Daily bias and reference levels

What a daily bias is

A daily bias is a directional hypothesis for the coming session, formed before it starts. Not a price prediction, not a target: a direction in which you agree to look for trades, and one in which you refuse to take them.

Its purpose is not to guess. Its purpose is to reduce the number of decisions you make under pressure. A trader who wrote their bias at eight is not wondering at three whether to buy or sell, only whether their setup is there.

It is a filter, and that is all. A filter can be wrong without being useless, exactly as a stop being hit does not make risk management useless.

The bias is written, and it is written beforehand. A bias formed at midday, after seeing the morning, is not a bias, it is a comment.

⚠️ A bias is not a conviction. The more attached to it you are, the less it serves you: its value comes from being revisable on a written criterion, not from being right.

Yesterday's levels, and why those

Three prices from the previous day anchor almost every method: the high, the low and the close. They have a quality nothing else has: they are fixed, public, and identical for everyone.

That is what makes them useful, not some magical property. A level everyone can see is a level where everyone places something, entries as well as exits. The lesson on liquidity and smart money explains what piles up there.

The day's open is often added to the list, as is the midpoint of yesterday's range. The more levels you add, the less each one counts, and beyond four or five you will always find a level near any price at all.

The discipline is therefore to choose few and keep them. Three levels held for fifty sessions beat twelve levels picked according to mood.

These prices are read off the evening before, in two minutes, and written down before the open. It is the least impressive work in this course and one of the few whose benefit shows within a week.

How to build a bias, honestly

The simplest method fits in one question: relative to yesterday's levels, where did price open, and which side did it stay on through the first hour?

An open above yesterday's high that holds there gives a bullish bias. An open below the low that holds gives a bearish one. An open inside the range gives no bias at all, which is a perfectly valid answer.

That third case is the one almost nobody accepts, and it is the one that protects most. A day without a bias is a day with less risk taken, not a day wasted.

Higher timeframe market structure provides a second method, slower and steadier: as long as the series of daily lows is rising, the bias stays bullish, whatever the morning's mood.

Pick one and write it down. Both work; alternating between them depending on what suits does not, and that is what everyone does.

The first imbalance of the day

One question comes up constantly in the ICT vocabulary and deserves a direct answer: does first presented fvg require daily bias, that is, do you need a direction before taking the session's first imbalance?

The honest answer is yes, for an arithmetic reason rather than a doctrinal one. The first imbalance of a day forms in either direction roughly as often. With no directional filter you take both, and you pay the spread on the half that fails.

The bias does not make the imbalance better. It halves the number of times you take it, and it preferentially removes the cases that were running against the day's move.

That is measurable, and it is the kind of measurement that takes one evening. Take thirty first-of-session imbalances, note which ran with your written bias, and compare the two groups.

The lesson on entry models handles the rest: once the direction is decided, the question of exactly where the order goes remains.

When the bias turns out wrong

A bias must have a written breaking point, otherwise it cannot be revised and turns into a belief. That point is usually one of yesterday's levels taken back the other way.

The mechanism is the same as for a stop: you decide in advance what would change your mind, and you change it when that happens, without negotiating with yourself.

What you must not do is flip the bias in the heat of it. An invalidated bullish bias does not automatically become a bearish one: it becomes an absence of bias, and an absence of bias means taking nothing until tomorrow.

That rule alone saves more money than all the others combined, because the worst sequences never come from a wrong reading. They come from a reading flipped three times in the same day.

⚠️ Record in your trading journal the morning bias AND the moment you abandoned it. The "bias changed mid-session" column is often the one that explains the month.

Practising: thirty days, a bias written the night before

Every evening, read off the three levels from the day just gone and write one line: bullish, bearish, or none, plus the price that would invalidate it. It takes two minutes.

The following evening, add two columns: did the bias play out, and was it invalidated during the day. Do not record trades yet, we are measuring the bias on its own.

After thirty days, count. If your bias lands right more often than chance on your instrument, you have a filter. If not you have a habit, and it is better to know before building on it.

Then look at the invalidation column. If it fills on more than one day in three, your bias method is too sensitive, and you should calibrate it on slower levels, daily rather than hourly.

This record combines with the sessions one: a bias held plus an hourly window give you a two-dimensional filter, which is already more than most traders apply after three years.