When evaluating a past event—like a vacation or a medical procedure—our brains don't average out the entire experience. Instead, we heavily judge it based on two points: the most intense moment (the peak) and the very end. In a famous experiment, patients undergoing a painful procedure rated the overall pain as lower if the doctor added an extra, slightly less painful minute at the very end.
The Snapshot Nature of Memory
Human memory does not function like a continuous video recording. When looking back on an episode from the past—whether an entire vacation, a meal at a restaurant, or an uncomfortable medical exam—the brain does not sum up the sensations of every second, nor does it calculate a mathematical average of the total time spent. Instead, the retrospective evaluation of an event is overwhelmingly shaped by two specific reference points: the moment of peak intensity and the final moments before the experience concluded. This psychological pattern is known as the peak–end rule.
A central consequence of this cognitive heuristic is a phenomenon called duration neglect. When people evaluate how pleasant or unpleasant a past experience was, the actual length of time the event lasted has remarkably little influence on their final judgment. An ordeal that lasts twice as long as another may be remembered as significantly less severe if its worst moment was mild and its closing seconds were relatively calm. Conversely, a brief event can leave a lasting negative impression if it ended abruptly on a painful note.
The Cold Water Experiment
The empirical foundations of the peak–end rule were established in the early 1990s by psychologists Daniel Kahneman, Barbara Fredrickson, Charles Schreiber, and Donald Redelmeier. One of the classic early demonstrations involved a laboratory experiment using the cold pressor test, where participants placed their hands in painfully cold water under controlled conditions.
In this study, participants were subjected to two distinct trials in a counterbalanced order. In the short trial, subjects kept one hand submerged in water held at 14 degrees Celsius for 60 seconds, which caused a noticeable and increasing level of discomfort. In the long trial, subjects kept their hand in water at 14 degrees Celsius for the same initial 60 seconds, but were then kept in the water for an additional 30 seconds while the temperature was subtly raised to 15 degrees Celsius. While 15-degree water is still cold and uncomfortable, it is measurably less painful than 14-degree water.
Objectively, the longer trial contained strictly more physical discomfort: it included the exact same 60 seconds of severe cold, followed by another half-minute of mild discomfort. Yet when participants were asked which of the two trials they would prefer to repeat for a third round, the majority chose the longer trial. Because the longer episode ended on a slightly more tolerable note, the remembered experience was judged to be less unpleasant overall, overriding the objective fact that it exposed them to extra seconds of pain.
Clinical Testing in Medical Procedures
To test whether this heuristic operated in high-stakes, real-world environments, researchers examined patients undergoing painful medical procedures, specifically colonoscopies and lithotripsy. In a series of clinical investigations led by Redelmeier and Kahneman, patients recorded their real-time pain levels at fixed intervals throughout their procedures using a handheld device.
After the procedures were completed, patients provided an overall retrospective evaluation of the total pain they had experienced. The findings mirrored the laboratory tests: retrospective ratings of pain did not track the total duration of the procedure, nor did they reflect an integration of all the pain recorded across time. Instead, the retrospective evaluation was strongly predicted by a simple average of the single most intense moment of pain (the peak) and the pain reported during the final minutes (the end).
Building on this insight, the researchers designed an experimental intervention. For a random subset of patients, physicians left the colonoscope stationary inside the body for approximately three extra minutes before removing it. Leaving the instrument stationary was mildly uncomfortable, but far less painful than the active maneuvering of the examination. Even though these patients spent more total time on the examination table and endured additional total discomfort, their retrospective ratings of overall pain were substantially lower than those of the control group. Furthermore, these patients were more willing to return for necessary follow-up screenings in later years.
Two Selves: Experiencing vs. Remembering
The peak–end rule highlights a fundamental divide in human cognition: the distinction between what Daniel Kahneman described as the experiencing self and the remembering self. The experiencing self lives in the present moment, feeling the continuous stream of sensory inputs as they unfold second by second. This self answers the question, 'How does it feel right now?'
The remembering self, on the other hand, is the retrospective narrator that keeps score and reflects on past events. It answers the question, 'How was it on the whole?' The remembering self does not have the capacity or the need to preserve every second of lived experience. Instead, it constructs simplified, highly compressed stories out of key highlights. Because decisions about the future are based on memories rather than real-time sensations, human behavior and future choices are governed almost entirely by the remembering self.
The Rule in Positive Experiences
While much of the early foundational research centered on pain and physical discomfort, the peak–end rule applies equally to positive and pleasurable experiences. How people evaluate a holiday, a film, a musical performance, or an interaction with a service provider is largely driven by the highest emotional peak and the final impression left behind.
In evaluations of material gifts or positive experiences, adding a small, low-value item at the conclusion can paradoxically diminish the overall perceived quality of the gift. Similarly, a vacation that features wonderful attractions can be severely downgraded in memory if it concludes with a frustrating airport delay or a difficult journey home. In contrast, an average experience that finishes on an exceptional high note is frequently recalled as an overall success.
Evolutionary Logic and Cognitive Limits
From an evolutionary perspective, the peak–end heuristic provides a highly efficient shortcut for memory storage. Storing an exhaustive, moment-by-moment log of every past encounter would demand immense cognitive resources. For survival, the most critical pieces of information about an environment or an organism are its extremes—the moments of maximum danger or reward—and the state of safety or risk when the encounter finished.
However, relying on this shortcut introduces notable cognitive biases. By neglecting duration, individuals may make choices that minimize peak discomfort at the cost of prolonged low-grade suffering, or choose shorter bursts of joy over extended periods of moderate contentment. Recognizing that memory edits experiences down to their peaks and endings helps clarify why retrospective evaluations so often conflict with the lived reality of time.
Key takeaways
•Retrospective judgments of past events are dominated by the most intense moment (the peak) and the final moment (the end), rather than the average or sum of the entire experience.
•Duration neglect causes people to largely ignore how long an event lasted when evaluating its overall quality or discomfort in hindsight.
•Human choices are guided by the remembering self—which relies on compressed heuristic snapshots—rather than the moment-by-moment experiencing self.