Trading Psychology: The Biases That Cost Money
Most trading mistakes are not analytical failures. They are predictable cognitive patterns that appear reliably under financial pressure.
Traders rarely lose because they could not understand a chart. They lose because knowing what to do and doing it under pressure are different problems.
Loss aversion
Losses hurt roughly twice as much as equivalent gains feel good. The consequence is systematic: traders hold losers hoping to avoid crystallising the pain, and close winners quickly to secure the relief. This inverts the risk-reward profile of the entire system, one trade at a time.
The disposition effect
The specific version of the above — a measurable tendency to sell winners and keep losers. It is one of the best-documented findings in behavioural finance and it converts positive-expectancy systems into losing ones without changing a single entry.
Confirmation bias
Once positioned, you notice supporting evidence and discount contradicting evidence. The chart you found convincing before entry becomes more convincing afterwards, not because anything changed but because you now need it to be true.
Recency bias
Recent events feel more likely to repeat. After three winners, position sizes creep up. After three losses, valid setups get skipped. Both are responses to noise, and both damage results at exactly the wrong moment.
Sunk cost
Adding to a losing position to "improve the average" is sunk cost reasoning wearing a technical costume. The capital already lost is gone and should not influence the next decision. Averaging down into a losing trade is among the most reliable account-destroying behaviours there is.
Overconfidence after wins
The most dangerous moment is not after a loss. It is after a run of wins.
Success feels like skill even when it was variance. Position sizes grow, rules loosen, and the eventual reversion arrives against a much larger position than the system was designed for.
Countermeasures that actually help
- Decide in advance. Entry, stop and target set before the position exists, when you have no stake in the answer.
- Fixed position sizing. Removes conviction — the least reliable input — from the sizing decision.
- Journal reasoning, not just outcomes. Patterns in your thinking only become visible in writing.
- Automate what you can. A resting stop order does not feel hope.
- Set a stopping rule. A daily or weekly loss limit that ends the session, defined while calm.
None of this eliminates the biases. They are features of human cognition, not defects to be trained away. The realistic goal is to build a process that limits how much damage they can do.
Educational content only. Not financial advice.
See the models in action
Reading about forecasting is one thing. Create a free account and get 5 AI predictions a month — no card required.
Start free
