There is a specific pattern in which traders hold losing positions far longer than any written rule would allow, then sell winners at the first flush of profit — often within days of entry. When asked to explain it afterward, they rarely say "I was afraid of losing." They say something closer to: "I didn't want to be wrong for having sold." The first formulation points to loss aversion. The second one points somewhere else: to the asymmetric weight the mind places on the regret of acting versus the regret of doing nothing. These are related but distinct forces, and conflating them produces a confused diagnosis and the wrong fix.

By the end of this article you will be able to distinguish two types of error — errors of commission (you did something that failed) and errors of omission (you did nothing and the market moved without you) — explain why the mind weights them differently in the short run, and use that distinction to interrogate hold decisions before regret, not after.

The Two Types of Regret

In 1982, Daniel Kahneman and Amos Tversky published a set of scenarios in Scientific American that captured an asymmetry most traders feel but rarely name. One scenario presents two investors: Paul owned shares in Company A and considered switching to Company B but did not; George owned Company B and switched to Company A. At the end of the period, each would have been $1,200 better off had he done the opposite. The finding was straightforward: participants judged that George — who acted — would feel more regret than Paul, despite identical financial outcomes. The loss was the same dollar amount. The pain was not the same.

This is the action-effect: negative outcomes produce more regret when they follow action than when they follow inaction. Feldman and Albarracín (2017) provided a formal theoretical account of the social-norms mechanism underlying this asymmetry in Journal of Experimental Social Psychology (vol. 69, pp. 111–120, DOI 10.1016/j.jesp.2016.07.009). Janet Landman's 1987 empirical work in the Personality and Social Psychology Bulletin (vol. 13, no. 4, pp. 524–536) found the same directional pattern: participants imagined greater regret following action than inaction, especially for negative outcomes.

The temporal dimension matters and is contested. Gilovich and Medvec (1994, Journal of Personality and Social Psychology, vol. 67, pp. 357–365) found that in the short run, people regret actions more than inactions; over longer horizons, inaction regrets tended to dominate — the things people did not do became the heavier burden. A 2022 replication study (Yeung and Feldman, Collabra: Psychology, vol. 8, article 37122) replicated the short-run action-regret pattern but did not replicate the long-run inaction dominance, which failed to reach statistical significance. A separate museum-based study with 2,600 visitors found that significantly more participants reported being more troubled by their most regrettable action than inaction — Richardson and Gilovich (2023) reported this figure as 58% (n = 743), published in Royal Society Open Science, vol. 10, issue 6, article 221574, DOI 10.1098/rsos.221574. The short-run asymmetry is well-supported; the claim that inaction regret dominates over a lifetime is original research with mixed replication. Both directions are relevant to trading — which is why this article handles them as distinct.

How the Asymmetry Works Inside a Trade

The action-effect produces a specific distortion in exit behavior. When a position is in profit, selling is an action. If the position continues to rise after you sell, you acted and it cost you — George's situation. The mind anticipates that regret strongly enough to bias toward holding: not because the thesis has strengthened, but because staying in avoids the category of "I acted and it went wrong." This is anticipatory regret operating on the winner side, pulling toward inaction exactly when the thesis may be exhausted.

When a position is losing, the mirror logic applies. Selling to cut a loss is an action. Holding is inaction. The fear of George's regret — "I sold it and then it recovered" — keeps the losing position open beyond any rational invalidation point. The trader is not analyzing the trade. They are managing the anticipated pain of being the one who sold at the bottom.

This is meaningfully different from loss aversion, which focuses on the asymmetric pain of losses versus gains in the abstract. Regret aversion focuses specifically on the attribution of that pain to a choice that could have been made differently. The loss does not just hurt; it is evidence against a decision, and that evidence is what the mind is running from. The distinction matters because the fix is different: loss aversion is addressed by reframing the math; regret aversion is addressed by examining whether the hold or exit decision is driven by thesis or by anticipated emotion.

Shefrin and Statman's landmark 1985 paper in The Journal of Finance (vol. 40, no. 3, pp. 777–790) identified regret avoidance as one of four psychological components of the disposition effect — alongside loss aversion, mental accounting, and self-control problems. They noted explicitly that none of these alone produces the full effect. Loss aversion is one piece; regret aversion is a related but distinct piece. Treating them as the same variable produces an incomplete map.

What the Pattern Cost During the 2000–2002 Crash

The NASDAQ Composite reached a closing high of approximately 5,048 on March 10, 2000, before declining 78% to a trough in October 2002. It did not surpass that 2000 peak on a closing basis until April 2015 — fifteen years later. The S&P 500 fell approximately 49% over the same period.

This episode is relevant not because of the index figures but because of the holding behavior Terrance Odean documented in brokerage account data. Using records from 10,000 accounts spanning 1987 through 1993, Odean found that outside of December, the proportion of gains realized (PGR) was 14.8% while the proportion of losses realized (PLR) was 9.8% — investors were roughly 51% more likely to sell a winner than a loser in any given period. In December, when tax-loss selling pressure arrived, the pattern reversed: PLR exceeded PGR. The implication is that the holding of losers was not driven by conviction about the fundamental case; it was sensitive to external cost structures in a way that a genuine fundamental hold would not be. ("Are Investors Reluctant to Realize Their Losses?" The Journal of Finance, vol. 53, no. 5, pp. 1775–1798, 1998.)

The dot-com decline produced a two-year environment in which a buy made in early 2000 would have shifted over time from an action-regret (buying was the wrong move) to an escalating inaction-regret structure (not selling before the decline accelerated, then not selling at each subsequent level). The exit that would have ended the loss required taking an action — selling — and incurring the regret type that the brain weights most heavily in the short term: the regret of having done something that made it worse. That single emotional friction, compounded across thousands of separate hold decisions, is the behavioral mechanism most consistent with the aggregate data. This interpretation is supported by the Shefrin-Statman and Odean frameworks, but no primary study isolated regret aversion as the specific causal mechanism for that episode in isolation.

Abu's Speed Run includes the March 2000 NASDAQ events. When you run those sequences, notice which hold decisions feel intuitively "safe" — and whether that safety feeling is thesis-based or based on avoiding the specific act of selling.

What the Asymmetry Costs

The asymmetric weighting of action versus inaction regret produces three distinct costs in process terms.

First, exit timing drifts toward what avoids commission regret rather than what the thesis supports. A position is held past its invalidation point not because new information arrived to justify the hold, but because selling feels like authoring a failure. Staying feels passive and therefore emotionally cheaper — even when the thesis is objectively broken.

Second, winners get cut early for the same reason, run in reverse. Selling a winner converts a paper gain into a realized gain, which feels like securing safety. But if the price continues higher, that sale becomes a source of action regret. The anticipation of that regret — felt before the decision, not after — biases the exit earlier than the setup warrants.

Third, and most damaging: the process of evaluating whether to hold or exit gets contaminated. Instead of asking "Does the original thesis still hold?" the trader asks a subtler and less useful question: "Which decision would make me feel worse if the price moves against it?" These questions produce different answers. The first one uses market information. The second one uses anticipated emotion as the variable.

The pre-defined exit is one of the structural solutions to this contamination. If the stop is written before entry — when regret is not yet on the table — then the exit decision is not made in the moment when regret aversion is highest. The pre-defined exit article covers the mechanics of writing an invalidation condition before entry; the discipline here is understanding why it matters emotionally, not just mathematically.

The Discipline: Classify the Hold Before It Becomes a Regret

Regret aversion is hardest to counter in real time because its influence is felt as judgment rather than emotion. "I shouldn't sell here" does not feel like fear; it feels like analysis. The discipline is to build a classification habit that forces the question before the decision calcifies.

When reviewing an open position, ask one question: Is my reason for holding this position a thesis-driven reason or a regret-avoidance reason? A thesis-driven reason references market conditions, price structure, or time-based criteria that were established before the position was entered. A regret-avoidance reason references what you would feel if you sold and the price moved in either direction. The distinction is the signal.

This is not a guarantee of better decisions — a thesis-driven hold can still lose, and the market does not reward honest reasoning. What the classification does is separate the decision from the emotional variable that contaminates it. If you can honestly label a hold as "thesis-driven," the hold is defensible. If the honest label is "I don't want to sell and be wrong," you have identified anticipatory regret as the driving factor — and that is not a valid hold reason.

Pre-commitment structures help enormously here. A written invalidation condition — "I exit if price closes below X" — converts the exit from an active choice (action, commission regret risk) into a mechanical consequence of a prior rule. The rule was made by a self that was not in the position yet; that self had no emotional stake in the outcome. A reversibility frame also clarifies the stakes: most exits are reversible. You can re-enter. The catastrophizing that regret aversion produces — "if I sell this, I lose forever" — does not hold structurally.

Finally: keep the decision journal current during a hold, not only at exit. A journal entry written while holding an underwater position will reveal the reasoning in real time. Reviewing it one session later often makes the distinction between thesis and regret-avoidance visible in the language — thesis-driven notes reference the market; regret-avoidance notes reference what the trader feels. The sunk-cost article covers the related pattern where prior investment in a position inflates the hold beyond rational grounds. Both biases frequently operate in the same trade.

Simulator Exercise: Speed Run Regret Mapping

This exercise runs inside a Speed Run or Arena session in Abu Terminal. It requires three hold decisions — moments in the run where you chose to stay in a position rather than exit.

After the session ends, locate three hold decisions in the debrief. For each one, write a short label: thesis-driven or anticipatory-regret-driven. A thesis-driven label means you can cite a specific condition that kept the position valid — a price level that had not been broken, a time window that had not closed, a catalyst that had not resolved. An anticipatory-regret-driven label means the primary reason you did not exit was concern about how it would feel if you sold and the price moved.

You are not graded on getting the label "right." The exercise is the labeling itself. The behavioral signal is asymmetry: if all three hold decisions receive the same label, run it again and be more precise. Most traders who do this honestly find the split is uneven — several thesis-driven holds and at least one regret-driven one, or vice versa — and the regret-driven holds tend to cluster at the positions that had been underwater the longest. That clustering is the pattern to carry into the next session.

A second run of the exercise: after labeling, check the outcomes. Did the thesis-driven holds tend to resolve better than the regret-driven ones? This is not a valid statistical test over three data points — sample size is too small to draw conclusions. What it builds is the habit of examining hold decisions by type rather than only by outcome. The goal of the drill is to make the regret-versus-thesis distinction a reflex, not a post-hoc rationalization.

The Speed Run sequences from the year 2000 are particularly effective for this exercise: the positions that declined the furthest tend to accumulate the most regret-driven hold pressure. Running them in simulator conditions — with no real money at stake — lets you feel the pull toward holding without committing to a decision your capital would have to absorb.

Regret Aversion and the Action / Inaction Asymmetry