How to Determine Your Daily Bias (A Repeatable Process)

Ask a struggling trader what their setup is and you'll get a clear answer. Ask which way they're leaning today and why, and you'll get a pause.

That pause is the whole problem. You can mark a perfect order block and take it in the wrong direction. You can find a textbook fair value gap on the side of the market that's about to get run. The entry model is the last decision in the sequence, and most people are treating it as the first one — because entries are what gets taught and bias is what gets waved at.

This is the process for the decision that comes first. Not a prediction of where price is going. A defensible default direction, with a level that tells you when you were wrong.

What a bias actually is

A bias is a working assumption about which side of the market you're willing to trade today, held with a defined invalidation point.

Three things follow from that definition, and all three get violated constantly.

It's not a forecast. You are not saying "EURUSD will go up today." You are saying "until price does X, I'll take longs and skip shorts." The difference matters because a forecast has no exit condition and a bias does.

It has the same shape as a trade. Idea, invalidation level, and a size of conviction. If you can't name the price at which your bias is wrong, you don't have a bias — you have an opinion, and opinions get defended rather than updated.

It's a filter, not a signal. Bias doesn't tell you when to enter. It halves the setups you're allowed to consider, which is exactly what it's for. The value isn't in the trades it finds; it's in the trades it stops you taking.

The three legitimate answers

Long bias. Short bias. No bias.

That third one is not a failure state and it isn't indecision. It's the correct output whenever the higher timeframe hasn't shown you anything — mid-range, no clean sweep, no displacement, nothing obvious being drawn towards. Some days genuinely don't have a readable direction, and the trader who forces one on those days is the trader who ends up long into a distribution because they needed a bias by 9am.

If you take one thing from this: "I don't know" is an answer you're allowed to give, and it should come up roughly as often as the other two. Most published content pretends it doesn't exist, which is how people learn that bias is something you always have rather than something you sometimes have.

The four questions

Work top-down, and answer them in order. The order matters — each answer constrains the next.

1. Where are we in the higher timeframe range?

Find the most recent meaningful swing high and low on your bias timeframe, and mark the midpoint. Price above the midpoint is in premium; below it is discount.

This single line does more work than any indicator. You want to be buying from discount and selling from premium, because that's the difference between entering where price is cheap relative to the range and chasing it after the move. If you catch yourself looking for longs in deep premium, the setup can be immaculate and the location still wrong — that's covered fully in premium and discount.

What this question gives you: a preference, not yet a bias. Discount says "longs are the trade I'd rather be taking." It doesn't say the low is in.

2. What is price being drawn towards?

Markets move between pools of resting orders. So ask: where are the obvious ones, and which side is more attractive?

Look for equal highs and equal lows, obvious swing points that everyone can see, the previous day's high and low, and the highs and lows of the current week. These are where stops sit, and stops are what price reaches for.

The read you want is comparative: which pool is closer, cleaner, and more obviously untouched? If there's a set of tidy equal highs 40 points above and nothing but messy structure 150 points below, the more likely near-term draw is up — regardless of how bearish the chart feels.

Be honest here, because this is where confirmation bias enters. If you already want to be short, every pool below looks compelling. The discipline is to mark the pools before you form a direction, and to mark both sides.

3. Has the higher timeframe actually shown its hand?

Preference and draw are context. This is the question that converts them into a bias.

You're looking for one of two things:

• A sweep followed by displacement. Price runs a pool — takes out an obvious high or low — and then moves away from it decisively, with a strong candle that leaves an imbalance behind. That's the signature that the run was the point, and the direction of the displacement is your bias. This is exactly the mechanic candle range theory is built on, and the sequence is set out in the power of three: accumulation, manipulation, distribution.
• A clean break of structure with follow-through. Not a wick through a level. A close beyond it, with intent, that then holds when price comes back to test it.

If you have neither — if price is chopping mid-range, or it swept a level and just sat there — go back to the top of this section and read the answer as no bias. A sweep without displacement is not a signal. It's price touching a level, which happens constantly and means nothing on its own.

4. What would make me wrong?

Name the price. Write it down.

For a long bias formed off a sweep of the sell-side and displacement up, the honest invalidation is usually a decisive close back below the low that was swept. If that happens, the "manipulation" you identified was just a move down, and you are on the wrong side.

The point of writing it down before the session is that it stops the goalposts moving. Bias without a written invalidation level tends to survive contact with contradictory price for far longer than it should, because each individual reason to hold on to it sounds reasonable in the moment.

Pick three timeframes and stop

The most common way this process goes wrong is timeframe sprawl — nine charts open, each saying something different, and a "bias" that's really an average of contradictions.

Three timeframes. One sets the bias, one finds the setup, one times the entry.

Style	Bias timeframe	Setup timeframe	Entry timeframe
Swing	Weekly	Daily	4H / 1H
Day trading	Daily	4H / 1H	15m / 5m
Intraday scalping	4H	1H / 15m	5m / 1m

Two rules that keep this clean:

Your bias timeframe must be meaningfully higher than your entry timeframe. If you set bias on the 15m and enter on the 5m, you don't have a bias — you have a slightly slower version of the same noise, and it'll flip three times a session.

Bias is set on the higher timeframe and not renegotiated on the lower one. The whole point of the hierarchy is that the lower timeframe is allowed to look wrong. It will look wrong. Price pulling back against your daily bias on the 5m is not new information; it's what a pullback is.

Where bias goes wrong

Anchoring. You formed a view on Monday and you're still defending it on Thursday, adding reasons as evidence accumulates against it. The written invalidation level is the antidote — it converts "am I still right?" from a judgement call into a check.

Circular bias. Using the timeframe you trade to set the bias for that timeframe. The bias is supposed to come from outside the noise you're trading. If it doesn't, it's just the chart telling you what it already did.

Confusing bias with conviction. A bias formed off a clean sweep and displacement at the edge of a weekly range is not the same object as a bias formed off "it's been going up." Both are directional. One deserves full risk and one deserves a skip. Grade yours: strong, weak, none — and let the grade decide whether you trade at all, not just which way.

Refusing to be flat. Covered above but worth repeating, because it costs the most money. There is no rule requiring a directional view every day.

Bias as a substitute for a setup. A bias is not permission to enter. It's permission to look for a specific kind of entry. You still need a location and a trigger — an order block, a fair value gap, a breaker — and the differences between them are laid out in order blocks vs fair value gaps.

A ten-minute pre-session routine

Do this before the session you trade, in this order, and stop when you hit "no bias":

1. Bias timeframe. Mark the current range high, low, and midpoint. Are we in premium or discount?
2. Mark the pools. Equal highs, equal lows, previous day's high and low, previous week's high and low. Both sides, before you have a view.
3. Look for the tell. Has a pool been swept with displacement away from it? Has structure broken and held?
4. Write the bias in one sentence, in this format: "Long while price holds above X, targeting the pool at Y." If you can't fill in both X and Y, you don't have a bias yet.
5. Drop to the setup timeframe and identify where you'd actually want to enter, in the right half of the range.
6. Note the level that ends it, and set an alert there rather than watching.

If step 3 gives you nothing, the correct output is "no bias, no trades unless the higher timeframe resolves." Write that down too — it's data, and days you correctly stood down are worth being able to count.

The sequence, once you have it

Bias is the first decision, not the only one. In order:

Bias — which way you're facing, from this process. Setup — the location and trigger that gets you in on that side. Invalidation — where the idea is wrong, which is what sets your stop. Not a fixed pip distance; the structural level that disproves the trade. Where to place your stop loss covers this per setup type, including what to do when the honest stop is too wide. Target — the pool you identified in question 2. That's the neat part: the draw on liquidity that gave you the bias is usually also the destination, which is why where to take profit starts from the same map.

When bias is done properly, the trade is largely specified before you look for an entry. When it's skipped, every one of those decisions gets made reactively with money on the line.

The whole sequence — bias, setup, entry, invalidation, target — is what the CRTLAB course teaches as one repeatable process rather than a collection of patterns, and if you're still assembling the vocabulary, ICT concepts explained for beginners is the place to start.

Test whether your bias is worth anything

Here's the part almost nobody does, and it's the only way to find out if any of this works for you.

Log your bias before the session, then log what happened. Your journal should have a bias column — direction, grade, and invalidation level — filled in before the first trade. After fifty sessions you can ask questions nobody can answer for you:

• What's your hit rate on strong-grade biases versus weak ones? If they're the same, your grading is decorative.
• Do you actually perform better on days you had a bias than days you didn't?
• On losing days, was the bias wrong — or was the bias right and the entry bad? These are completely different problems with completely different fixes, and without the column you cannot tell them apart.

That last one is the payoff. Most traders trying to fix "my strategy isn't working" are fixing the wrong half of it. How to journal your trades covers the fields that make this measurable, and a bias column is the cheapest one to add.

The bottom line

A daily bias is a working assumption with an invalidation level, not a prediction. Build it top-down: where you are in the range, where the liquidity sits, whether the higher timeframe has actually shown its hand, and what would prove you wrong. Use three timeframes, set the bias on the highest, and don't renegotiate it on the lowest.

And keep "no bias" on the table. The trader who is flat on unreadable days and full-size on clear ones will beat the trader with a view every single morning — not because their reads are better, but because they're only taking the reads worth taking.

FAQ

What is a daily bias in trading? A working assumption about which direction you're willing to trade for the session, held with a specific price level that invalidates it. It's a filter that halves the setups you'll consider, not a prediction of where price will close and not a signal to enter.

How do you determine daily bias in ICT? Top-down, in four steps: locate price within the higher timeframe range to establish premium or discount, mark the liquidity pools on both sides to identify the likely draw, check whether a pool has been swept with displacement away from it, and write down the level that would prove the read wrong. If step three shows nothing, the answer is no bias.

Which timeframe should I use for daily bias? One meaningfully higher than the one you enter on. Day traders typically set bias on the daily, find setups on the 4H or 1H, and enter on the 15m or 5m. If your bias timeframe is close to your entry timeframe, the bias will flip repeatedly during the session and isn't doing its job.

Can I trade without a daily bias? You can, but you'll take setups on both sides of a market that's only paying on one. The more useful version of this question is the reverse: you should be willing to not trade when you have no bias. Days without a readable higher timeframe signal are days to stand down, and forcing a direction on them is a reliable way to lose money.

What's the difference between bias and a trade setup? Bias is direction and context; the setup is location and trigger. Bias says "I'm only interested in longs today, above this level, aiming at that pool." The setup is the order block, fair value gap or breaker that actually gets you in. A bias without a setup isn't a trade, and a setup taken against your bias is the most common avoidable loss there is.

How often should my bias change? Only when your invalidation level is breached, or when the bias timeframe produces a new signal. If your bias changes multiple times within a session, you're reading it from a timeframe that's too low. Intra-session pullbacks against your direction are expected and are not a reason to flip.

Why does my bias keep being wrong? Two likely causes, and the journal tells you which. Either you're forming bias without a genuine tell — treating a sweep with no displacement as a signal, which is just price touching a level — or your bias is fine and your entries are poor, in which case your losses will cluster on trades that went the right way eventually. Logging bias direction and grade before the session is what separates them.

Does a daily bias work for scalping? Yes, but shift the whole ladder down: set bias on the 4H, find setups on the 1H or 15m, and enter on the 5m or 1m. The principle is unchanged — the bias timeframe must sit above the noise you're trading. Be aware that lower-timeframe biases invalidate faster and more often, so the invalidation level matters even more.