A backtest is a simulation of how a strategy would have done if you'd started trading it some time ago. It's hypothetical, and it guarantees nothing about the future. But with a large enough sample it's still the best tool we have for deciding whether a strategy has an edge worth risking money on. Read properly, that is. Most people only look at one number.
Here's how I read one.
1. **Win rate**. The odds of a trade making money. 6 winners out of 10 is a 60% win rate. Useful, but meaningless on its own, and you'll see why in point 3.
2. **Expected value per trade**. The average profit or loss across all trades, losers included. You want this positive, and not barely positive. It's what you're being paid for taking the risk.
3. **Risk-reward, the R**. The ratio between your target and your stop loss. This is the number everyone skips, and it changes everything. A 50% win rate at 1:1 R makes you exactly nothing over the long run. The same 50% win rate at 1:2 R makes you money. Win rate and R only mean something together.
4. **Total trades in the test**. Small samples lie. A 20-trade backtest can look brilliant on pure luck. A few hundred trades starts to mean something. Below that, the strategy is statistically not significant, however good the curve looks.
5. **Evaluation window**. How long the test ran. For daily-timeframe strategies I want 10+ years, so the strategy has lived through more than one kind of market. A strategy tested only on a bull run hasn't really been tested.
6. **Sharpe ratio**. For intraday strategies you can mostly ignore it. Happy to go deeper on this if anyone's interested.
And the practical part, actually placing orders from a backtest: the test already defines your entry, your stop and your target. Your only job at order time is to place all three together and size the position so the loss at the stop is a fixed fraction of your capital. The moment you override those numbers mid-trade, you're no longer trading the strategy you tested.
Curious what others use as a minimum trade count before trusting a backtest.