What happens if I combine several EAs with similar strategies?

They lose money at the same time. The portfolio equity curve looks like the worst individual EA curve, multiplied by total capital. Real diversification requires that the EAs are not in drawdown simultaneously.

How do I measure correlation between EAs in practice?

The most direct way: export the trades of both EAs from the MT report, calculate daily or weekly P&L, and compute the Pearson correlation coefficient. Values above 0.6 are a warning sign.

How high should the combined maximum drawdown be set?

It depends on your risk tolerance, but a practical rule is that the combined maximum drawdown of the portfolio should never exceed 20–25% of total capital. Traders working under prop-firm rules typically use 5–10% as a hard limit.

EA portfolio management: correlation & combined drawdown

Why a strong individual EA can fail in a portfolio

You have an EA with a stable equity curve, solid Sharpe ratio, and manageable drawdown. You add a second, equally solid EA. Suddenly the portfolio drawdown is larger than either EA's individually. What happened?

The mistake almost always lies in unnoticed correlation: both EAs react to the same market regime, lose money in the same periods, and add their drawdowns together rather than smoothing them out.

Real portfolio management means that a good individual system only realises its full value when combined with complementary systems.

What correlation means in a trading context

Correlation measures whether the P&L time series of two EAs tend to move in the same direction.

Correlation near +1: The EAs win and lose at the same time. No diversification benefit.
Correlation near 0: P&L movements are independent. Genuine smoothing effect.
Correlation near -1: When EA A loses, EA B gains and vice versa. Ideal — but rarely natural.

Common causes of high correlation:

Same trading symbols (EURUSD and GBPUSD often move in sync)
Similar strategy types (two trend-followers both lose in ranging markets)
Same timeframes (all EAs react to the same daily close trigger)
Shared news sensitivity (neither EA has a news filter)

Measuring correlation in practice

The quickest method: export the trade report from MT4/MT5 as a CSV, calculate daily P&L (closed trades), and compute the Pearson correlation coefficient between the equity time series of each EA.

In Excel or Python, a simple =CORREL(series_A, series_B) or df[['EA_A','EA_B']].corr() is sufficient.

Interpretation guide:

Correlation	Assessment
> 0.7	High overlap — almost no diversification
0.3 – 0.7	Partial correlation — limited smoothing
−0.3 to 0.3	Good independence
< −0.3	Natural hedge — rare

For a statistically meaningful result you need at least six months of parallel runtime with 100+ trades per EA.

Calculating combined maximum drawdown

The combined maximum drawdown (max combined DD) is the key risk metric for a multi-EA portfolio. It measures what percentage of total capital the portfolio has lost simultaneously.

Simplified approximation for uncorrelated EAs:

If EA A has a max DD of 10% and EA B has 12%, and they are completely uncorrelated, the combined DD is not 22% but closer to √(10² + 12²) ≈ 15.6% — the losses do not fully overlap statistically.

Reality check: During crisis periods (flash crashes, news gaps, weekend gaps) theoretical uncorrelation breaks down. EAs that are independent under normal conditions often lose simultaneously when it matters most. Always calculate a stress scenario — for example, all EAs simultaneously at their historical maximum DD. That is the worst-case exposure.

Capital allocation: how much for each EA?

No EA should receive so much capital that its maximum drawdown threatens the overall portfolio.

Rule of thumb for beginners:

Divide total capital equally across all EAs — simple and avoids inadvertent overweighting.
Set a hard limit: no single EA should manage more than 30–40% of portfolio capital.

Risk normalisation (advanced): Standardise the lot sizes of all EAs so each carries an identical risk per trade — for example 0.5% of total capital. This means you are comparing like for like. Details on lot-size calculation can be found in the guide on position sizing and per-trade risk.

Which EAs can be sensibly combined

Good combinations link complementary market-regime profiles:

Trend-follower + mean-reversion: when the trend-follower loses in ranging conditions, the mean-reversion EA profits.
Different symbols with low macro correlation: EURUSD and USDJPY have historically lower correlation than EURUSD and GBPUSD.
Different timeframes: a scalper (M5) and a swing trader (H4) respond to different signals.
Different session times: an EA that trades only the Asian session and one focused on the London session rarely overlap in drawdown.

Never combine multiple EAs of the same type on highly correlated pairs without a thorough correlation analysis.

Drawdown limits and automatic stop

In real portfolio operation — especially on a VPS without constant manual monitoring — an automated drawdown stop is important.

Implement a portfolio guard: a separate script or EA that checks total equity daily (or per candle) and closes all EAs when the portfolio drawdown limit is breached. This protects against the most common scenario — one EA enters an extreme loss streak while the others keep running into an already damaged account.

For prop-firm accounts this limit is often mandated by the firm's rules. More on this in the guide on passing a prop-firm challenge with an EA.

Monitoring and rebalancing

A portfolio requires regular review:

Monthly: update the correlation matrix for all EAs — markets change their regimes.
Quarterly: check whether each EA's equity curve still tracks the walk-forward expectation. Deviations are a signal of a strategy break.
After capital changes: adjust the lot sizes of all EAs proportionally — not just the growing EA.

You can track your own combined drawdown across EAs continuously with our free portfolio analysis tool. To see what a real multi-EA account looks like — including 22% drawdown and a 1.53 profit factor — view the live multi-EA showcase.

Conclusion

A multi-EA portfolio is not a safety net created simply by adding more systems. It requires thought: which EAs have independent P&L sources? What combined drawdown is acceptable? How is capital fairly distributed? Answering these questions before deploying real money builds a portfolio that can survive genuine stress periods. The foundation for this is solid infrastructure — a reliable VPS and brokers with transparent execution are not optional; they are a prerequisite. Check our broker reviews to see which providers meet the bar. To size each EA's lots, use our position-size calculator.