Blog / Risk Choice, Reference Points, and the Difference Between Gains and Losses

Risk Choice, Reference Points, and the Difference Between Gains and Losses

A choice can have the same expected value under two descriptions while its outcomes occupy different psychological domains relative to a starting point.

October 17, 2026 · 4 min read · By Cepingku Editorial Team

When people discuss a risky choice, they often begin with the final amount. That can leave out the comparison point that gives the amount meaning. In prospect theory, Kahneman and Tversky describe value as attached to gains and losses relative to a reference point, rather than simply to final wealth. Their 1979 paper is a description of experimental choice patterns. It does not supply a personal investing rule, a fixed loss-aversion multiplier, or a forecast of what any one person will choose.

Start by naming the reference point

Consider two fictional options expressed as future final balances. Option A ends at 50 for certain. Option B ends at 100 with a 50% chance and at 0 with a 50% chance. Both have an expected final balance of 50, but they do not offer the same distribution. With a starting balance of 0, A is a certain gain of 50 and B is an uncertain gain: either 100 or 0. With a starting balance of 50, A is no change, while B includes a possible loss of 50 and a possible gain of 50. The numbers have not changed; the comparison point has.

That distinction also prevents a common accounting mistake. If the 50 in Option A is described as an added payment rather than a future final balance, a person who already has 50 ends at 100, not at no change. Before calling an outcome a gain or loss, write down whether each number is a final balance, a change from today, or a change from another stated baseline.

Question to checkWhy it changes interpretation
What balance or condition is the comparison point?The same final balance can be a gain, a loss, or zero change relative to different starting points.
Are the amounts final balances or added changes?A final balance of 50 and an additional 50 have different implications for someone who starts at 50.
Are the options being compared by expected value alone?Equal expected values do not make their possible outcomes, certainty, or psychological domain identical.

A symmetric teaching example

A second fictional set makes the domains explicit. In the gain domain, compare a certain +50 with a 50% chance of +100 and a 50% chance of 0; each has an expected change of +50. In the loss domain, compare a certain -50 with a 50% chance of -100 and a 50% chance of 0; each has an expected change of -50. These are deliberately symmetric teaching examples. They do not mean that gains and losses feel symmetric, nor do they establish how a particular person should act.

Kahneman and Tversky called the group-level pattern in which preferences can reverse across those domains the reflection effect: people often show risk aversion for gains and risk seeking for losses. The wording matters. “Often” describes a tendency found in the research setting, not a prediction for an individual, a specific amount, or a real-life decision. A person may also respond differently when the stakes, time horizon, information, or personal constraints change.

Expected value is one property, not the whole choice

Expected value summarizes a probability-weighted average. It does not state how likely each outcome is, whether one outcome is certain, or which reference point a person has adopted. The same paper also discusses certainty and isolation effects: features shared by options, or the way a choice is presented, can affect preferences. That is a reason to slow down before turning one answer into a stable trait or treating a single calculation as a complete decision method.

For a practical reflection, take one choice you are considering and make two columns. In the first, state the current reference point and label every outcome as a final balance or a change. In the second, list the possible outcomes and their probabilities without collapsing them into an average. Then ask which part of the description feels most salient: keeping what is already present, avoiding a drop, securing a gain, or preserving flexibility. This exercise clarifies the framing of a choice; it is not investment advice.

Use a questionnaire as a prompt, not proof

The Risk Preferences Reflection can help you notice how you answer questions about uncertainty, constraints, and pressure. It does not provide financial advice, predict returns, or establish that this site’s questionnaire has the validity of the research discussed here. Treat a result as a prompt to revisit your examples and assumptions, especially the reference point you may have left unstated.

Want to see where you land?

Take the Risk Preferences Reflection and get your results in minutes.

Take the Risk Preferences Reflection →

← Back to the blog