You have just opened a long position. Over the next few hours, almost without noticing, your behaviour changes: you read the articles that speak well of the asset carefully and skim the critical ones; in the forums you look for the comments of people who are bullish like you and dismiss the bearish ones as chronic pessimists; on the chart you spot every signal that supports your thesis and find a reason to file away the ones that contradict it. You are no longer analysing the market. You are defending an idea.
That is confirmation bias: the systematic tendency to seek out, interpret, remember and give weight to information that confirms what we already believe, while ignoring or discounting whatever contradicts it. In everyday life it produces stubborn opinions and pointless arguments. In trading it produces something more concrete: losses. This guide covers what the bias actually is, the experiment that exposed it, the specific ways it sabotages traders, and the procedures that work against an enemy you will never catch in the act on yourself.
The experiment that exposed the bias
The systematic study of the phenomenon owes a great deal to the English cognitive psychologist Peter Wason, whose experiments in the 1960s became classics. The best known is the 2-4-6 task, published in 1960. Participants were told that the sequence "2, 4, 6" followed a rule, and that their job was to discover it by proposing other three-number sequences. For each one, the experimenter would say whether it fitted the rule or not.
The real rule was trivial: any three numbers in ascending order. Most participants immediately formed a narrower hypothesis, such as "even numbers increasing by two", and then proposed almost exclusively sequences that would confirm it: 8-10-12, 20-22-24, and so on. Receiving a steady stream of yes answers, they became convinced they had solved it, and confidently announced the wrong rule. Very few did the logically powerful thing, which was to propose sequences that could have destroyed their own hypothesis, such as 1-2-3 or 10-9-8.
The deeper lesson is that confirmation bias is neither stupidity nor laziness. It is the default way the human mind tests a hypothesis. Looking for confirmation is cognitively cheap and emotionally pleasant; looking for refutation is hard work and mildly painful, because every refutation is a small wound to the ego. The scientific method, with its insistence on falsification, was built precisely as an artificial correction to this natural tendency. Good trading needs the same correction.
Three places where the bias operates
Confirmation bias does not act at a single point in the thinking process. It acts at three, and all three show up on a trading desk.
The first is selective search. We actively choose the sources that agree with us. The trader who is long an asset follows the bullish analysts, joins the groups where everybody holds the same position, and arranges their reading so that concordant opinions are what they mostly meet. Social platforms amplify this by showing us more of what we engage with, but the initial choice is ours.
The second is distorted interpretation. Faced with the same ambiguous information, we read it in whichever way favours our thesis. A mixed quarterly report becomes "excellent if you look at the right details" for the trader who is long and "the beginning of the end" for the one who is short. A classic 1979 study showed that people holding opposite views, when shown the same set of mixed evidence, both walked away more convinced of their original position: each side judged the favourable evidence solid and the unfavourable evidence flawed. Evidence does not persuade. It gets recruited.
The third is selective memory. We remember the times we were right far better than the times we were wrong. The trader vividly recalls the call that worked and quietly forgets the three that did not. It is also the mechanism that inflates the reputation of market gurus: the accurate forecasts are celebrated and repeated, the failed ones slide into oblivion. Put the three together and you get a self-feeding view of the world: we look for confirmation, we read everything as confirmation, and we remember only the confirmations.
What it costs a trader
The first and most expensive damage is marriage to a losing position. When a trade goes against them, the affected trader does not re-examine the thesis; they go looking for ammunition to defend it. They reread the analysis that convinced them, find fresh bullish articles, and reinterpret the drawdown as an opportunity to average down. Every contrary piece of information gets neutralised as manipulation or as the market not understanding yet. The losing position is held, or increased, well past any reasonable point. Confirmation bias is the perfect accomplice of loss aversion: one supplies the emotional reason not to close, the other supplies the intellectual justification. Both belong to the wider family covered in our guide to cognitive biases in trading, and both are examined from the risk side in cognitive risk management.
The second damage happens before the entry. A trader who falls in love with an idea stops doing analysis and starts doing advocacy: building the case in favour of the trade rather than hunting for the reasons it might be wrong. On the chart they see the patterns that support the thesis and miss the ones that do not; among the indicators they select the one giving the desired signal and ignore the rest. The result is a trading plan that looks extremely solid, because it was assembled to look solid rather than to be tested.
The third damage is a fossilised strategy. The bias does not only protect individual trades, it protects entire belief systems. A trader convinced their method works remembers the winners as proof of the method and files the losers as exceptions or as the market misbehaving. Without an honest look at the data, a mediocre system, or one that no longer suits current conditions, can survive for years in its author's mind while it drains the account a little at a time.
Echo chambers turn a private bias into a group one
Confirmation bias is as old as the human mind, but the environment traders operate in today amplifies it as never before. In a few clicks anyone can build a perfectly homogeneous information diet: only concordant voices, only analysis pointing the desired way, only fellow holders. Communities built around a single asset drift easily into echo chambers where dissent is treason, the sceptic is an enemy, and every piece of news is instantly reframed in favour of the shared thesis.
These environments weld individual confirmation bias to herd behaviour, and the combination is dangerous: belonging to the group makes changing your mind psychologically more expensive, because leaving a position stops being a private admission of error and starts feeling like abandoning the team. Several of the best-known retail disasters of recent years grew in exactly this soil.
The countermeasure is not isolation. It is deliberately building an information diet tilted against your own instinct: following serious analysts who disagree with you, reading the opposing case first, spending time where dissent is normal. It also helps to look at hard positioning data rather than opinions. Our retail sentiment board shows how the crowd is positioned across 29 markets, updated every three hours, which is a useful reality check when your feed has convinced you that everyone sensible agrees with you. If all your sources agree with you, you do not have sources. You have mirrors.
Procedures that beat it
The first defence, true of every bias, is knowing it exists and that it applies to you as well. Awareness alone is not enough, though, and there is even a documented trap called the bias blind spot: we spot biases easily in other people and deny them in ourselves. What works is procedure, meaning structured habits that force the mind to do what it will not do spontaneously, which is to go looking for the refutation.
The most powerful technique is to ask, before every trade: what would have to happen to prove this idea wrong? That is Wason's question applied to the market. If you cannot imagine any condition that would invalidate your thesis, you do not have a trading thesis, you have a belief. Writing the invalidation point down in advance, as a specific price level, data release or event, and tying the stop loss to it turns falsification from an abstract exercise into concrete protection for the account. When that level is reached the discussion is over: the thesis was wrong, you are out, regardless of whatever new justifications your mind has produced in the meantime.
This is one reason a written plan beats an improvised one. Our daily AI forecast publishes each plan before the session opens, with entry, stop and targets fixed in advance, and then checks it against the real candles of that day and publishes the outcome, win or lose. Whatever you think of any individual plan, the format itself is the discipline this bias requires: the level is decided when you have nothing at stake, and the result is scored afterwards whether it flatters you or not.
A second technique is the formal devil's advocate: before entering, write the best possible case against your own trade, the way an intelligent trader holding the opposite position would write it. Not two token lines, but the strongest version of the contrary thesis. If your idea still stands afterwards, it is a stronger idea. If it wobbles, you have just avoided a loss.
A third is an honest trading journal. Recording the thesis, the invalidation point, the outcome and the lesson for every trade forces a confrontation with the real numbers of your performance instead of the edited version selective memory keeps. The journal has no bias; it is the antidote to the narrative.
Finally, a hygiene rule for open positions: be suspicious of yourself whenever you notice that you are "doing research" on a trade that is already open and already losing. Nine times out of ten you are not analysing, you are looking for comfort. The real analysis happens beforehand, with a cold head and a written invalidation point. Afterwards the job is not to find new reasons to stay, but to check whether the original reason for entering still holds.
The same trap outside the markets
It is worth widening the lens, because recognising the bias away from the screen makes it easier to catch on the screen. Confirmation bias is the engine of polarisation: on any contested subject, people with opposite views consume different sources, read the same facts in opposite ways, and each come away more certain. It is also a tool for anyone selling something. Guru forecasts, horoscopes and certain marketing narratives work because the audience remembers the hits and forgets the misses, doing the confirmation work on the seller's behalf.
For an investor, the most treacherous ground is the long-horizon story: this asset can only go up, this sector is the future, that country is in terminal decline. Such theses are vague enough to recruit any headline as confirmation and identity-forming enough to make questioning them painful. The remedy is the same as for a single trade, scaled up: ask periodically what concrete evidence would change your mind, then go looking for it. Anyone who cannot answer is not investing in a thesis. They are campaigning for one.
An enemy that works from the inside
Confirmation bias is among the most dangerous biases for a trader precisely because it never announces itself. From the inside, looking for confirmation feels exactly like staying informed, going deeper, doing your homework. Wason showed more than sixty years ago that this is the default setting of the human mind: we test ideas by trying to confirm them, not to break them. On markets, where wrong ideas cost real money, that default setting is a luxury.
The defence is not hoping to become immune, because nobody is. It is building procedures that do the work instinct refuses: the falsification question before every trade, the invalidation point written down and wired to the stop, the formal devil's advocate, the journal kept honestly, and an information diet that deliberately includes the voices you disagree with. These are uncomfortable habits, because every refutation you go looking for is a small wound to the ego. But in trading the choice is exactly that: small voluntary wounds to the ego today, or large involuntary ones to the account tomorrow. The traders who last have almost always chosen the first.
