Running Multiple Expert Advisors: Correlation, Exposure and Portfolio Construction — an EA Portfolio Management Guide
EA portfolio management is the discipline of running several expert advisors on one account as a single, coherent risk unit rather than a collection of independent bets. Five EAs do not equal five times the returns — they equal five times the exposure if correlation is ignored. This guide covers strategy-family correlation, aggregate drawdown, position sizing and the framework for adding or removing EAs. Written for operators who already run automation and want the portfolio layer done properly.
EA portfolio management: why five EAs are not five times
The most common mistake in multi-EA trading is assuming that each additional strategy adds an independent profit stream. Returns do not compound across strategies; exposure does. If your five EAs all trade EURUSD, all use momentum entries and all trade in the same direction during a trend, you have one large EURUSD bet split across five vehicles.
Correlated losses compound faster than correlated gains. A trend-following EA and a breakout EA look like different animals on paper, but both take positions after directional movement. When a reversal hits, both stop out in the same session. The equity chart shows one deep dip where you expected five shallow ones. This is why EA portfolio management starts with honesty about what your strategies actually share, not with what their marketing pages claim. See the FCA’s consumer guidance on CFDs.
EA portfolio management: correlation between families
Strategy families — trend, mean-reversion, breakout, range, divergence, price-action and grid — have different structural relationships to the market. Trend followers and mean-reversion systems are naturally negatively correlated in trending markets: one rides the move, the other fades it. Breakout and range strategies alternate as volatility regimes change. The value of the mix is not that each family makes money but that their loss periods rarely coincide.
Measure it rather than assuming it. Record each EA’s daily equity changes and calculate the correlation coefficient between every pair. Treat any pair above 0.7 as one position and consolidate the sizing. Our EA portfolio management service does this analysis across every running strategy, flagging overlaps before they compound into a concentrated loss. Re-run quarterly: strategies uncorrelated in a trending environment converge during range conditions, and vice versa. Correlation is a property of a market regime, not a permanent property of a strategy.
EA portfolio management: aggregate drawdown
Every strategy vendor shows a drawdown figure. None of them show yours, because your drawdown is the peak-to-trough decline of the combined account equity — all open positions from every EA at once. A 15% per-strategy drawdown across four strategies can overlap and produce a 30% aggregate drawdown, and overlapping losses are exactly what happens in a market shock, because strategies converge when volatility spikes.
The recovery maths is unforgiving: a 30% drawdown needs a 43% gain to break even; 50% needs 100%. Set your drawdown cap and daily loss limit at portfolio level, not per strategy, and enforce them at account level. One EA should never breach a limit the others respect. Aggregate exposure monitoring — combined long and short per instrument — catches the quiet version: five small positions in one pair add up to one large exposure nobody approved.
Position sizing across strategies
Position sizing in a multi-EA portfolio starts with a risk budget, not a per-strategy decision. A workable default: risk up to 1% of the account per strategy, with the sum of all open-trade risks capped at a portfolio limit of 5-10%. Size the budget by correlation, not by conviction. Two uncorrelated strategies at 1% each carry less combined risk than two correlated strategies at 0.7% each, because their loss events rarely overlap. Risk per trade should scale with equity automatically — use percentage-based sizing, not fixed lots, so every strategy’s exposure breathes with the account.
Hold back part of the budget as reserve. A fully allocated portfolio cannot take the opportunities that made you want automation, and reserve capacity lets you add a genuinely uncorrelated strategy when one appears. If a strategy’s realised risk runs persistently above its budget — more slippage, wider stops, longer trade duration than modelled — cut its allocation before you cut the strategy.
When to add a strategy and when to remove one
Add a strategy only when the correlation analysis shows a gap, never because a sales page impressed you. The best candidates are from a family you do not already run, on an instrument you do not already trade, with a position-holding pattern different from the rest of the portfolio. The product matcher helps identify strategies that complement an existing line-up rather than duplicating it. Add one at a time at 20-30% of its target allocation, observe for two to four weeks live, then scale up. Never add multiple untested EAs simultaneously — if results disappoint, you will not know which one to blame.
Remove a strategy when any of three things happen: its correlation with the rest of the portfolio climbs above 0.7 and stays there; its drawdown contribution consistently dominates the aggregate figure; or live performance diverges from its stated characteristics through a full market regime, not just a rough month. Portfolio construction is a maintenance activity, not a one-off event. Quarterly rebalancing — re-measuring correlation, re-weighting allocations, retiring redundant strategies — is what keeps the portfolio doing what it was built to do.
Frequently asked questions about EA portfolio management
Why does running five expert advisors not produce five times the returns?
Returns do not compound across strategies, but exposure does. If your five EAs are correlated, they take the same trades at the same time, so you hold one concentrated position through five vehicles. Five uncorrelated strategies may smooth the equity curve, but each additional strategy adds complexity, spread costs and drawdown surface. Expect diversification of risk, not multiplication of profit.
How do I measure correlation between my expert advisors?
Record each EA’s daily returns or equity changes and calculate the correlation coefficient between every pair, using either Pearson or rank correlation. Pairs above 0.7 are effectively one bet in two shells and should be treated as a single exposure. Re-run the analysis quarterly because strategies that diverge in a trending market tend to converge when conditions change.
What is aggregate drawdown and why does it matter more than per-strategy drawdown?
Aggregate drawdown is the peak-to-trough decline of the combined account equity, including all open positions from every EA at once. A 15% drawdown in one strategy and a 15% drawdown in another can overlap and produce a 30% combined loss. Because recovery maths is non-linear, set drawdown and daily-loss limits at the portfolio level, not per strategy. The ESMA investor corner and the CFTC’s leveraged-trading resources set out the wider risk context.
How should I size positions across multiple expert advisors?
Allocate a risk budget first, then divide it. A common approach is 1% of the account per strategy, capped so the sum of all open-trade risks stays within your portfolio limit, typically 5-10%. Reduce allocations to strategies whose correlation with the rest of the portfolio has risen, and keep a reserve for new additions. Risk per trade should scale with equity, not stay fixed in lots.
Build the portfolio layer properly. See how AlgoTM EA portfolio management handles correlation analysis, aggregate exposure and portfolio-level drawdown limits across every strategy you run on MT4 and MT5.
Risk disclosure: Trading foreign exchange, commodities, CFDs and cryptocurrencies carries a high level of risk and may not be suitable for all investors. Past performance is not indicative of future results. AlgoTM provides trading tools and technology only and does not provide investment advice, portfolio management or guaranteed returns.
