You open the chart. You see support. You switch to the 15-minute: support holds. You switch to the 5-minute: it still holds. You go long.
What you just did is not analysis. It is collecting evidence for a verdict already reached. The decision was made, and you kept changing timeframes until one agreed with you.
That is confirmation bias, and it is probably the most expensive one, because it is the only bias that looks like serious work.
What is confirmation bias?
Confirmation bias is the tendency to seek out, remember and overweight information that supports what we already believe, and to dismiss information that contradicts it.
Peter Wason isolated it in a famous 1960 experiment. Participants are given the sequence 2, 4, 6 and asked to find the rule by proposing other sequences to test it. Almost everyone proposes 8, 10, 12, then 20, 22, 24, and confidently announces "even numbers increasing by two". The real rule was simply "any three increasing numbers". Almost nobody thought to test a sequence that was supposed to fail.
That is exactly what happens when you switch timeframes until you find the one that confirms your directional bias.
What does it look like on a trading account?
Confirmation bias does not shout. It takes forms that sound reasonable:
| What you tell yourself | What is actually happening |
|---|---|
| "I'm confirming on another timeframe" | You switch timeframes until one goes your way |
| "I'll add an indicator to validate" | You stack indicators until one signals your entry |
| "The market will turn, the fundamentals are good" | You are losing and looking for a reason not to cut |
| "This trader says the same thing I do" | You scrolled until you found someone who agrees with you |
| "It's just noise" | A contrary signal showed up and you relabeled it as noise |
The common thread: in every case, the contrary information existed and was visible. You did not miss it. You set it aside.
Why is it more costly in trading than elsewhere?
In most jobs, confirmation bias wastes time. In trading, it loses money, for three reasons.
The market supplies unlimited evidence. On any instrument, there is always a timeframe, an indicator or an analyst that supports your thesis. The raw material of the bias is endless.
Feedback is noisy. A bad trade can win and a good trade can lose. The outcome therefore does not correct your belief. It often does the opposite: a bad trade that wins reinforces the bad belief.
The bias gets worse when you are losing. The further the position goes against you, the stronger the need to be right, and the harder you look for evidence. That is when your judgment degrades the most, and when it costs the most. This mechanism combines with loss aversion, which already pushes you not to cut.
Three countermeasures that hold up in real conditions
Advice like "stay objective" is useless: nobody knows they are biased while they are. What works are procedures that run before the bias kicks in.
1. Write the opposite thesis before you enter
Before every position, write one sentence describing what should happen if you are wrong, and at what level you will admit it. Not just "I have a stop": the scenario itself.
Long ES on a retest of 5,820 support. If support breaks on higher volume than the previous breakout, my read is wrong: this is not a retest, it is distribution. Exit immediately, without waiting for the stop.
It looks trivial. In practice, a good share of positions do not survive the exercise, because writing the invalidation forces you to admit you did not have one.
2. Fix your timeframes before you look at price
Decide in your plan which timeframes you use, in what order, and what for: one for context, one for entry. Any timeframe checked outside that list, after seeing price, is a search for confirmation, not analysis.
3. Review your losing trades looking for the ignored signal
This is the most unpleasant exercise and the most profitable. On each losing trade, ask one question: what contrary information was available at the time of entry? Not "what went wrong": the market did not go wrong, it did what it does. You are looking for the signal you saw and relabeled as noise.
Do it on twenty trades and a pattern appears. Almost always the same signal, almost always at the same moment.
The role of the journal
Confirmation bias has one weakness: it does not survive a timestamped written record. As long as your thesis stays in your head, it keeps rewriting itself. Afterwards, you will remember having "felt" that it was going to break. Written down before entry, with its invalidation, it becomes a fact you can no longer renegotiate.
That is the point of a journal: state the thesis before the open, then compare it with the outcome with no way to rewrite it. If you are new to this, start with the difference between a trade blotter and a trading journal.
Key takeaways
Confirmation bias is not fixed with willpower. It is worked around with procedure: an opposite thesis written before entry, timeframes fixed in the plan, and a systematic review of losing trades looking for the ignored signal.
It is one of the four biases that do the most damage to a retail account, alongside FOMO, loss aversion and overconfidence. For the wider picture, see 10 trading psychology mistakes and their fixes.
Sources: Wason, P. C. (1960), "On the failure to eliminate hypotheses in a conceptual task", Quarterly Journal of Experimental Psychology, 12(3), 129-140. Nickerson, R. S. (1998), "Confirmation Bias: A Ubiquitous Phenomenon in Many Guises", Review of General Psychology, 2(2), 175-220.
