Trading psychology: biases, tilt, routines and when to stop

A plan on paper says when to open a trade, where it is wrong and how much is at risk. Between the plan and the result stands the person who has to carry it out, with money at stake and a price moving on the screen.

Psychologists and economists have documented, over several decades, ways in which judgement departs from plain arithmetic when outcomes are uncertain. These are not faults of character and they are not confined to beginners. They are tendencies of ordinary thinking, and they show most when the stakes feel high and the feedback is quick, which describes a trading screen well.

Knowing their names does not remove them. It makes them easier to recognise in one’s own record, which is where the tools further down this page come in.

Loss aversion. A loss is felt more keenly than a gain of the same size. In trading it tends to show as closing a winning trade early, to make the gain safe, and holding a losing one, because closing it would make the loss real. The pattern of selling winners and keeping losers is well enough known to have a name, the disposition effect.

Recency. What happened last is given more weight than what happened over a longer time. Three wins in a row feel like skill and three losses feel like a broken method, although a run of either length is ordinary in any series of uncertain outcomes.

Anchoring. A judgement leans on the first figure met, whether or not that figure has any bearing. In trading the anchor is often the price at which a position was opened, or a recent high or low: a trade is held until it is “back to where I got in”, although the market keeps no record of where anyone got in.

Confirmation. Once a view is formed, evidence that agrees with it is noticed and remembered, and evidence against it is explained away. A trader who is long finds the reasons to stay long.

Overconfidence. People tend to rate their own judgement, and the precision of their own forecasts, higher than the results justify. It grows after a good stretch, and it tends to show as larger positions and more trades.

• Each of these bends a decision in a direction that can be predicted, which is why they can be planned for.
• None of them is a signal. Recognising a bias says something about the decision, and nothing about what the price will do.

Tilt is a word borrowed from card players. It describes a state in which emotion, usually after a loss or a run of losses, has taken over from the plan: decisions come faster, sizes grow, and the rules that were clear in the morning are set aside.

Revenge trading is its commonest form. A loss is felt as something to be won back at once, from the same market, so the next trade is opened quickly and often larger. The trade is not taken because the plan calls for it. It is taken because of the previous trade.

The arithmetic is unforgiving. A larger position after a loss puts more of a smaller balance at risk, and a loss needs a larger percentage gain to recover than the loss itself was. Tilt after a win exists too: a run of gains can lead to the same abandonment of size and rule, from the opposite mood.

Common early signs are physical and behavioural: a quickened pulse, moving a stop that was fixed, opening a trade within moments of closing one, watching the balance instead of the chart. They are easier to act on when they have been written down in advance as reasons to pause.

A routine moves decisions to a time when nothing is at stake. What may be traded, how much is at risk on each trade, where a trade is wrong and what ends the session are settled beforehand and written down. During the session the only question left is whether the conditions written down have been met.

A routine usually has three parts. Before: a look at the calendar of scheduled releases, the plan for the day and an honest note of one’s own state, since tiredness and distraction are conditions too. During: the plan, followed. After: the record, filled in while the reasons are still remembered.

A checklist is the routine in its shortest form: a few lines, read before each trade, each answered yes or no. Is this one of the conditions in the plan? Is the size the planned size? Is the level at which the trade is wrong already set? Its value is that it is dull. It asks the same questions on a good day and on a bad one, and a no is an answer that needs no further thought.

A routine does not make a method profitable. It makes the results a fair test of the method, because the method is what was actually traded.

Memory is a poor record of trading. It keeps the vivid trades and loses the dull ones, and it rewrites the reasons afterwards. A journal is the correction: for each trade, what the plan was, what was done, what happened and what state the trader was in.

Its use is in the reading, not the writing. Read over some weeks, a journal can show whether losing trades were held longer than winning ones, whether size rose after losses, whether results differ by time of day or by mood, and how often the plan was not followed. Those are the biases above, found in one’s own figures.

A journal records what happened. It cannot say what will happen, and a small number of trades shows very little: a pattern in twenty entries may be chance.

The decision to stop is best made before the session, because the moment it is needed is the moment judgement is least reliable. A stopping rule is a line written in advance. Common forms are a limit on the day’s loss, a limit on the number of losing trades in a row, a limit on the number of trades, and a fixed time.

A stopping rule also covers the good day. Some traders stop after a set gain or a set number of trades, not because gains are dangerous but because the plan described a number of decisions, and the decisions after that are unplanned.

Stopping does not recover a loss and does not prevent the next one. What it limits is the damage done in the state least suited to deciding. The amounts are for each person to set: nothing here says what a limit should be.