“A PF of 1.5 or above is the benchmark for a good EA” — this is a number you will inevitably encounter when researching FX EAs. But is it really true? Choosing an EA based on PF alone will almost certainly lead to regret. This article breaks down the true meaning of Profit Factor (PF) and the pitfalls that PF by itself cannot reveal.
The Profit Factor (PF) Formula
PF = Total Profit / Total Loss
For example, with 1,000,000 yen in profit and 600,000 yen in losses, the PF is 1.67. Exceeding 1.0 means you are theoretically in positive territory.
Interpreting PF by Range
| PF | Rating | Interpretation |
|---|---|---|
| < 1.0 | ❌ Unacceptable | Already losing money. Not worth evaluating. |
| 1.0 – 1.3 | ⚠️ Noise Range | A small spread difference or slippage can push it into negative territory. |
| 1.3 – 1.5 | ○ Possible | Worth considering as a candidate with additional verification in live conditions. |
| 1.5 – 2.0 | ◎ Good | Suitable for live trading. Confirming max DD and trade count is a prerequisite. |
| 2.0 – 3.0 | ★ Very Good | However, too good to be true. Suspect over-optimization or period-specific performance. |
| 3.0+ | 🚨 Red Flag | High likelihood of curve fitting or a martingale/averaging EA. |
Why Relying on PF Alone Will Get You Burned
1. PF Is Unreliable When the Trade Count Is Low
A PF of 3.0 from 10 trades versus a PF of 1.5 from 1,000 trades — the latter is overwhelmingly more reliable. Statistical significance is determined by trade count. Evaluate with a minimum of 100 trades, preferably 300 or more.
2. Always Look at PF Together with Max Drawdown
Even a PF of 2.0 paired with a max DD of 60% will almost certainly blow up your account in live trading. The PF / Max DD ratio (high PF, low DD) is what matters. The benchmark is PF ≥ 1.5 and Max DD ≤ 30%.
3. The Meaning Changes Depending on Lot Control
Averaging and martingale EAs are structurally prone to showing high PF. However, because they are designed to “win big when winning and drag out losing streaks until the account blows up,” the PF figure does not reflect reality (Why Averaging EAs Face Serious Drawdowns). Treat the PF of variable lot EAs with more skepticism than fixed lot EAs.
4. PF Varies Significantly by Time Period
A long-term backtest from 2015 to 2023 might show PF 1.8, yet isolating 2020 alone could reveal PF 0.7. Look at how PF fluctuates across different market environments. Checking PF trends on a yearly and quarterly basis reveals which market conditions an EA thrives in and which it struggles with.
5. Balance with Risk-Reward Ratio
Even with the same PF of 1.5, an EA with a 70% win rate × RR 1:0.7 and one with a 35% win rate × RR 1:3 are completely different in character. The former has fewer losing trades and is easier on the nerves; the latter wins big when it does, but losing streaks are brutal. Determine which suits your trading style.
A Practical Checklist for Reading PF
- Are there 300 or more trades?
- Is the max DD 30% or below?
- Is it a fixed lot or variable lot EA?
- Is the annual PF trend stable? (Are there any extreme outlier years?)
- Does the win rate and RR balance match your trading style?
- Does the market environment of the backtest period resemble the intended live trading period?
- Was the backtest run with realistic (worst-case) spreads?
FX AI Lab’s Verification Stance
All EA verification data published by our lab includes the following without exception.
- Trade count (minimum 300)
- PF + Max DD + Win Rate + RR
- Annual PF trend
- Backtest spread conditions (realistic worst-case values)
- Forward test results (reproducibility in live conditions)
We never rely on marketing claims like “PF 2.0, 95% win rate” alone. We explicitly show in data which market conditions the EA works in and which it does not. That is what we believe true “verification” means.
Related Links
This article is intended for informational purposes only and does not constitute a recommendation of any specific trade or instrument. FX and EA trading carries the risk of capital loss. Please refer to our Risk Disclosure for details.