Strategies
Backtesting: How to Validate Your Trading Strategy Before Risking Capital
Backtesting gives you a statistically valid test of your strategy's historical performance — separating genuine edge from wishful thinking before real money is risked.
Key rules
- Minimum 100 trades is required for statistical validity — 20-30 is not enough to measure real edge
- Define rules in objective, testable terms before looking at historical charts — avoid look-ahead bias
- Calculate expectancy: (win rate × avg win) − (loss rate × avg loss) — must be positive for any strategy to work long-term
- Measure the maximum drawdown — if you can't survive the worst historical drawdown psychologically, the strategy won't work for you
- Forward test with paper trading for 50–100 trades before risking real capital — validate that you can execute it in real time
Backtesting is the process of applying your trading rules to historical data to evaluate how the strategy would have performed. Done correctly, backtesting answers the fundamental question: "Does this strategy have a statistically valid edge, or am I imagining one?"
**Why Backtesting Matters** Every strategy looks great in hindsight when you're cherry-picking examples. Systematic backtesting forces you to test ALL instances of the setup, including the ones that failed. Only when you've seen the full distribution of outcomes — including the worst drawdown periods — can you commit to executing the strategy with confidence.
**Manual Backtesting Process** Step 1: Define your setup rules in objective, testable terms. "Bullish pin bar at a prior support level after a 3+ bar decline, with RSI below 40" — this is testable. "Price looks like it wants to bounce" — this is not.
Step 2: Scroll through historical charts and mark every instance of the setup occurring. Do not look at the right side of the chart while identifying setups — this introduces massive look-ahead bias.
Step 3: Record the outcome: entry price, stop, target, actual exit, profit/loss in R multiples.
Step 4: Analyse the statistics: win rate, average win (R), average loss (R), expectancy, maximum drawdown, longest losing streak. A strategy with 45% win rate but 2.5R average winners has positive expectancy: (0.45 × 2.5) - (0.55 × 1) = 0.575R per trade.
**Forward Testing (Paper Trading)** After backtesting, forward test in a simulated account for 50–100 trades before risking real capital. This tests whether you can identify and execute the setup in real time, not just in the clean environment of historical charts.
**Common Backtesting Mistakes** 1. **Look-ahead bias**: You knew what happened next when marking the setup. Use a chart replay tool. 2. **Overfitting**: Optimising rules until they perfectly fit historical data — they'll fail on new data. 3. **Insufficient sample size**: 20 trades tells you nothing. Need 100+ to have statistical validity. 4. **Ignoring costs**: Account for spread, commission, and slippage in every trade.