Plan more than one outcome
Write down the conditions that support an idea and the evidence that would invalidate it. Examine trend structure, relevant economic events, correlated exposure and changing volatility before deciding how a scenario fits your process.
Review the limits of the evidence
A backtest describes how a rule behaved under specified historical assumptions. Examine costs, slippage, sample selection and the possibility of overfitting. Keep the distinction between a favourable result and a repeatable process clear.