Two strategies each return $10,000 over a 12-month backtest. One does it through a smooth, consistent equity curve with a 6% maximum drawdown. The other achieves the same result through a series of near-blowups, recovering each time by luck. Net profit: identical. Quality: completely different.

Profit tells you the destination. It tells you nothing about how you got there, how repeatable the journey is, or whether you could survive doing it again. These metrics do.

The Analysis Framework: Four Tiers

1

Survival Metrics — Does this strategy avoid ruin?

Max drawdown, max drawdown duration, recovery factor

2

Edge Metrics — Is there a genuine, repeatable edge?

Profit factor, expectancy, Sharpe ratio

3

Consistency Metrics — Does it perform reliably over time?

Monthly win rate, rolling profit factor, performance by market condition

4

Context Metrics — Is the result real or manufactured?

Total trades, trade frequency, modeling quality, in-sample vs OOS comparison

Always evaluate from the bottom up — survival first, context last. A strategy that fails Tier 1 doesn't need further analysis. A strategy that passes all four tiers with solid numbers is worth serious consideration.

Tier 1: Survival Metrics

Maximum Drawdown ESSENTIAL
Max DD = (Peak Equity − Trough Equity) / Peak Equity × 100
The largest peak-to-trough decline in account equity. This is your worst-case scenario in the historical data — and as covered in our drawdown article, live trading typically exceeds this by 1.5–3×. Evaluate it not just as a number but as a question: can I psychologically and financially survive 2.5× this drawdown?
Conservative
< 15%
Acceptable
15–25%
High Risk
> 30%
Recovery Factor ESSENTIAL
Recovery Factor = Net Profit / Max Drawdown
How much profit the strategy generates for every unit of drawdown it requires. A recovery factor of 1.0 means the strategy earned exactly as much as its worst loss. Higher is better — this metric rewards strategies that generate strong returns without requiring large drawdowns to do so.
Weak
< 2.0
Good
2–5
Strong
> 5.0

Tier 2: Edge Metrics

Profit Factor ESSENTIAL
Profit Factor = Gross Profit / Gross Loss
The ratio of total winning trade profits to total losing trade losses. A profit factor of 1.5 means the strategy makes $1.50 for every $1.00 it loses in aggregate. This is more reliable than net profit because it measures the quality of the edge, not just the accumulated result. Watch for overfitting: a profit factor above 2.5 on in-sample data deserves skepticism.
Marginal
1.0–1.3
Solid
1.4–2.0
Suspicious
> 2.5
Expectancy ESSENTIAL
Expectancy = (Win Rate × Avg Win) − (Loss Rate × Avg Loss)
The average amount you expect to gain or lose per trade. A positive expectancy means the strategy is profitable in aggregate regardless of the sequence of wins and losses. This is the most direct measurement of edge and should always be positive — if it isn't, no amount of risk management can save the strategy.
I used to evaluate strategies entirely by net profit and win rate. I deployed three systems that looked great by those metrics. All three lost money live. When I went back and calculated profit factor and expectancy properly, all three were red flags I had completely missed.

Tier 3: Consistency Metrics

Monthly Performance Distribution

A strategy that made $10,000 in one month and lost $3,000 across the other 11 months has a very different risk profile than one that made approximately $830 every month. Break the backtest results into monthly segments and examine the distribution. How many losing months were there? What was the worst month? Does performance cluster around certain market conditions?

Rolling Profit Factor

Calculate the profit factor on rolling 3-month windows throughout the backtest. If the rolling PF varies wildly — strong in some periods, weak in others — the strategy may be market-regime dependent. If it's consistently above 1.0 across all windows, the edge is more likely to persist.

Key Test

Calculate your strategy's profit factor on just the last 20% of your backtest period — the most recent data. If it's substantially lower than the full-period PF, your strategy may be losing its edge at the exact moment you'd be deploying it live.

Tier 4: Context Metrics

Trade Count

A backtest with fewer than 100 trades is statistically insufficient. With small sample sizes, even a genuinely poor strategy can produce impressive-looking results by chance. Aim for at least 200–300 trades before drawing conclusions. Longer backtests aren't always better if the market conditions changed dramatically during that period.

Modeling Quality

In MT4/MT5, the modeling quality percentage tells you how accurately the tick data was reconstructed during the backtest. A modeling quality below 90% introduces significant noise into the results. For strategies with short-duration trades or tight stops, even 99% modeling quality may not fully represent real execution conditions.

Putting It Together: The One-Page Analysis

When evaluating any strategy, run through this sequence before considering the net profit number:

1. Can I survive 2.5× the max drawdown? If no, stop here. 2. Is the recovery factor above 2.0? 3. Is profit factor between 1.4 and 2.5? 4. Is expectancy positive? 5. Is monthly performance reasonably consistent? 6. Are there at least 200 trades? 7. Is there out-of-sample validation?

A strategy that passes all seven is worth forward testing with minimal capital. A strategy that fails any of the first four questions is not ready for live deployment — regardless of how impressive the net profit looks.

Run the Full Analysis on Your EA

EA Analyzer Pro calculates profit factor, recovery factor, expectancy, drawdown metrics and more — free, from your MT4/MT5 backtest report.

Open EA Analyzer Pro →