MetaTrader 5’s Strategy Tester is one of the most powerful tools available to a retail trader — and one of the easiest to fool yourself with. A backtest is a hypothesis about the past, and with enough tweaking almost any strategy can be made to look brilliant on data it has already seen. Testing honestly is a discipline in itself.
Separate the data you tune from the data you trust
The single most important habit is out-of-sample testing. Optimise your parameters on one slice of history, then evaluate — without further tweaking — on a slice the strategy has never touched. If performance collapses out of sample, you did not find an edge; you found a curve fit. The gap between the two is your honesty check.
Model the costs that erode real returns
A backtest that ignores spread, commission, swap and slippage is fiction. On MT5, use "Every tick based on real ticks" quality where you can, set realistic spreads for the session and instrument, and account for the fact that fills at scale are worse than the mid-price you see on a chart. Frictionless results do not survive contact with a live account.
Beware the flattering traps
Overfitting is the obvious one — dozens of parameters tuned until the equity curve is a straight line. But watch also for look-ahead bias (using data that would not have been available at the time), survivorship in your instrument list, and sample sizes too small to mean anything. A strategy with forty trades tells you almost nothing.
Judge robustness, not perfection
A strategy that holds up decently across many parameter settings, instruments and market regimes is worth more than one that is spectacular under a single fragile configuration. Stress it: shift the parameters slightly, run it on a different period, add cost. What survives that scrutiny is closer to real.
None of this guarantees future performance — no test can. This is educational content, not financial advice; treat every backtest, including anything you run for Chart Radar, as evidence to be questioned rather than proof.