Candle backtests have to guess the path inside every bar

[Amanhasnoname]

A one-minute bar only records open, high, low, and close. The order in which price hit the high and the low inside that minute is unknown, and the tester has to make an assumption.

That assumption decides outcomes for scalping. If a trade has a 3 pip stop and a 3 pip target, both inside the same bar's range, the result depends on which was touched first. Different testers resolve it differently, and the difference can swing a strategy from profitable to losing.

Bars with wide ranges are the worst offenders. They contain several possible paths, and the tester picks one, often optimistically.

Testing at tick level removes the guesswork, but the data needs to be good. Testing at one-second or tick granularity on real data shows how many of the apparent wins were ambiguous.

A quick check is counting the trades where both stop and target fell within one bar. If those make up a large share, the result is not trustworthy.

What share of backtest trades had stop and target inside the same bar, and how did tick testing change the results?