Cognitive biases in investing and trading

Confirmation, anchoring, recency, herding and hindsight bias explained with market examples, the research behind each, and a simple check for every one.

Part of the Psychology and Discipline track on Agenticks. About 10 minutes, written for a intermediate reader.

Read first: Common beginner mistakes (honest).

Last reviewed .

Key takeaways

A cognitive bias is a predictable way judgment drifts away from what the evidence supports. It is not a sign of low intelligence. It is a mental shortcut that usually saves time and sometimes misleads, and the shortcuts that serve us well in daily life can be costly when money, uncertainty and time pressure arrive together. The U.S. Securities and Exchange Commission keeps a plain list of investing behaviors that research links to weaker results, including active trading, holding losers and selling winners, chasing past performance, and manias and panics. Most of those behaviors trace back to a handful of biases. This lesson covers five of them. Each comes with an example for a long-term investor, an example for an active trader, and a check you can use.

A bias is a shortcut, not a character flaw

Professionals show these biases too. The goal is not to become a person without biases. The goal is to know which shortcut you lean on and to put a rule or a check in its path before it reaches your money.

Confirmation bias Confirmation bias is looking for, noticing and remembering the evidence that agrees with what you already believe, while explaining away the evidence that does not. An investor who owns a company reads every upbeat article about it and skims past the warning in the annual report. A day trader who is long sees each small bounce as proof, and treats the break of the level that was supposed to invalidate the trade as noise. The check: before you act, write the strongest case for the other side in two or three sentences. If you cannot, you have not looked.

Anchoring Anchoring is leaning on the first number you saw and adjusting too little away from it. In a classic experiment, people watched a wheel of fortune stop on 10 or on 65, then estimated the share of African countries in the United Nations. The group that saw 10 gave a median guess of 25%. The group that saw 65 guessed 45%. The wheel had nothing to do with the question. In markets the anchor is usually the price you paid, or a past high. An investor waits for a fallen stock to get back to even before selling. A trader treats yesterday's high as a fair price simply because it is on the screen.

Worked example: waiting to get back to even

An investor buys 100 shares at $50 and the price falls to $35. Round, illustrative numbers.

Amount paid (100 x $50)$5,000
Value now (100 x $35)$3,500
Loss so far (1,500 / 5,000)-$1,500 (-30%)
Gain needed to reach $50 again ((50 - 35) / 35)+42.9%
What the anchor asks forA 42.9% rise just to break even

The $50 is a fact about the buyer, not about the company. Whether to hold depends on what the investment is worth now, not on what it cost.

Recency bias Recency bias is giving the latest events too much weight when forecasting. Greenwood and Shleifer studied six surveys of investor expectations from 1963 to 2011 and found that people expected higher stock returns after returns had been high, and that those expectations tended to be too optimistic at exactly those moments. For investors, that is the pull to add money after a strong year and to stop after a bad one. For traders, it is tripling size after three winning trades in a row, as if three trades said much about the next hundred. Our lesson on sample size explains why they do not. The check: before acting on a streak, look at a much longer stretch of history and ask what the last few months would look like inside it.

Herding Herding is doing what many others are doing because they are doing it. In brokerage records covering tens of thousands of households, individual investors' trades turned out to be strongly correlated: many bought the same stocks at the same time, and kept doing so month after month. Herds produce the manias and panics on the SEC's list. The investor version is buying a stock because it is everywhere on social media. The trader version is jumping into a move that a chat room is excited about, after most of it has already happened. The check: wait a set period, such as a day, before acting on anything you heard from a crowd, and write down your own reason in one sentence.

Hindsight bias Hindsight bias is the feeling, once you know how something turned out, that it was obvious all along. Fischhoff found that people told the outcome of an event rated it as far more predictable than people who were not told, and they barely noticed their own judgment had shifted. Every crash looks obvious afterwards. Every chart shows the perfect entry once the move is over, which is why a screenshot proves nothing. The check: write your reason for a decision before the result is known. A record made in advance cannot be rewritten by the outcome.

Confirmation bias
Reading only the bullish posts about a stock you already own
Anchoring
Refusing to sell until the price returns to what you paid
Recency bias
Doubling your size after three winning trades in a row
Herding
Buying a stock because everyone in a chat room is buying it
Hindsight bias
Saying a crash was obvious once it is over

Turning checks into habits Knowing about a bias does little on its own. What helps is a routine that runs whether or not you feel biased at the time: Argue the other side in writing before any new position. Decide your exits before you enter. A sell rule chosen in advance removes the anchor from the decision. Zoom out. Compare any recent streak with years of history, and spread risk so no single story dominates, as in our lesson on diversification. Pause before following a crowd. Keep a record. A trading journal captures your reasons before the outcome, which is the only reliable cure for hindsight. The next lesson looks at a bias with its own body of research: why losses hurt more than gains feel good, and what that does to selling decisions.

A trader writes down, before entering, the price at which the idea is wrong and the reason for the trade. Which two biases does this habit work against most directly? Anchoring and hindsight bias An exit decided in advance replaces the purchase price as the reference point, and a reason written before the result cannot be rewritten by hindsight.

You can name the five biases

You can now recognize confirmation, anchoring, recency, herding and hindsight bias in market decisions, and you have a check for each one.

Check your understanding

You bought a stock at $40 and it now trades at $30. You decide to hold until it is back to $40 before selling. Which bias is this most likely to be?
Anchoring. The $40 is the anchor. The market does not know or care what you paid, so it is a poor guide to whether the stock is worth holding now.
Which habit most directly works against confirmation bias?
Writing down the strongest case against your position before you act. Confirmation bias means seeking evidence that agrees with you. Deliberately building the opposite case forces the missing evidence into view.
Survey research on investor expectations (Greenwood and Shleifer, 2014) found that expected stock returns tend to be highest when:
Recent returns have been high. Expectations follow recent returns. The study also found they were often too optimistic at exactly those times, which is recency bias in data.
Why is hindsight bias dangerous for someone learning to trade?
It makes past moves look predictable, so you overrate how well you could have called them. If every chart looks obvious afterwards, you learn the wrong lesson about your own skill. Recording your reasoning before the outcome is the cure.

Common questions

What are the most common cognitive biases in investing?
Five that show up again and again are confirmation bias, anchoring, recency bias, herding and hindsight bias. Each is a mental shortcut that can push an investor or trader to act on a story instead of the evidence.
Can you get rid of cognitive biases?
Not completely, because they are part of how people think. What works better is designing checks around them: written rules, exits decided in advance, a journal, and a pause before acting on a crowd or a recent streak.

Terms defined in this lesson

Continue

Sources

Written and reviewed by the Agenticks Learn editorial team. How we write and check lessons.

Agenticks Learn is general education. It is not financial, investment, tax or legal advice, it does not take your situation into account, and nothing on it is a recommendation to buy, sell or hold any security or to open any account. Examples use illustrative numbers. Investing involves risk, including losing money you put in, and past performance does not predict future results. Tax rules and limits change; the official sources linked in each lesson are the final word.