Why Can't I Cut My Losses? The Psychology Behind Holding Losing Trades

The scene

The position is down. Not catastrophically — just enough to notice, enough to check twice a day instead of once. The plan, made calmly before the position existed, was simple: if it drops past a certain level, close it. That level has come and gone.

And yet the finger doesn’t move to the sell button. Instead, a familiar sentence surfaces: it’s only a paper loss until I sell. The thesis gets a quiet re-examination. Maybe the timeline was too short. Maybe the market just needs a session or two. The number on the screen stays red, and the plan — so clear a week ago — stops being the thing that governs the decision.

This is not a failure of arithmetic. Nobody forgets that red is worse than green. What’s happening is a well-documented mismatch between what a plan says on paper and what a brain does at the actual moment of decision, and it has been studied for decades.

Losses don’t weigh the same as favorable moves

The foundational finding here comes from Kahneman & Tversky’s 1979 paper on prospect theory. Their core result: people don’t evaluate outcomes on a neutral scale where an upside move of a given size and a downside move of the same size carry equal weight. Losses loom larger. A decline registers as psychologically heavier than an equivalent increase feels good.

That asymmetry alone explains part of the hesitation — closing a losing position means fully registering a loss that, while unrealized, still hasn’t been “felt” at full weight. But Kahneman & Tversky’s work goes a step further, into what’s sometimes called the reflection effect: people tend toward risk-averse choices when facing upside outcomes, and risk-seeking choices when facing losses.

Translated to a chart: once a position has moved from green to red, the decision-maker isn’t in the same psychological territory anymore. They’re now choosing between a certain, smaller loss (sell now) and an uncertain outcome that includes the chance of no loss at all (hold and see). Prospect theory’s finding is that in this specific domain — the domain of losses — the uncertain option tends to look more attractive than it would if the exact same odds applied to favorable outcomes. The plan said “close it here.” The moment says “there’s still a chance it comes back.” Those are two different reference points doing two different kinds of math, and the second one tends to win in the moment.

The paper loss fallacy

Layered on top of that asymmetry is a mental accounting trick: the belief that an unrealized loss isn’t “real” yet. As long as the position stays open, the loss lives in a kind of suspended state — visible, uncomfortable, but not final. Selling is the act that converts it from a number on a screen into a closed, countable event.

This is worth sitting with: it reveals what actually triggers the discomfort. It isn’t the decline itself — that already happened, days or weeks ago, priced into the account regardless of whether the position is open or closed. What triggers the discomfort is the realization — the moment the mental account closes and the outcome becomes final. Holding the position doesn’t change the economic reality. It postpones the psychological event. The paper loss fallacy is the story that gets told to justify that postponement: it isn’t a loss until it’s sold, so don’t sell.

Breakeven as a gravitational pull

There’s a third piece, and it’s the one that shows up most clearly in trade data: the pull toward the entry price. Not the stop level, not a technical level, not a level with any predictive information at all — the price paid. Traders describe waiting “just to get back to even” as though breakeven were a meaningful signal about where the price is headed next, when it’s really just the point where the reference frame flips back from loss to favorable territory.

That flip matters because of the asymmetry described above. Near the entry price, a trader is right at the boundary between the two domains prospect theory describes — the one where losses loom large and risk-seeking behavior tends to creep in, and the one where things look calmer. The pull to “just get back to even” isn’t really about the price level. It’s about wanting to exit the loss domain through the door of a small favorable outcome rather than the door of a realized decline.

What this looks like in the data

Odean’s 1998 research on individual investor accounts gave this pattern a name and a measurement: the disposition effect. Looking at real trading records, Odean found a tendency to sell positions that had appreciated relatively quickly, while positions that had declined were held meaningfully longer — a measurable skew in exit timing between winning and losing positions, showing up consistently across accounts rather than as an occasional quirk.

That’s the empirical signature of everything described above. The asymmetric weighting of losses, the paper loss fallacy, the pull of the entry price — none of these are visible in the moment they’re happening. What is visible, after the fact, is the shape they leave behind: a gap between how long a position that appreciated was held and how long one that declined was held, showing up again and again in the same account.

The mirror, not the fix

None of this explains why any single position was held past its plan. Behavioral research doesn’t diagnose a single decision — it describes a pattern that tends to recur across many of them, and explains the mechanism behind why that pattern is so stable across traders and markets.

What’s interesting is that the mechanism is invisible from inside a single trade, but not invisible across a body of trades. A holding-time distribution — how long positions that declined were held versus positions that appreciated — is just a measurement. It doesn’t say what should have happened. It says what the footprint looks like, and lets the shape of that footprint speak for itself. That distinction — between narrating a plan and measuring the pattern that actually played out — is closer to what behavior measurement, done honestly, is for.