Backtesting & forward testing

A backtest runs a trading strategy over historical prices to see how it would have behaved. It is the cheapest way to reject a bad idea — and the easiest way to fool yourself about a mediocre one.

Backtesting in MetaTrader

The MetaTrader Strategy Tester replays history bar by bar or tick by tick. Key settings that change results dramatically:

• Modelling quality — “every tick based on real ticks” is far more realistic than “open prices only”, especially for scalpers and EAs with tight stops.
• Spread — a fixed low spread flatters results. Use a realistic or variable spread.
• Commission and swap — must be included; holding positions overnight has a cost or credit.
• Date range — short ranges hide bad regimes. Include calm and volatile years.

In-sample, out-of-sample, walk-forward

1. In-sample: the period used to design and tune the strategy.
2. Out-of-sample: a later period kept untouched until the design is frozen. One look only — reusing it turns it into in-sample data.
3. Walk-forward: repeatedly optimise on one window and test on the next, rolling forward through history. It better imitates how a strategy is actually maintained.

Forward testing

After a backtest, a strategy is usually run on a demo account or a very small live account for months. Forward testing reveals things history cannot: real execution, requotes, slippage, server outages and broker-specific conditions. Demo fills are often better than live fills, so even a good demo record is not the end of the story.

Classic backtesting mistakes

• Look-ahead bias — using information that would not have been available at the time (for example the close of the current bar).
• Survivorship and selection — showing the one configuration that worked out of hundreds tested.
• Curve fitting — too many parameters tuned to one period.
• Ignoring costs — the most common reason a backtest edge disappears live.
• Too few trades — 40 trades cannot tell skill from luck.