MT5 EA generator: from rules to expert advisor

What an EA generator should — and should not — do

A good generator converts a validated set of rules into executable MQL5 code without losing the meaning of those rules. It should preserve your entry conditions, exit conditions, stop placement and position sizing exactly as tested, and it should make the cost assumptions visible.

It should not silently change the strategy, hide the generated code behind a compiled .ex5, or claim results that the backtest never produced. If the generator cannot show you the source, you are not buying automation — you are buying a black box.

From strategy rules to MQL5 code

The translation pipeline has four predictable parts, and each one is a place where edges leak.

• Signal mapping: your backtest's entry logic (indicators, price conditions, time filters) becomes OnTick or OnBar evaluation code. The trigger must fire on the same bar the backtest assumed — one-bar offsets here silently change results.
• Risk mapping: stop loss, take profit and position sizing rules become trade management code. A 1% risk rule in the backtest must mean the same 1% on a live account.
• Cost mapping: spread, commission and slippage assumptions from the backtest are baked into the EA's settings, so the live version trades in the same economic world as the test.
• Edge cases: weekends, gaps, broker-specific fills and partial closes. A robust EA handles them without violating the original rules.

Cost parity: why spread and slippage must match

Most retail strategies live or die on costs. A backtest that assumed a 0.2 pip spread on a pair that actually trades at 1.0 pip will look profitable in the report and lose money in reality. The EA must carry the same cost model as the backtest, or the 'parity' between them is fiction.

Before running a generated EA live, compare three numbers: the average spread your broker actually shows, the slippage on a demo account, and the cost assumptions in the backtest. If any differs materially, re-run the validation with the real numbers.

The paper trading bridge

Backtest and live trading differ in ways no backtest can fully model: broker data feed, order routing, spread widening under news, and execution timing. Paper trading is the bridge that exposes these differences at zero cost.

Run the generated EA on a demo account for 1-3 months. Track the key metrics from the backtest — win rate, profit factor, drawdown — and compare. Small drift is normal; large drift means something in the translation or the cost model is wrong. Fix it before a single real dollar is attached.

Red flags in EA generators

• Only delivers a compiled .ex5 with no source you can audit.
• Shows a perfect backtest but refuses to expose walk-forward or out-of-sample results.
• The EA's cost settings cannot be configured to match your broker.
• Promises guaranteed returns or shows only in-sample curves.
• No paper trading stage in the workflow, or 'paper' that trades in the backtest engine rather than a real broker demo feed.