Tick data versus candles: what the backtest never saw
Posted: Sat Oct 31, 2026 8:52 pm
A one-minute candle compresses hundreds of ticks into four numbers: open, high, low, close. Everything inside that minute disappears, including the order of the high and the low.
That order matters enormously for scalpers. If a trade has a 3 pip target and a 3 pip stop, and the candle shows both levels were touched, the result depends on which came first. Candle data cannot say.
Backtesting engines handle it with assumptions. Some assume the stop hits first, others the target, others interpolate. The results differ widely, and the optimistic ones inflate performance.
Tick data preserves the sequence, so the test can follow the actual path. It also includes the spread changes at each moment, which candle data lacks.
For strategies with targets of a few pips, candle-only backtests are not reliable. They can be a rough screening tool, but the conclusion should come from tick-level tests.
What kind of data goes into backtests, and has switching to tick data ever changed a verdict?
That order matters enormously for scalpers. If a trade has a 3 pip target and a 3 pip stop, and the candle shows both levels were touched, the result depends on which came first. Candle data cannot say.
Backtesting engines handle it with assumptions. Some assume the stop hits first, others the target, others interpolate. The results differ widely, and the optimistic ones inflate performance.
Tick data preserves the sequence, so the test can follow the actual path. It also includes the spread changes at each moment, which candle data lacks.
For strategies with targets of a few pips, candle-only backtests are not reliable. They can be a rough screening tool, but the conclusion should come from tick-level tests.
What kind of data goes into backtests, and has switching to tick data ever changed a verdict?