Backtesting a trading strategy means applying fixed entry, exit, and risk rules to historical price data to see how the method would have behaved. Done well, it is a filter that kills weak ideas before they cost live capital. Done badly, it is a story generator that produces a perfect equity curve and a live account that looks nothing like it.
The goal is not to prove you are a genius. The goal is to discover whether the rules are specific enough to test — and robust enough to survive a different stretch of market.
Write Rules a Stranger Could Follow
If two people cannot mark the same entries on the same GBP/USD or FTSE chart, you do not have a strategy. You have vibes. Backtesting requires definitions: which timeframe, what triggers a long or short, where the stop sits, how the target or time exit works, and what happens around major news if you care.
Ambiguous language — “enter on strength”, “cut when it looks wrong” — cannot be tested. Translate feelings into levels, candle closes, or indicator thresholds you can apply the same way every time.
Simple Forward-Looking Logic Beats Kitchen-Sink Optimisation
Start with one idea and one market. Add filters only when you can explain why they belong. Every extra condition can improve the past by accident. A curve that only works when five indicators align in a rare combination is usually fitting noise.
Use a period long enough to include different regimes: quiet weeks, volatile London opens, and stretches where your setup barely appears. A backtest that only covers a one-way trend will flatter trend-following rules and mislead you for the next choppy month.
Sample Size Traps
Twenty trades prove almost nothing. Winning eight of ten in a short window can be luck. You want enough trades that a few outliers do not decide the story — and you want to see maximum drawdown, not only average win.
Also separate in-sample from out-of-sample. Build or tweak rules on one date range. Then lock the rules and run them on a later range you did not use for design. If performance collapses on the holdout, you fitted the first window.
A free traders assessment can help you check whether your live journal already shows the same setups you think your backtest celebrates — or a different, messier behaviour.
Costs, Slippage and the Spread
Retail backtests that ignore spread, commission, and slippage invent edge. On FX and index CFDs, the round-turn cost matters, especially for short-hold ideas. Assume fills that are slightly worse than the ideal signal price. If the edge disappears under modest friction, it was never an edge for your account size.
News windows matter too. A rule that enters at the exact print of UK data will look clean in a spreadsheet and ugly in live execution.
Why Curve-Fit Equity Curves Fail Live
Optimisers love parameters that dodge historical losses. Markets do not owe you the same dodge next month. Overfitted systems often show tiny stop distances, perfect indicator periods, and almost no losing streaks — until live trading begins.
Prefer rules you understand economically or behaviourally: for example, fade extremes only in a defined range regime, or follow breakouts only with a clear invalidation. Understanding is not proof. It is a defence against worshipping a line on a chart.
Forward Test Before You Scale
After a respectable historical pass, trade the rules on demo or tiny live size for enough sessions that boredom sets in. Compare live results to backtest expectations on win rate, average R, and rule breaches. If you cannot follow the rules live, the backtest is academic.
Before you automate or increase size, a free traders assessment is a useful pause to see whether process gaps — not “the market changed” — explain the gap between test and reality.
Conclusion
To backtest a trading strategy, define rules a stranger could apply, include costs, demand enough trades, and validate on data you did not use to invent the method. Perfect historical curves are a warning light, not a certificate. Live discipline still decides whether any tested edge shows up in pounds.
Samuel and Co Trading treats backtesting as homework before risk, not as proof of future profit. Test simply. Forward-check honestly. Scale only what you can execute without rewriting the rules mid-trade.
