6 min read

Why time flies when you're having fun

Ten minutes waiting for a delayed train and ten minutes into a good conversation are the same ten minutes. Nobody experiences them that way. The mechanism behind the difference is simpler than you'd expect, and it has a twist at the end.

Attention is the valve

The standard explanation adds one component to the pacemaker-accumulator model: a gate between the pacemaker and the accumulator, controlled by attention. Pulses only accumulate while you're attending to time. Attend hard, and the gate is wide open, pulses pile up, and the interval registers as long. Attend to something else, and the gate narrows, pulses are lost, and the same interval registers as short.

This is a genuinely testable claim, and the test is the dual-task experiment. Give someone an interval to judge while also making them do something demanding, and their duration estimates shrink, reliably and in proportion to how demanding the second task is. Nothing about the interval changed. Only the share of attention pointed at it did.

So “time flies when you're having fun” is slightly mislabelled. Fun isn't the operative variable. Absorption is. Time flies during a tax return too, provided the return is complicated enough to hold you. What reliably makes time crawl is having attention available to spend on time itself, which is exactly the condition of being bored: an unengaging task and nowhere else for your attention to go.

The twist: memory reverses the sign

Now the part that makes the folk saying confusing. Ask someone about a week-long holiday and you often hear both halves of a contradiction: it flew by, and it feels like ages ago. Both are true, because they're answers to two different questions handled by two different systems.

In the moment, felt duration tracks attention to time, so an engaging week passes quickly. In retrospect, felt duration tracks how much you remember, so an engaging week, dense with novel and distinguishable events, expands. A featureless week at home produces the mirror image: it drags while you live it and collapses to nothing when you look back, because there are almost no distinct memories to give it length.

This is the single most useful thing in the literature for understanding your own life, and it's worth restating plainly:

Fear doesn't actually slow time down

People who have been in car crashes describe the world going into slow motion, and the obvious interpretation is that under threat the brain speeds up its sampling rate, like a high-speed camera. This has been tested more carefully than you'd imagine, including by dropping volunteers in free-fall while they attempted to read a display flickering just above the normal threshold of perception.

If perception genuinely accelerated, the display would have become readable. It didn't. Yet the same volunteers, asked afterwards how long their fall had lasted, overestimated it substantially. The conclusion was that frightening events aren't experienced in slow motion; they're remembered in unusual density, and the recall of a richly encoded few seconds feels like the recall of a longer stretch. Slow motion is a memory effect wearing a perception costume.

The oddball effect

A smaller, stranger cousin of the same finding: in a stream of identical images, one unexpected image appears to last longer than the others, even when every image is shown for exactly the same time. Novelty inflates duration. One explanation is that a novel stimulus recruits more neural processing, and the brain reads “more processing” as “more time”.

Turn that around and you get one of the better explanations for why adult years feel short. Repetition is compression. A week where everything is familiar is a week your brain can encode cheaply, and cheap encoding is what shortness feels like in retrospect.

What this means when you're playing

A timing game is a pure prospective task: you always know the question is coming. So the attentional gate is under your control, and how you point it is a real strategic choice.

Two things follow. First, whatever you do during a round, do it consistently. If you stare at the dot with an empty mind on one round and run an internal monologue on the next, you've changed how wide the gate is opened, and your two estimates are drawn from different distributions. Consistency of attention matters more than intensity of attention.

Second, dividing your attention shortens felt duration, which means a distracted round will tend to come in short: you'll stop the clock early, or call a five-second interval four. If you notice a run of undershoots and you were half-listening to something else, that's not your internal clock breaking. That's the gate doing exactly what the model predicts.