Which timeframe to choose

Updated 2026-09-29 Читать по-русски

A timeframe changes not the precision of the answer but the question itself. The same shape on an hourly and on a daily chart describes events of different magnitude, and their continuations differ in magnitude too.

What changes with the timeframe

Length in wall-clock time. The same number of candles is two days on an hourly chart and several months on a daily one.

Depth of history. Lower timeframes hold more candles, so more windows to compare against: more analogues are found. Higher ones have shorter histories — decades become a few thousand candles, and there may be no matches at all.

The nature of the noise. An hourly chart contains intraday swings that do not exist as separate events on a daily one. That is not "more detail" but a different subject.

How to choose

From the horizon of your decision, not from habit.

If you care about the next few hours, a higher timeframe answers a different question. If you care about weeks ahead, a lower one drowns in noise and finds hundreds of "similar" stretches with no bearing on anything.

A sensible guide: the scale of the sample should be comparable to the scale of the question.

Why more data is not better

The temptation to drop to a lower timeframe is understandable: there are always more matches down there. But the abundance does not come from the situation becoming clearer — it comes from short stretches being more numerous in any history and more often alike.

It is the same trap as with too short a sample: many matches are a sign of a weak criterion, not of solid statistics.

Results cannot be compared across timeframes

Outcome shares on an hourly and on a daily chart are different quantities. They have different horizons, different sample composition and different event magnitudes.

Matching directions on two timeframes confirm nothing either: it is the same series cut differently, and the two are not independent observations.

A common mistake

Cycling through timeframes and stopping at the one where the picture looks good. That is not choosing a tool but choosing an answer — and almost any view can be confirmed if you keep changing the scale.

Related questions

In depth

See it on a live chart

The same markup on a prepared example: the chart, the objects it found and what followed.

Open the example