Why the Same EA Loses on a Different Broker

"It was profitable in the backtest on broker A, but when I ran it on broker B it lost." This is a common EA story. Most people blame the EA, but the cause is usually not the EA — it is the difference between brokers. Same logic, different environment, different result. Let us break down what causes it.

Cause 1: Spread

The biggest factor is spread. For the same symbol, spread can differ several-fold between brokers.

This is lethal for thin-margin EAs. A scalper or grid EA with a small take-profit gives up a large fraction of each win to the spread. A spread just 1.5× wider can flip profit into loss. A trend follower with a large take-profit, by contrast, is relatively immune.

Cause 2: Execution and slippage

The price at which your stops and limits fill depends on the broker's execution model.

• On brokers with slow execution or heavy slippage, you fill at worse prices than assumed.
• Frequent requotes make entries disappear entirely.
• On brokers whose spread blows out during news, the entry at that moment is not worth taking.

A backtest assumes idealized execution. The further a real broker's execution is from that ideal, the wider the gap between live and backtest.

Cause 3: Symbol specification (the overlooked one)

This is the most overlooked. Even the same "gold" has different symbol specs across brokers.

• Symbol name: XM = GOLD, Exness = XAUUSDm, HFM = XAUUSD, and so on. A hard-coded EA simply won't run if the name differs.
• Contract size: the size of one lot differs between brokers.
• Tick value / tick size: the money value of one pip differs.

An EA that hard-codes these as fixed values will miscalculate lot size on another broker and take several times the intended risk. Even with a profitable backtest, an oversized lot in live trading can end the account in one move.

How to spot broker dependency

Before you buy or run, check:

1. Are there backtests on multiple brokers? A single-broker result may be optimized to that one broker.
2. Is the spread assumption stated? Distrust backtests run on unrealistically tight spreads.
3. Does the EA read symbol specs dynamically (_Symbol, dynamic tick-value calculation) rather than hard-coding them?
4. Is it thin-margin? Always demo-test a small-take-profit EA on the broker you actually use.

What makes an EA broker-robust

Structurally, the least broker-dependent EAs are:

• Large profit target (trend following, breakout): spread has relatively little impact.
• Dynamic symbol specs: works with any broker's symbol name and contract size automatically.
• Built-in spread ceiling: does not enter the moment the spread widens.

Every one of our EAs uses _Symbol instead of a hard-coded name, computes tick value and tick size dynamically, and includes a spread-ceiling filter — so it runs with the same logic on XM, Exness, HFM and any other MT5 broker.

Check it yourself

We publish a free indicator that shows your spread right now and how far it has widened relative to its own average. Watch when and how much your broker's spread widens, and you can judge whether that broker suits EAs at all.

Spread Guard (free indicator)

The real data for our large-target, broker-robust trend EAs is on the full verification page. We do not put thin-margin scalpers or grids at the front of the lineup precisely to avoid broker dependency.

Summary

• The same EA gives different results on different brokers (the cause is the environment, not the EA).
• The three big factors: spread, execution/slippage, symbol specification.
• Thin-margin EAs are the most broker-dependent; large-target EAs are the most robust.
• What to check: multi-broker testing, spread assumptions, dynamic symbol specs.
• Always demo-test a thin-margin EA on your own broker.

Related: How spread and slippage affect EA returns · Diversifying an EA across multiple brokers · Why a 90% win rate is a warning sign