Retrospective-evaluation heuristic
Peak-End Rule
Retrospective evaluations of an episode can be disproportionately influenced by its most intense moment and its ending.
remembered value approximates mean(peak, end)
This is a descriptive approximation, not a universal estimator. Duration, trend, goals, anticipation, and later narration can all matter.
The timeline keeps total momentary experience visible while the remembered summary gives extra weight to the peak and end.
(%)
The animation runs automatically, pauses on the conclusion, and then repeats. The main control changes the scenario rather than scrubbing the timeline.
- CHANGE
- Ending quality
- WATCH
- remembered evaluation
- MEANING
- The timeline keeps total momentary experience visible while the remembered summary gives extra weight to the peak and end.
The whole timeline and the remembered episode can disagree.
A moving experience trace feeds two summaries: time-average experience and a peak-end reconstruction.
What it actually says
Research on painful and affective episodes found that later global judgments sometimes gave little weight to duration and more weight to salient snapshots.
The result concerns memory and evaluation, not a command to manipulate endings. Ethical design must still consider every moment and total burden.
"A useful law compresses a pattern. It does not erase the conditions that make the pattern true."
How the idea developed
The modern form emerged through observation, argument, and later refinement. The timeline separates the first insight from the version now used in textbooks and practice.[1]
Fredrickson and Kahneman report duration neglect in affective episodes.
Kahneman and colleagues show preference for a longer painful episode with a better end.
Research tests boundary conditions in health, service, and experience design.
How the pattern works
The relation becomes useful only when its mechanism, measurement process, and operating range are visible.
Salient moments are more retrievable.
The final state frames episode closure.
Global judgment is built rather than replayed.
This is a descriptive approximation, not a universal estimator. Duration, trend, goals, anticipation, and later narration can all matter.
Where it earns its keep
Applications are strongest when the law changes a decision, measurement, model, or experiment rather than merely providing an analogy.
Separate experienced and remembered burden
ApplicationCollect continuous and retrospective measures.
Do not discount total pain.
Repair endings without ignoring the middle
ApplicationClosure affects later recall.
Track duration and cumulative cost too.
Where it stops working
Peak-end weighting varies with attention, episode structure, delay, expertise, and the question asked.
"Only the ending matters"
Better: The peak, duration, trend, and goals may matter."Make bad experiences longer"
Better: The classic result is not a general design recommendation.Sources and further reading
Original publications and serious secondary scholarship are prioritized over summaries.
- Fredrickson and Kahneman - Duration NeglectFoundational 1993 experiments.https://pubmed.ncbi.nlm.nih.gov/8355141/
- Kahneman et al. - When More Pain Is Preferred to LessThe better-ending experiment.https://doi.org/10.1111/j.1467-9280.1993.tb00589.x
- Redelmeier and Kahneman - Patients Memories of Painful Medical TreatmentsField evidence and clinical context.https://doi.org/10.1097/00001648-199607000-00002