Multi-Timeframe Confirmation in a Trading Bot: The Higher Timeframe as a Filter
One of the most common extensions to an existing strategy is a filter that takes the higher timeframe into account: only allow entries when the larger timeframe points the same way. The principle behind it fits in one sentence - the higher timeframe decides what gets traded, the lower one decides when.
The basic principle
- The higher timeframe sets the permitted direction, via a trend filter or a structural condition
- The lower timeframe provides the actual entry trigger
- Signals against the higher-timeframe direction never get executed in the first place
Common timeframe combinations
- Swing-oriented: 4 hours for direction, 1 hour for the setup, 15 minutes for the entry
- Day trading: 1 hour for direction, 15 minutes for the setup, 5 minutes for the entry
- Short-term: 15 minutes for direction, 5 minutes for the setup, 1 minute for the entry
The technical pitfall: forming or closed bar?
The most important implementation detail is almost always overlooked in a first draft: does the bot read the currently forming higher-timeframe bar, or only the closed one? If the forming bar is evaluated, the filter state can flip back and forth within the same signal window - meaning identical situations sometimes allow an entry and sometimes don't. For traceable, reproducible results, this decision has to be made deliberately and written into the rule set explicitly.
Extending an existing strategy with a timeframe filter?
Describe your rule set - the feasibility check and quote are free and non-binding.
Request a projectWhat the filter costs you
A timeframe filter doesn't only remove losing trades - it inevitably removes part of the winners too, since it simply reduces the number of permitted signals. Whether that trade-off pays off can't be answered generally, and certainly shouldn't be claimed upfront: it depends on the specific rule set and has to be compared in testing, with and without the filter, rather than assumed to be an improvement.
The general limits of automated strategies are covered in more detail in our Risk & Liability Notice.
Frequently Asked Questions
How many timeframes make sense?
Two to three is common in practice - one for direction, optionally one for the setup, and one for the entry. More levels mainly increase the number of conditions that must be true at once, and with it how rare the signals become.
Should the filter evaluate the forming or the closed higher-timeframe bar?
Both are workable, but it has to be decided deliberately. The closed bar gives stable, reproducible results but reacts later. The forming bar reacts earlier but can still change its state within the same signal window.
Can an existing bot be extended with a multi-timeframe filter afterward?
Generally yes - it's one of the most common retrofits, as long as the filter condition is clearly defined.
Does a higher-timeframe filter automatically improve results?
No. It removes signals against the higher-timeframe direction, but that removes winners as well. Whether the overall ratio improves has to be tested for the specific rule set.
A timeframe filter is quick to describe and still detail-sensitive to build - above all around the question of which higher-timeframe state applies at the moment the signal is checked. Those details decide whether a strategy behaves traceably later on.
Have an idea for ATAS?
Tell us about your project - you'll get a free, non-binding quote within 24-48 hours.
Request a Free Quote