Pomegra Wiki

Peak-end rule

Peak-end rule is the tendency to evaluate an experience based on its most intense moment (peak) and how it ends (end), rather than on the average experience or total duration. An investment experience of 10 years with an average annual return of 8% but a peak return of 20% and a final year return of 5% is judged by the 20% peak and 5% end, not by the average 8%.

Related to recency bias and selective memory. Discovered by Kahneman and colleagues in studies of pain and pleasure.

The classic experiment

Researchers had subjects submerge a hand in cold water under two conditions:

Condition A: 60 seconds in cold water (painful). Condition B: 60 seconds in cold water, then 30 more seconds in slightly warmer water (still cold, but less so).

Condition B is worse (more total time in cold water), but subjects reported it as less unpleasant. Why? The end (warmer water, less painful) was better in B, and the final moment dominated the memory.

This demonstrates peak-end rule: the overall experience is judged by the peak moment (both equally painful) and the end (B is less painful at the end).

Peak-end rule in investing

Regret about bear markets. An investor holds stocks through a bull market (peak returns are high and positive), but then experiences a bear market (end is negative). She judges the entire experience by the recent bear market, even if the overall 10-year return is still positive.

Example: Portfolio returns are +15% (year 1), +20% (year 2, the peak), +10% (average), …, -10% (year 10, the end). She judges the experience as bad because the end is negative, even though the average is +10% and the total return is positive.

Fixation on lowest points. An investor’s portfolio hit a -30% low in 2008 (the peak pain), but has recovered. Yet, the memory of that -30% low dominates how she feels about the portfolio, even though it has since recovered fully and grown.

Recency in satisfaction. A portfolio that has been strong all year but falls 5% in the final month will feel worse than a portfolio that was weak all year but rose 5% in the final month. The end moment weights heavily.

Peak-end rule and portfolio evaluation

This bias can cause investors to misjudge portfolio performance and satisfaction:

  • A 10-year portfolio with average +10% annual returns but ending down 15% feels like a failure, even though the total return is strong.
  • A portfolio that peaks at $500,000 but ends at $400,000 feels like a loss, even though it is up overall.
  • A portfolio that ends strong (last month is up 10%) feels better than its historical average suggests it should.

Peak-end rule and decision-making

Peak-end rule can bias future decisions. If an investment experience ended badly, you might avoid similar investments, even if the average return was good. Conversely, if it ended well, you might overcommit.

Distinguishing peak-end from recency bias

Recency bias is the tendency to overweight recent data in your forecast. Peak-end rule is the tendency to remember and evaluate past experiences by their peak and end, not their average.

They are related but distinct. Recency bias affects future judgment; peak-end rule affects memory of past experience.

Defenses against peak-end rule

  • Track average performance. Rather than remembering or evaluating by peak and end, track the average annual return or average experience. This is more representative.
  • Separate the experience from the memory. Your memory of a portfolio (dominated by peak and end) is not the same as the actual experience (which is the average). Acknowledge this gap.
  • Use longer time horizons. Over longer periods, peak and end matter less. A 30-year portfolio’s peak and end are less dominating than a 3-year portfolio’s.
  • Reframe the end positively if possible. If a portfolio ends down but you plan to continue investing, reframe: “The end is down, but that means I am buying at lower prices.” This changes the emotional meaning of the end.
  • Use rules rather than memory. A mechanical rebalancing rule or a pre-set asset allocation does not rely on memory of peaks and ends. It operates regardless of how the recent experience felt.

See also

Wider context