Recency, the gambler's fallacy and sunk cost

Three more biases finish the set. They are smaller than loss aversion but they are the ones you will meet most often on an ordinary Tuesday, and each has a clean research origin.

Recency and availability

In 1973 Tversky and Kahneman described the availability heuristic: people judge how likely something is by how easily examples come to mind. Recent, vivid and emotional events come to mind easily, so they feel more probable than they are.

For a trader this means the last few trades dominate your sense of what the market is doing and what you are capable of.

• Three winners in a row and every setup looks clean; size drifts up.
• Three losers in a row and the same setups look suspicious; you skip the one that would have worked.
• A single violent reversal against you last week and you now exit every trade early "because you've seen what can happen."
• A big move you missed on Monday and by Wednesday you are chasing anything that resembles it.

None of this is analysis. It is memory weighting. The market on trade 41 does not know what happened on trades 38, 39 and 40.

The defence is to make the long sample as available as the short one. A running expectancy over your last 100 trades, displayed at the top of your journal, is a number your brain can weigh against the last three. Module 4 builds this.

The gambler's fallacy and its cousin

Tversky and Kahneman's 1971 paper "Belief in the Law of Small Numbers" showed that people, including trained scientists, expect small samples to look like the population they come from. Flip a fair coin five times and get five heads, and most people feel that tails is "due." It is not. The coin has no memory.

Traders run this in both directions.

The gambler's fallacy proper: after four losing trades, the fifth "has to" win, so it gets sized up. Nothing about the previous four changes the odds on the fifth. If your system wins 45% of the time, the fifth trade wins 45% of the time.

The hot-hand version: after four winners, you are "in the zone," so the fifth gets sized up. Gilovich, Vallone and Tversky's 1985 study of basketball shooting found that fans and players strongly believed in hot streaks while the shot data showed sequences consistent with chance. (Later work by Miller and Sanjurjo found a subtle bias in that analysis and a possible small hot-hand effect in basketball. Whether one exists in your trading is a question for your log, and for most people the answer is no.)

Both versions share a structure: a run of outcomes is treated as information about the next one, and the trader changes size in response. Changing size on streaks is one of the fastest ways to turn a positive-expectancy system into a negative one, because the size increase lands, on average, on a perfectly ordinary trade.

Key idea: Your last few trades are not information about your next trade. Recency makes them feel like it, and the gambler's fallacy makes you act on it, usually by changing size. Fixed size is the antidote.

Sunk cost

Arkes and Blumer's 1985 paper "The Psychology of Sunk Cost" showed that people who had paid more for something used it more, even when the money was unrecoverable and the decision should have depended only on what happens next. Their examples included theatre tickets and a fictional company that kept funding a doomed project because it had already spent so much.

Trading has three common sunk-cost traps.

The position. "I've already lost $400 on this, I'm not closing now." The $400 is gone whether you close or not. The only question is whether the next hour's expected value is positive at this size. Usually the honest answer is that you would not open this position now, at this price, if you were flat, and if you would not open it, you should not be holding it.

The research. You spent the weekend building a thesis on one market. Monday's price action says the thesis is wrong. The weekend feels wasted if you do not trade it, so you trade it anyway. The weekend is spent either way.

The career. Two years and $30,000 into trading, the thought of stopping feels like admitting all of it was for nothing. This one is the heaviest and Module 6 gives it a full lesson. For now: what you have spent is not a reason to spend more. The only question is whether the next year has positive expected value, and that depends on evidence, not on what the last two cost.

A useful test for any sunk-cost decision: describe the situation to yourself as if you had just arrived. "I am flat. This market is here. Would I enter now, with this stop, at this size?" If no, close it.

Putting the module together

Each bias in this module is a specific, named, replicated finding, and each produces a specific behaviour:

You will not stop having these; Kahneman, who spent his life studying them, said he still felt them. What you can do is design a process in which the biased decision has already been made, correctly, before the moment arrives. That is the rest of this course.

Try it: Write each of the six biases on a card with one sentence about how it shows up for you specifically. Keep the cards by your screen for a week. Each time you catch one in the act, put a tally on the card. At the end of the week you will know which two or three are yours, and those are the ones to build rules around first.

Recap

• Recency (availability heuristic, Tversky and Kahneman 1973) makes the last few trades feel like information about the next one; display a long-sample expectancy to counterbalance it.
• The gambler's fallacy and hot-hand belief (belief in the law of small numbers, 1971; Gilovich et al. 1985) lead to sizing up on streaks; fixed size is the fix.
• Sunk cost (Arkes and Blumer 1985) keeps traders in positions, theses and careers because of what has already been spent; ask "would I enter this now, flat?"
• All six biases are universal and will not go away; the goal is a process that makes the decision before the bias can.
• Track which biases are yours for one week and build rules for those first.