In 1979, Daniel Kahneman and Amos Tversky published a paper in Econometrica that quietly dismantled a century of economic assumptions about rational agents. Prospect Theory described something most gamblers already knew: losing £100 produces a psychological sting roughly twice as intense as the pleasure of gaining £100. The asymmetry is not a character flaw. It is architecture.
Watch it operate. A position drops 30% and the holder does nothing — not because they’ve reassessed the fundamentals and decided the thesis still holds, but because selling would make the loss real. The paper loss is theoretical. It might recover. It might. Meanwhile, a different position rises 25% and gets sold within the week. The gain is right there, tangible, and the fear of watching it evaporate overwhelms any rational assessment of whether the business has further to run. Kahneman and Tversky called this loss aversion. The practical shape of it is a portfolio that gradually accumulates its worst performers and sheds its best. You trim the flowers and water the weeds. Not because you’re foolish — because your nervous system is doing exactly what evolution designed it to do, which is to treat threats as more urgent than opportunities.
The trouble compounds. Imagine you’ve held that losing position through a twelve-month drawdown, and then markets begin to recover broadly. Everything else in your portfolio is climbing. That single losing name is flat. What do you feel? Relief that it stopped falling, probably. And here is where the next bias enters. You look at the last three months of upward market movement and some part of your pattern-recognition machinery whispers: this is the new normal. Markets go up now. The recovery will continue. Your losing position will come back.
This is recency bias — the brain’s tendency to construct predictions almost entirely from whatever happened most recently. It is particularly ruinous at extremes. In March 2009, after eighteen months of relentless decline, retail investors pulled record sums from equity funds. The S&P 500 had fallen 57% from its peak. Every recent data point confirmed that stocks destroy wealth. Investing more felt insane. That month was the bottom. By 2013, the index had tripled. The same mechanism works in reverse: by late 2021, three years of extraordinary returns had convinced a generation of new investors that stocks — particularly technology stocks — moved in one direction. Recency bias is cheapest in the middle. At the tails, where it matters most, it pushes behaviour in precisely the wrong direction at precisely the wrong moment.
But say you’ve navigated that. Say you held through the downturn and resisted the euphoria. You still have to contend with what happens after you’ve formed a view. A friend recommends a company. You do your research. You buy. Now you have a position — and with it, an identity. You are someone who owns this stock. You believed in it enough to act. From this point forward, every article you read about the company passes through a filter. Positive coverage confirms your judgement. You linger on it. Negative coverage gets contextualised, explained away, minimised. The filter is invisible. It doesn’t feel like bias — it feels like discernment.
Research from the University of California, Berkeley, published in 2012, demonstrated that professional analysts with strong prior ratings on a stock adjusted their price targets significantly less in response to contradicting earnings data than to confirming data. These are people whose literal job is dispassionate analysis. The confirmation bias operated anyway. The implication is uncomfortable: expertise does not inoculate you. It may even make the problem worse, because experts have more sophisticated tools for rationalising away inconvenient evidence.
Now stack another layer. You’ve formed your view, you’ve filtered the evidence, and you’re looking at the stock’s price history. It traded at $200 eighteen months ago. It’s $80 today. Bargain. Except the $200 figure is doing work that has nothing to do with the company’s current earnings, competitive position, or cash flow. It is an anchor — a reference number that your brain seized on and now uses as the baseline for “what this stock is worth.” Tversky and Kahneman demonstrated anchoring in 1974 with a rigged roulette wheel and a set of questions about African nations in the United Nations. The random number on the wheel shifted people’s estimates by 20 to 30 percentage points. In markets, the anchor is usually a prior price, and it is why the phrases “52-week high” and “52-week low” appear on every brokerage screen despite being analytically meaningless. A stock at its 52-week low is not cheap. A stock at its 52-week high is not expensive. These are statements about the past, not about value. But they feel like statements about value, and the feeling is what moves money.
By this stage, you are several cognitive layers deep. You held the loser too long. You projected recent performance into the future. You filtered new information through your existing belief. You anchored to an irrelevant number. And now the sunk cost arrives.
You’ve held this position for two years. You’ve spent dozens of hours researching it. You’ve defended it to your partner, to your friends, possibly on a forum somewhere. Walking away now doesn’t just mean accepting a financial loss — it means accepting that all that time, all that emotional investment, produced nothing. The rational calculation is clean: past costs are gone; only future expected returns should determine whether you hold. But rationality is not the operating system. Identity is. And your identity has become tangled up with this decision in ways that make the exit feel like confession.
The final layer is the quietest and possibly the most expensive. Overconfidence. Not arrogance — something subtler. In calibration studies conducted by Baruch Fischhoff in the 1970s and replicated dozens of times since, people who report being “90% confident” in a factual prediction turn out to be correct approximately 70% of the time. The gap between felt certainty and actual accuracy is consistent, measurable, and apparently immune to training. In markets, overconfidence manifests as higher trading frequency (because you believe you know something the price doesn’t reflect), lower diversification (because you believe your analysis of specific names is superior), and chronic underestimation of tail risks (because you believe you would see the crisis coming). Terrance Odean’s 1999 study of 10,000 brokerage accounts at a major discount firm found that the most active traders underperformed the least active by 7.1 percentage points annually. They weren’t unlucky. They were confident.
These six biases don’t operate independently. They layer, reinforce, interlock. Loss aversion keeps you in a bad position. Recency bias tells you it’ll recover. Confirmation bias filters the evidence to support that view. Anchoring gives you a false reference for what “recovery” even looks like. Sunk cost makes leaving feel like failure. And overconfidence assures you that your read on the situation is better than the market’s read.
The honest thing to say is not that you should “be aware” of these patterns — awareness is a weak intervention against machinery this deep. The honest thing is that Kahneman himself, in interviews late in his career, said he had not gotten appreciably better at avoiding his own biases despite fifty years of studying them. He just learned to distrust certain feelings — the feeling of certainty, specifically. That one, he said, was almost always the signal to slow down.
This article is for educational purposes only and does not constitute financial advice. Related reading: How to Tell if a Financial Influencer Is Worth Listening To