Most new traders build a strategy from a handful of screenshots that looked good, then trade it live with real money. That's not a strategy — it's a guess wearing a costume. Backtesting is the process of testing your rules against historical data to find out, before you risk anything, whether your idea actually has an edge.
What Backtesting Actually Proves
A backtest doesn't guarantee future results. What it does is answer a narrower, more useful question: does this specific rule set, applied mechanically over hundreds of historical setups, produce a positive expectancy? If the answer is no, you've just saved yourself months of losses.
The Three Jobs of a Backtest
- 1.Validate — does the strategy have a measurable statistical edge?
- 2.Calibrate — what stop, target, and position size fit the strategy's real behavior?
- 3.Condition — build the psychological confidence to follow the plan under pressure.
Traders often backtest to feel confident. The real goal is truth. If 200 historical trades say your strategy loses money, that's the value of the exercise — it stopped you before live capital did.
Why Skipping It Is So Common — and So Costly
Backtesting is tedious. Scrolling through years of charts, logging every entry, and tallying results by hand is boring compared to the dopamine hit of placing a live trade. But the traders who survive years in this business are almost universally the ones who did the boring work first.
- ▸Without a backtest, every losing streak feels like proof the strategy is broken.
- ▸With a backtest, you know what a normal losing streak looks like — and can hold the line.
- ▸Without data, position sizing is a guess. With data, it's a calculation.
What a Good Backtest Requires
- ▸A fully written, mechanical rule set — entry, stop, target, filters.
- ▸Clean historical data across a range of market conditions (trend, range, high/low volatility).
- ▸A large enough sample size to be statistically meaningful (see the sample size article).
- ▸Honest, unbiased execution — no cherry-picking only the winning setups.
"Buy when it looks bullish" cannot be backtested honestly because you'll unconsciously pick winners in hindsight. Every rule needs to be objective enough that two different people would take the same trades.
Expectancy: The Number That Matters
The single most important output of a backtest is expectancy — the average amount you can expect to win or lose per trade, expressed in R (multiples of your risk).
Expectancy (R) = (Win% × Avg Win R) − (Loss% × Avg Loss R)A strategy with a 40% win rate and an average winner of 2.5R against an average loser of 1R still has positive expectancy: (0.40 × 2.5) − (0.60 × 1) = 1.0 − 0.6 = +0.4R per trade. Backtesting is how you discover numbers like this instead of guessing.
The Bottom Line
Backtesting doesn't remove risk from trading — nothing does. What it removes is the risk of trading a strategy that was never going to work in the first place. That single filter eliminates the majority of ways new traders blow accounts.