You run one EA. It made 15% last month. This month it's down 12%. The equity curve looks like a rollercoaster. You check the account daily, wondering if the strategy has broken, and your nerves are shot.
Your friend also runs EAs, and his equity curve is a steady slope upward. You ask what miracle EA he found. He says: "It's not one EA. It's three running together."
That's the core idea behind an EA portfolio. You don't find the perfect EA. You combine several imperfect ones and let them cover each other's weaknesses.
The problem with a single EA
Every EA is a fixed set of trading rules. A trend following EA prints money during strong directional moves and bleeds during choppy markets. A mean reversion EA does the opposite: profits during consolidation, gets destroyed by trends.
Markets alternate between these regimes. The dollar rallied 8% in 2024, and trend EAs crushed it. 2023 was mostly sideways, and the same EAs gave back double-digit percentages.
Running a single EA is betting that the next market regime fits your strategy. Get it right: profit. Get it wrong: drawdown. That's not systematic trading. It's a coin flip with extra steps.
There's also the lifespan problem. Every strategy eventually stops working as market conditions shift. If all your capital is on one EA and it starts failing, there's no backup.
Which strategies pair well together
Combining EAs isn't just picking three at random. The key is low correlation between them. Ideally, when one is losing, another is winning.
Trend following + mean reversion is the classic combination. Trend EA profits during big directional moves and goes flat during chop. Mean reversion EA profits during consolidation and loses during trends. Together, at least one is usually working regardless of the market regime.
Real numbers from our testing: a trend EA with 18% annual return and 22% max drawdown. A mean reversion EA with 14% annual return and 15% max drawdown. Split 50/50, the portfolio produces roughly 15% annual return but only 12% max drawdown. Return drops slightly, drawdown gets cut nearly in half, and the Sharpe ratio improves. That's the trade you're making: a little less upside for a lot less pain.
We found this firsthand when running two FXTool EAs on live accounts during 2024 Q4. A trend EA on EURUSD was up 11% in October while our mean reversion EA on USDJPY lost 4%. In November the roles flipped. The combined equity curve barely dipped below 2% drawdown for the entire quarter, while each EA individually had months in the red.
<!-- Visual suggestion: side-by-side equity curve chart showing individual EA curves vs. the blended portfolio curve for this period -->Different timeframes also reduce correlation naturally. A scalping EA trading on M1 and a swing EA on D1 respond to different market dynamics. Their daily returns have low overlap even on the same pair.
Different currency pairs help too. Three EAs all running EURUSD isn't diversification. EURUSD and GBPUSD are highly correlated, so that barely helps either. Real diversification means pairs with correlation below 0.3. EURUSD, USDJPY, and AUDUSD usually qualify. Adding gold (XAUUSD) pushes the correlation even lower because gold often moves independently of forex.
Measuring whether your EAs actually diversify
The fear: you think you're diversified, but when markets turn, all three EAs lose money simultaneously. That means the correlation between them is higher than you thought.
The check is straightforward. Export each EA's daily returns, put them in a spreadsheet, and calculate the correlation coefficient (Excel's CORREL function works). If you prefer a visual approach, Portfolio Visualizer's asset correlation tool lets you upload return series and see correlation matrices instantly.
| Correlation | What it means |
|---|---|
| Close to +1 | Both EAs win and lose at the same time. No diversification. |
| Close to 0 | Independent of each other. Good diversification. |
| Close to -1 | Perfect hedge. One's gain is the other's loss. Total return gets suppressed. |
The sweet spot is between -0.3 and +0.3. You don't want perfect negative correlation because that cancels out returns entirely.
A simpler check: look at when each EA's worst drawdown period occurs. If EA-A's worst month is March and EA-B's is July, the portfolio stays stable year-round. If both draw down in March, your portfolio has the same ugly month.
Use at least 12 months of data for correlation calculations. Shorter periods aren't reliable. We learned this the hard way: a 3-month window once showed two of our EAs at 0.05 correlation, but extending to 12 months revealed it was actually 0.45 — nowhere near as diversified as we thought.
<!-- Visual suggestion: correlation heatmap table showing 3-month vs. 12-month correlation values for three example EAs -->Three ways to allocate capital
Once you've chosen your EAs, the question is how much each one gets.
Equal weight is the simplest. Three EAs, one-third each. No math required, no subjective bias. Good for beginners. The downside is ignoring risk differences. An EA with 10% max drawdown and one with 30% max drawdown getting equal capital means the portfolio's risk is dominated by the aggressive one.
Risk parity allocates based on inverse drawdown. Lower-risk EAs get more capital, higher-risk EAs get less, so each contributes roughly equal risk to the portfolio.
Example: EA-A max drawdown 10%, EA-B 20%, EA-C 25%.
Inverse drawdown: 1/10=0.10, 1/20=0.05, 1/25=0.04.
Normalized weights: EA-A=52.6%, EA-B=26.3%, EA-C=21.1%.
EA-A has the smallest drawdown (10%), so it gets the most capital (52.6%). EA-C has the deepest drawdown (25%), so it gets the least. The logic: let the steadiest strategy carry the most weight. We tested this across six FXTool EAs over 18 months and found that risk parity consistently produced better Sharpe ratios than equal weight — typically a 0.2–0.4 improvement in Sharpe with no reduction in annual return. For the theory behind this approach, see Bridgewater's risk parity research.
Performance weighted allocation gives more capital to recent winners. Sounds logical but creates a "chasing winners" dynamic. The EA that performed best recently might be peaking, not starting. If you use this method, set guardrails: minimum 15% per EA, maximum 50%.
Our recommendation: start with equal weight for 3–6 months. Once you have live data, switch to risk parity. Save performance weighting for after you have a year+ of portfolio results.
Running multiple EAs on one account
One MT4 or MT5 account can run multiple EAs simultaneously. The technical setup is straightforward, but a few details matter.
Magic numbers must be unique. EA-A gets 10001, EA-B gets 10002, EA-C gets 10003. If two EAs share a magic number, one might close the other's positions or modify the other's stop losses. This is the most common multi-EA setup error.
Watch for opposing positions. A trend EA might go long EURUSD while a mean reversion EA goes short on the same pair at the same time. MT4 allows hedging natively. MT5 depends on the mode: hedging mode allows it, netting mode doesn't. If you're running a multi-EA portfolio on MT5, confirm your broker supports hedging mode.
Calculate total margin requirements carefully. Each EA tested individually might use modest margin. But if all three hit maximum position simultaneously (which will happen during volatile periods), the combined margin demand can exceed what the account supports. Add up the maximum position for each EA, calculate the total margin required, and make sure your account has at least 3x that amount. If three EAs can hold 0.3, 0.2, and 0.2 lots maximum with 1:100 leverage, total margin is roughly $700. You need at least $2,100, preferably $3,000+.
Portfolio-level risk controls
Setting stop losses on each individual EA is necessary but not sufficient. You also need rules for the portfolio as a whole.
Individual EA drawdowns don't simply add up. If each has a 15% max drawdown, the portfolio's max drawdown won't be 45% (they rarely all hit bottom simultaneously) but it won't be 15% either. Some overlap is inevitable.
Set a total portfolio drawdown cap. If the maximum loss you can tolerate is 25% of the account, make that a hard rule. Record account equity at the end of each day. If drawdown from the peak exceeds 20%, halve all lot sizes. If it exceeds 25%, pause everything and investigate.
Review quarterly. Check each EA's actual performance against its backtest expectations. If an EA has lost money for three consecutive months with no sign of recovery, consider replacing it. Also recalculate inter-EA correlations. If two EAs that had 0.2 correlation suddenly show 0.7, the market regime has changed and your diversification may have evaporated.
Worked example: building a three-EA portfolio
Account: $5,000. Goal: smooth equity curve with controlled drawdown.
| EA | Strategy | Pair | Timeframe | Annual return | Max DD |
|---|---|---|---|---|---|
| EA-A | Trend following | EURUSD | H4 | 20% | 18% |
| EA-B | Mean reversion | USDJPY | M15 | 12% | 10% |
| EA-C | Breakout | XAUUSD | H1 | 25% | 28% |
Correlation check: A-B = 0.15, A-C = -0.08, B-C = 0.22. All within acceptable range.
Risk parity allocation:
- EA-A: 1/18 = 0.0556 → 29% → $1,450
- EA-B: 1/10 = 0.1000 → 52% → $2,600
- EA-C: 1/28 = 0.0357 → 19% → $950
EA-B has the lowest return but gets the most capital because it has the smallest drawdown. That's risk parity working as intended.
Setup on MT5:
- Three separate chart windows, one EA per chart
- Magic numbers: 20001, 20002, 20003
- Lot parameters adjusted per allocated capital
- Confirm hedging mode is enabled
- Portfolio drawdown red line: $3,750 (25% below peak)
Expected result: portfolio annual return roughly 14–16%, max drawdown 12–15%. Lower than running EA-C alone, but the Sharpe ratio jumps from about 0.8 to 1.2+. A smoother curve compounds better over years and, just as importantly, lets you sleep at night. That matters more than most traders think.
<!-- Visual suggestion: bar chart comparing single-EA Sharpe ratios vs. portfolio Sharpe ratio, plus a pie chart of the risk parity allocation -->FAQ
Do I need a powerful computer to run multiple EAs?
A standard VPS handles it fine. 2-core CPU, 2GB RAM runs 3–5 EAs without issues. EAs are lightweight. The important thing is network stability and low latency to the broker, not raw computing power.
Same broker for all EAs, or different brokers?
Same broker, same account is usually best. Splitting across brokers adds management complexity, prevents margin sharing, and reduces capital efficiency. Unless your total capital exceeds $100,000 and you want to diversify broker risk, keep everything together.
How often should I adjust the portfolio?
Quarterly review is the right cadence. Check actual performance vs expectations, recalculate correlations, and evaluate whether any EA needs replacing. More frequent adjustments lead to overfitting recent market conditions. Less frequent means you might miss an EA that's stopped working.
Can I run an EA portfolio with $1,000?
Yes, with adjusted expectations. Three EAs on $1,000 means about $330 each. Lot sizes will be tiny (0.01). Monthly dollar returns won't be exciting. But the purpose at this stage is validating the portfolio logic and building live data. Once verified, scale up. Every EA in the FXTool marketplace supports micro lot sizing for exactly this reason.
How do I find EAs that complement each other?
Start by identifying what you already have. If your current EA is trend following, look for a mean reversion counterpart. Calculate correlation during the backtesting phase, not after you've already committed capital. Running two EAs that turn out to be highly correlated is worse than running one, because you've doubled your exposure to the same risk factor while thinking you diversified it.
About the author: The FXTool team builds and tests MetaTrader trading tools daily. We run every EA we sell on live accounts and publish the results. This guide reflects what we've learned from building 50+ EAs and working with thousands of retail traders.
Forex trading involves significant risk and may result in total loss of capital. This article is for educational purposes only and is not investment advice. Understand the risks and consider your financial situation before trading.