EA Portfolio Strategy — Diversify Risk by Combining Multiple EAs
2026-05-18
The approach of 'find one great EA and run it' leads to large drawdowns when the market environment that EA excels in ends. Just as institutional investors combine multiple non-correlated strategies, the portfolio mindset is equally important for individual EA trading.
Contents
Why Combine Multiple EAs
EA performance depends on market conditions. Trend-following EAs generate large profits in strong-trend markets but lose repeatedly in sideways markets. Conversely, range-breakout EAs excel after ranging periods when volatility expands, but generate excessive stop-outs during trending markets.
❌ Problems with a Single EA
- A change in market conditions causes a sudden drawdown
- More than 60% of the year is not the EA's "ideal" market
- Psychologically difficult to continue trading
- Risk concentrates on a single strategy
✅ Benefits of Multiple EAs
- Drawdowns are distributed over time
- Profit opportunities exist in any market environment
- Maximum drawdown is mathematically reduced
- Trading can continue even when one EA underperforms
What Is the Correlation Coefficient — The Key to Combining EAs
The correlation coefficient (−1 to +1) shows how similarly the returns of two EAs move. A lower correlation coefficient means a higher portfolio effect and less overlap in drawdowns.
| Correlation | Meaning | Portfolio Effect | Recommendation |
|---|---|---|---|
| 0.0–0.3 | Near zero correlation | Maximum drawdown smoothing | ⭐⭐⭐ Ideal |
| 0.3–0.6 | Low–medium correlation | Some risk diversification benefit | ⭐⭐ Acceptable |
| 0.6–0.8 | Medium–high correlation | Limited diversification effect | ⭐ Use with caution |
| 0.8–1.0 | High correlation | Equivalent to taking 2× the risk | ❌ Not recommended |
Our 3-Strategy Portfolio Example
The three XAUUSD H1 EAs available on this site are designed to profit in mutually different market conditions.
Stable core profit engine
XAUUSD M30 突破+D1/H4 分形8策略。在強趨勢行情中高表現。Exness 7.7年回測 ×92.3・PF1.51(v2.35)。收益的支柱。
Best market
Strong trending market with high ADX
Worst market
Sideways market with low ADX
Drawdown smoother / complementary role
USDJPY 唐奇安通道突破(H1)。與黃金資產/價格走勢獨立的低相關補充。Dukascopy 真實tick 9.4年驗證 PF1.55。
Best market
Markets with a sustained directional trend on H1
Worst market
Directionless ranging markets
High-precision trend-follower / profit engine
BTCUSD 4小時線(H4)趨勢跟隨。加密資產的價格驅動因素與黃金/外匯不同，可強化分散配置。實盤8年回測 PF1.72。
Best market
H4趨勢明確的行情
Worst market
H4無趨勢的震盪行情
兩隻加密趨勢EA加一隻外匯/日圓趨勢EA(2019–2025真實MT5全對近乎無相關)。將資金向最強腿傾斜的最佳配比(BTC65%/ETH25%/USDJPY10%)把CAGR提升到32.5%，而回撤維持在8.4%。
| EA | CAGR | 最大回撤 |
|---|---|---|
| BITCOIN COMET | +40.6% | 12.3% |
| ETHEREUM TREND | +14.7% | 11.6% |
| MEGAMAX USDJPY | +13.9% | 10.6% |
| 等權 (各1/3) | +24.3% | 6.7% |
| 推薦配比 65/25/10 | +32.5% | 8.4% |
BTCUSDm / ETHUSDm H4 + USDJPYm H1・2019–2025 真實MT5。依真實淨值曲線最佳化配比: 65/25/10的報酬/回撤最佳(CAGR32.5%/回撤8.4%)，優於等權(24.3%/6.7%)。各腿皆設停損（BTC/ETH腿由EA執行，券商端無SL），無網格無馬丁。
下載 Multi-Asset Trend Pack →Lot Sizing and Practical Risk Management
When running multiple EAs, the most important thing is ensuring the combined risk does not exceed your tolerance. If each EA is set to risk 1% per trade, three EAs entering simultaneously results in 3% combined risk.
| Scenario | Risk per EA | Max combined risk | Recommended use |
|---|---|---|---|
| Conservative (recommended) | 0.5% | ~1.5% | 3 EAs running simultaneously — safe operation |
| Standard | 0.7% | ~2.1% | Balanced approach |
| Aggressive | 1.0% | ~3.0% | Only after confirming EA performance in live trading |
Common Mistakes
❌ Running multiple EAs of the same type on the same currency pair
例：同時運行兩個黃金趨勢跟隨型EA。兩者在相同的漲跌中同時盈虧，相關性超過0.8，只會讓風險加倍，沒有任何分散效果。
❌ Using the same Magic Number
If different EAs share the same Magic Number, one EA may accidentally close another EA's positions. Always assign separate values.
❌ Comparing EAs with different backtest periods
Profit Factor and win rates from a 10-year backtest and a 2-year backtest are not directly comparable. Always re-verify under the same period and conditions before deciding on a combination.
❌ Stopping one EA when drawdowns overlap
When both EAs lose at the same time, the psychological urge to stop the losing one is strong — but this destroys the portfolio. Define the combined DD tolerance in advance and set rules for what to do when that limit is reached.
FAQ
Q: How many EAs is optimal to run simultaneously?
A: 2–4 verified EAs (10-year backtest + at least 4 weeks of forward testing) is realistic. More than that makes management complex and reduces the attention each EA deserves. If you are new to this, start with one stable EA and add more gradually as you verify performance.
Q: Is it better to combine EAs on different currency pairs?
A: Different currency pairs can still move similarly in the same market environment (risk-on/risk-off). For example, GOLD (XAUUSD) and USD/JPY can be correlated due to USD strength. True diversification comes from spreading across strategy types (trend-following / breakout / range), not just currency pairs.
Q: How do I calculate the correlation coefficient?
A: You can use Excel's Data Analysis tool or Python's `pandas.DataFrame.corr()`. List each EA's daily P&L (daily +/−) and compute the correlation. You will need backtest data covering the same period. MT5 Strategy Tester reports can be exported on a daily basis for this purpose.
Q: What should I do if one EA enters a deep drawdown?
A: Follow the rules you set in advance for when the "allowed DD" is exceeded. For example, decide before you start: "If a single EA exceeds 15% drawdown of account balance → pause and re-examine parameters." Avoid emotional decisions.
Q: Can I run multiple EAs on the same account?
A: Yes, you can. However, always assign separate Magic Numbers to each EA. MT5 uses the Magic Number to identify each EA's positions even on the same chart. Also make sure the combined open positions do not compress the account's margin level.
Q: Can I add EAs from other vendors to the portfolio on this site?
A: You can, but verify the correlation with third-party EAs as well. In particular, adding multiple EAs with similar strategies on the same currency pair can unintentionally concentrate risk. Compare backtest data over the same period before deciding.
Q: Can I include Beta EAs in the portfolio?
A: Run a 10-year backtest in MT5 Strategy Tester and confirm PF ≥ 1.2, 100+ trades, and max DD ≤ 15% before moving to forward testing. After confirming performance close to the backtest level over at least 4 weeks of forward testing, start on a live account with a small size (risk ≤ 0.3%).
Q: What overall Profit Factor should I target for the portfolio?
A: You cannot simply add up the PF of each EA. Combining low-correlation EAs reduces overall portfolio drawdown, but also moderates total profit somewhat. Use these as rough benchmarks: each EA ≥ PF 1.3 individually, overall portfolio max DD ≤ 20%, and portfolio Sharpe Ratio ≥ 0.5.