How to Trade Multiple Timeframes
To trade multiple timeframes, use exactly two: a higher one that decides which direction you may trade and a lower one that decides when. Write down what happens when they disagree, because the answer is almost always to do nothing.
Multi-timeframe analysis means using a higher timeframe to decide direction and a lower one to decide timing. Two charts, two jobs. The discipline is entirely in the number of charts and in what you do when they disagree.
Before you start
Exactly two timeframes, each with a written job. One says which way you are allowed to trade; the other says when. Written down.
A rule for what happens when they disagree, decided before they do. The answer is almost always no trade, and deciding that afterwards means deciding by preference.
A ratio between them wide enough that they are answering different questions. Four to six times is the usual range; adjacent timeframes mostly duplicate each other.
The steps
1. Choose the higher timeframe from your holding period
If you hold for days, the daily is your context chart. If you hold for hours, the hourly is. The holding period decides, not habit.
2. Take the lower timeframe from a wide ratio
Daily and hourly, or hourly and ten-minute. Adjacent timeframes show almost the same information twice, which is the same problem as having three.
3. Write the job of each chart
“The daily says up or down. The hourly says when.” Two sentences. Without them both charts end up being consulted for both questions.
4. Read the higher timeframe first, and only once
Establish the permitted direction at the start of the session and do not revisit it after seeing the lower chart. Re-reading it after a setup appears is how the context gets adjusted to fit.
5. Take the entry from the lower timeframe only
The setup, the level and the stop all come from the timing chart. The context chart has already done its job and contributes nothing further.
6. Do nothing when they disagree
A lower-timeframe setup against the higher-timeframe direction is not a trade. That is the rule doing the only thing it exists to do.
7. Resist adding a third chart
A third timeframe feels like more information and functions as a tiebreaker you consult only when you dislike the answer. Two charts can produce a no; three effectively cannot.
How to tell it worked
Exactly 2 timeframes are open, with a written job each.
The higher one was read 1 time per session, before the lower one.
0 trades were taken against the higher timeframe’s direction in the last 30 days.
And the ratio between them is at least 4 times.
Why three charts stop working
Because at any moment they will not all agree. With three, some subset always supports any given trade, so the framework can produce a yes for everything and a no for nothing.
And because the lowest chart is usually noise. On this site’s shared series direction runs average 2.01 bars, so a very fine timeframe is showing you two-bar movements that reverse constantly.
What the higher timeframe is actually for
Removing half your trades. That is its entire function: it declines every setup pointing the wrong way, without any judgement about the setup’s quality.
Not improving the entries it permits. A setup is as good or as bad as it was; the context chart adds nothing to it beyond permission.
Which means the test of whether it is working is how often it says no. A higher timeframe that has never declined a trade is not filtering — it is decoration on a chart you were going to trade anyway.
Choosing the pair
Start from how long you actually hold, not from how long you would like to. A position held for three days is a daily-chart trade whatever the entry looked like on a five-minute chart.
Then divide by four to six for the timing chart. Daily and hourly. Hourly and ten-minute. Weekly and daily. Each pair shows genuinely different information rather than the same swings at two resolutions.
Check that the higher chart produces a small number of readings. If your context timeframe changes direction twice a week, it is not supplying context — it is a second timing chart, and the framework has collapsed into one.
And keep the pair fixed. Switching to a lower context chart after being filtered out of a trade you wanted is the most common way this rule quietly stops existing, and it never feels like abandoning the system at the time.
Reading them in the right order
Context first, always, and before the market opens where possible. The direction for the session is established when there is no setup in front of you to influence it.
Then the timing chart, repeatedly, throughout the session. That is the one you watch; the other has already spoken.
And if the context genuinely changes, it changes at the next session boundary rather than in the middle of a decision. A daily direction that gets revised at eleven in the morning was revised by something on the hourly, which is exactly backwards.
The original data
Of the 24,971 unique videos in research/search-study-corpus.jsonl, 2 mention multiple timeframes in
the title, at a median of 92,559 views — one of the largest audiences per video anywhere in this
corpus. Timeframe selection appears in a handful more. The counts come from site/corpus_count.py.
2 videos at a 92,559 median. Almost no coverage and an enormous audience for what exists — the clearest single demand signal in this data, on a technique that takes two charts and two sentences to describe.
The answer to the question on that chart is that your rule already answered it. The higher timeframe decides direction — and making an exception because the lower-timeframe setup looks excellent is removing the filter on exactly the trades it was built to decline.
When it fails
The failure is the third chart added for confirmation, and it removes the framework’s ability to say no. Two timeframes disagree, which under the rule means no trade. A third is opened to break the tie — reasonably, it feels like more information. From then on, whichever two agree with the intended trade become the ones consulted. The system has not been abandoned; it has been converted into a search for agreement, and agreement is always available somewhere.
The second failure is adjacent timeframes. They show the same thing twice.
A third is re-reading the context chart after a setup. It gets adjusted to fit.
A fourth is taking entries from the higher chart. Its job was direction.
A fifth is no written conflict rule. The conflict then resolves by preference.
And a sixth is a filter that has never declined a trade. It is not filtering.
Related
Multi-timeframe analysis covers the framework in general. Timeframes explains what each one shows and hides. And top-down analysis is the broader version of the same idea.
Three timeframes was where I went wrong for a long time. With three charts open I could always find one supporting the trade I wanted, and I genuinely believed I was being thorough. Two charts with defined jobs can tell me no, and that is the entire difference.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.