Holding Period: Winners vs. Losers — What the Timing Gap Reveals

Two positions are open at the same time. One is up 4%. One is down 4%. The winning one gets closed within the hour — “lock it in, don’t let it round-trip.” The losing one stays open. The thesis, supposedly, hasn’t changed. Or maybe it’s just that closing it would make the number real, and leaving it open lets it stay theoretical a little longer.

Almost every trader has lived this exact split-screen moment, usually without noticing that it’s a pattern at all. It just feels like two separate, reasonable decisions made on two separate days. But pull up the holding period for every winning position against every losing position across a few months, and the split-screen moment turns out to have been happening constantly, quietly, in the background — a habit dressed up as a series of one-off judgment calls.

A gap with a name

This asymmetry — closing positions that have appreciated sooner than positions that have declined — is one of the more thoroughly documented patterns in behavioral finance. It’s called the disposition effect: a tendency to sell winners early and hold losers longer, first named and measured by Terrance Odean.

In “Are Investors Reluctant to Realize Their Losses?” (Journal of Finance, 1998), Odean worked through a large set of discount-brokerage account records — real trades, real timestamps, not surveys or simulations — and found that investors closed out positions that had risen in value at roughly one and a half times the rate at which they closed positions that had fallen. Not occasionally. Not under stress. A general tendency across the account data.

That ratio is worth sitting with. It doesn’t say traders never held a winner or never cut a losing position. It says that, in aggregate, across thousands of decisions, the exits skewed — appreciating positions got closed out meaningfully faster and more often than declining ones did. A tilt, not a rule. But a tilt large enough, and consistent enough, to show up clearly in the account data of ordinary retail investors decades ago, in a market that looked nothing like today’s.

What the tilt looks like up close

Zoom into a single account and the effect stops being an abstract ratio and starts looking like a shape — a distribution of holding times with two different curves stacked on top of each other. Picture, hypothetically, an account where every winning position is charted by how long it stayed open, and every losing position is charted the same way. If the disposition effect is present, the winner curve bunches up short — hours, a day, maybe two — while the loser curve stretches out longer, with a tail that runs into days or weeks past where the winner curve has already emptied out.

Nobody sets out to build that shape on purpose. It accumulates from a string of decisions that each felt like the correct call at the time: taking the win because “it might reverse,” staying in the loss because “it just needs to come back.” Each individual choice is defensible. The pattern is only visible once you stop looking at trades one at a time and start looking at the holding-time distribution as a whole — winners on one side, losers on the other, and the gap between them.

Why the gap tends to form

The usual explanation in the behavioral-finance literature centers on loss aversion — the idea that a decline registers more heavily than an equivalent rise, so closing out a losing position feels like it costs something extra, something beyond the number on the screen. Selling a winner locks in a good outcome. Selling a loser locks in a bad one. If those two experiences aren’t symmetric, the exits won’t be symmetric either — even when the underlying price moves are.

It’s worth being precise about what this explains and what it doesn’t. It’s a mechanism for why the timing gap tends to appear, not a claim about what any specific holding-period gap means for any specific account, and not a claim about what letting a position run longer or shorter does to an outcome. The disposition effect describes a pattern in when people exit. It says nothing about whether that pattern is helping or hurting in any individual case — that depends on the position, the setup, the market, and a dozen variables a holding-time chart alone can’t see.

Where the pattern is worth having a mirror for

The reason this gap is interesting isn’t that it’s dramatic. It’s that it’s invisible from inside a single trading day and completely visible from a slight distance. In the moment, closing a winner and holding a loser are two unrelated decisions, made hours apart, about different tickers, for reasons that felt specific to each one. Only when the holding times get lined up side by side — winners against losers, weeks or months of them at once — does the asymmetry stop looking like a coincidence and start looking like a signature.

That’s the general shape of behavioral measurement: not judging any single exit, just making the aggregate pattern legible enough to see rather than remember. Odean’s account data made that signature visible across a whole population of investors; the same shape is available in a single trader’s own history, if the holding times ever get pulled out of the noise of individual trades and set next to each other. biaX is built around exactly that kind of retrospective view — surfacing the footprint, not scoring it. What a trader does with the gap once they can see it is, deliberately, left entirely up to them.