How trading algorithms are tested

Modelling modes in MT5

• "Every tick based on real ticks" — the broker's stored real ticks with the real spread are used; the most accurate mode.
• "Every tick" — ticks are generated from minute bars; less accurate, especially in sharp moves.
• "1 minute OHLC" and "Open prices only" — fast modes for a rough check; unsuitable for the final evaluation of grids and close-level strategies.

Backtest and forward test

A backtest checks the strategy on history. If parameters were tuned on the same history the result is inflated (overfitting). A forward test checks it on a period not used for tuning (the MT5 tester can set aside a forward part), then on a demo or a small live account.

What to look at in the report

• Maximum equity drawdown (not only balance — for grids they differ a lot).
• Minimum Margin Level and whether there were Stop Out episodes.
• Number of trades and average profit per trade relative to costs.
• Robustness: how the result changes with small changes in parameters and period.

Tester limitations

• History does not repeat: future moves can be stronger than past ones.
• The tester does not reproduce all real execution behaviour (queues, delays, rejections).
• Future swap and spread may differ.

The FOX rule

FOX publishes strategy-tester results only as test results, with the mode and period stated — never as client returns or a promise of profit.