Peak-End Rule

Six good days disappear behind lost luggage on the final night. When someone asks about the trip, the airport answers. The clearest peak-end rule examples are this ordinary: a sharp moment and the exit become the story, while most of the hours lose their vote.

In peak-end rule psychology, memory gives extra weight to the moment that felt most intense and to the way the episode ended. Its length, and much of the middle, weigh less than a running total would suggest. Researchers call the length part duration neglect. You are not averaging the hour. You are keeping a couple of snapshots and answering from those.

A cold-water study makes the oddity concrete. One trial was a minute with a hand in water cold enough to hurt. The other trial was that same minute, plus 30 more seconds while the water was warmed slightly. It still hurt, and it hurt less than the minute before. Later, a significant majority chose to repeat the longer trial. The memory with the gentler ending beat the memory that contained less pain.

Peaks and endings are real. A ruined last night is still a ruined night. A cruel final conversation is still something that happened. The bug is letting those snapshots stand in for the episode when you decide whether to go back, quit, recommend, retell, or design the next one. The middle lived through the minutes. The choice is being made by the highlight reel.

The same compression works on good experiences. A glowing last hour can gild a long, flat, or miserable stretch, and a single peak can become the whole vacation in the story you tell. You then choose the sequel from the story.

• "The vacation was a disaster." One ruined final night, six decent days behind it.
• "The relationship was perfect." The ending was tender, and a long unhappy middle has gone soft.
• "I'll never go back to that clinic." The peak was the wait, and the care itself has dropped out of the score.
• "The talk failed." The last question went badly. The forty minutes before it were fine.
• "We should book them again." Launch day felt great, and the months of rework are hard to feel as a sum.
• "That film was wonderful." The ending landed. The sag in the middle is already gone.

The argument: an hour of useful repair ends on one sharp sentence. By the next morning the hour is the sentence. The decision to stop talking is a decision about the exit, wearing the costume of a decision about the relationship.

The procedure: discomfort that eases at the end is remembered as milder than a shorter bout that stops at the worst moment. People then choose, or recommend, from the memory rather than from the minutes they actually paid.

The trip: lost luggage on the last day rewrites a week. The story you tell friends, and the question of whether you would go back, both answer to the airport, not to the days.

The handoff at work: a project with a grim middle and a smooth launch is filed as a win. The next estimate gets built from the launch feeling, so the middle is scheduled as if it had been short.

You repeat experiences that were worse in total because they ended gently, and you abandon experiences that were better in total because they ended awkwardly. Restaurants, clinics, trips, jobs, and conversations all get this treatment. The choice feels like taste. It is often a highlight reel.

Stories harden around the same two frames, so other people inherit your edit. A partner, a colleague, or a customer hears a disaster or a triumph and cannot find the week they lived. Trust thins when the retelling and the calendar disagree.

People who design experiences learn the wrong lesson if they only polish the exit and the peak. Those moments matter. A long careless middle still happened to the person who was in it, even if the memory will underweight it later.

Memory keeps what is intense and what is available. The end is available because nothing came after it to cover it. Duration is a poor thing to feel as a sum: an extra twenty minutes of mild discomfort does not announce itself as "more total" the way a spike announces itself as "the point."

When you choose again, you ask the remembering version of yourself, not the version that was in the middle of the minutes. That version answers with the snapshots it kept. The answer is fluent, so it feels like the experience.

Kahneman, Fredrickson, Schreiber, and Redelmeier's 1993 Psychological Science paper ran the cold-water comparison: 60 seconds at 14 °C, versus those 60 seconds followed by about 30 seconds as the water rose to 15 °C, still uncomfortable and less so. A significant majority chose to repeat the longer trial. The paper's reading is that retrospective evaluations of an unpleasant episode are dominated by the discomfort at the worst moment and at the final moment, with duration playing a small role.

Fredrickson and Kahneman's 1993 paper on film clips found the same shape for pleasant and unpleasant scenes. People rated affect while the clips played, then gave a global rating afterward. Duration's effect on the global rating was small, and smaller still when the rating was made from memory. The global rating behaved like a weighted average of snapshots of the experience, as if length had barely been part of the event.

• The story of the episode is basically the ending, plus one intense scene.
• You cannot say how long the middle lasted without guessing.
• A gentler exit made a longer ordeal seem like the better one.
• You are choosing a sequel from the story, not from a simple log of the days.
• Someone who was there remembers a different episode than the one you are grading.

Keep the ending in the score. Stop letting it be the score.

• Before you decide "never again" or "let's repeat that," name the middle in one plain sentence: what most of the time was actually like.
• If the stakes are a trip, a job, a clinic, or a relationship, jot a few marks while you are in it. Memory will not volunteer the length.
• Ask which choice you would make if the ending had been ordinary and the middle unchanged.
• When you design something other people will endure, fix the peak and the exit, and also ask what the long middle feels like from inside it.
• If your retelling and another person's calendar disagree, treat that gap as data about the edit, not as proof they weren't paying attention.

Should I ignore how things end?

No. Endings are part of what happened, and a bad ending can be a reason on its own. The check is whether you are judging the ending or the episode. If a short log of the middle would change the choice, the snapshots are doing too much of the work.

Against the peak-end rule, the choice has to survive contact with the middle.

The ruined last night

The tender exit

The smooth launch

Practice this pattern in the Reframing App - capture the trigger, label it (like Peak-End Rule), check evidence, and write a more balanced thought.