpe.
Names, formulas and settings differ between platforms. A Sharpe ratio from one tool is not directly comparable with a Sharpe ratio from another unless both use the same return frequency and assumptions, so compare strategies inside the same tool.
Net Profit and Total Return
Net profit (net P&L) is gross profit minus gross loss, after the commissions and slippage the backtest was configured with. Total return expresses it as a percentage of starting capital.
Both depend on position size and starting capital, so on their own they say nothing about risk. Always read them next to maximum drawdown.
Win Rate (Percent Profitable)
Win rate = winning trades ÷ total trades.
A high win rate feels good and proves little. A strategy that wins 70% of the time with an average win of $100 and an average loss of $300 loses money:
0.70 × $100 − 0.30 × $300 = $70 − $90 = −$20 per trade
Low-win-rate strategies can be profitable when the winners are much larger than the losers. Read win rate together with average win and average loss.
Average Trade and Largest Win or Loss
Average trade = net profit ÷ number of trades. It shows how much room the strategy has for costs. If the average trade is only slightly larger than your commission plus slippage, small changes in execution can erase the edge.
Largest win and largest loss show concentration. If one trade produced most of the net profit, the result depends on repeating that one event.
Profit Factor
Profit factor = gross profit ÷ gross loss (the loss as a positive number).
A strategy with $30,000 of gross profit and $20,000 of gross loss has a profit factor of 1.5. Above 1.0 the strategy made money over the test; below 1.0 it lost.
There's no universal "good" profit factor. A very high value from a small number of trades is more often luck or overfitting than an edge.
Maximum Drawdown
Maximum drawdown is the largest decline in equity from a peak to a later low. If equity peaks at $50,000 and falls to $40,000 before making a new high, the drawdown is $10,000, or 20%.
Drawdowns take more to recover from than they cost:
| Drawdown | Gain needed to get back to the peak |
|---|---|
| 10% | 11.1% |
| 20% | 25% |
| 30% | 42.9% |
| 50% | 100% |
Two cautions:
- The historical maximum isn't the worst case. A future drawdown can be deeper.
- Platforms measure it differently. Some use closed-trade equity, others include open-trade swings. Check which one you're reading.
Sharpe Ratio
Sharpe ratio = (average return − risk-free rate) ÷ standard deviation of returns, usually annualized.
It measures return per unit of volatility. Higher is better, but the number depends heavily on how it's calculated:
- Return frequency. Daily, weekly and monthly returns give different values for the same strategy.
- Annualization. It's typically scaled by the square root of the number of periods per year.
- Risk-free rate. Some tools subtract one, some don't.
Its limitations:
- It treats upside volatility as risk, the same as downside volatility.
- It assumes returns are roughly symmetric and independent. Strategies with rare, large losses can show a high Sharpe until the loss arrives.
- With few trades or a short test period, the estimate is noisy.
Rules of thumb for a "good" Sharpe exist, but they depend on the horizon, the market and the calculation, so don't treat any threshold as universal.
Sortino Ratio
The Sortino ratio is the Sharpe ratio with one change: the denominator uses only downside deviation, the volatility of returns below a target (often zero). A strategy with large upside swings and small losses scores better on Sortino than on Sharpe.
Probabilistic Sharpe Ratio
The probabilistic Sharpe ratio estimates the probability that a strategy's true Sharpe ratio is above a benchmark, often zero, given the observed Sharpe, the number of return observations, and the skewness and kurtosis of those returns. The same Sharpe ratio is more convincing over a long track record with well-behaved returns than over a short one with fat tails.
Alpha and Beta
Beta measures how much a strategy's returns move with a benchmark's returns. Alpha is the part of the return that exposure to the benchmark doesn't explain. A strategy with high beta and little alpha is mostly a leveraged version of the benchmark.
MAE and MFE
- Maximum adverse excursion (MAE): the largest unrealized loss a trade showed before it closed.
- Maximum favorable excursion (MFE): the largest unrealized profit a trade showed before it closed.
They help with stops and targets. Winners that routinely go deep against you before recovering suggest a stop that barely survives. Trades that show large MFE and close small suggest the exit gives back most of the move.
How to Read a Backtest Report
- Count the trades. There's no magic minimum, but the fewer the trades, the more of the result is luck.
- Check the costs. Confirm commission and slippage are realistic for the market and size.
- Compare drawdown with what you can tolerate, in dollars and in percent.
- Check profit factor and average trade after costs.
- Look for concentration. Check the largest win and the monthly returns: did one month or one trade make the year?
- Check out-of-sample results. If you optimized, compare in-sample with out-of-sample or walk-forward results. Large degradation is a warning.
- Match the fill model to the strategy. Bar-based fills can misjudge intrabar stops and targets. Test those with tick data where it's available.
Red Flags
- A very high Sharpe ratio or profit factor from few trades.
- An equity curve with almost no drawdowns. This often points to look-ahead bias, a bug, or overfitting.
- Results that collapse when realistic costs are added.
- Optimized parameters that only work at one exact value.
Metrics in Roboquant Backtests
Roboquant runs a compiled strategy against historical data, which comes from CME today, with the fill model you choose: OHLCV bars, tick fills that follow the trade tape, or order-book fills that walk market depth (tick and order-book data on supported symbols, dates and plans). Each result includes the summary metrics listed above, entries and exits on the chart, equity and drawdown curves, the trade list with commissions and slippage, and strategy logs. Replay lets you watch a run on the chart.
In the Optimize tab you can validate parameters with in-sample/out-of-sample splits or walk-forward testing, depending on your plan. The backtesting docs cover fill models, costs and optimization in detail.
Related Reading
- Natural Language to Trading Code: How AI Builds Your Strategies
- Turn a YouTube Trading Strategy Into Code You Can Backtest
- Best Pine Script AI Generators in 2026
Try It
Run a backtest and read these metrics on your own strategy.
Trading involves risk of loss. Backtest results are hypothetical and do not guarantee future performance.