Probability Weighting Function

Transform objective or stated probabilities into subjective decision weights, typically overweighting small probabilities, underweighting middle and high probabilities, and treating the zero and certainty boundaries nonlinearly.

Core Idea

A probability weighting function maps an objective or stated probability p into a subjective decision weight π(p) before a chooser aggregates uncertain outcomes. In expected utility, probability enters linearly. In prospect-theoretic and related models, the effective weight need not equal the probability. Small probabilities are often overweighted, middle and high probabilities underweighted, and the transitions from impossibility to possibility or from uncertainty to certainty can receive disproportionate importance.

The function is a choice representation, not necessarily a report of believed frequency. A person can state a probability accurately yet behave as though it carries a different decision weight. The probability transform is also separate from the value function over gains and losses. This separation lets analysts distinguish attraction to a small chance from curvature or asymmetry in the outcomes themselves.

Structural Signature

• The uncertain outcomes — mutually exclusive consequences represented by the chooser.
• The objective or stated probability — p attached to an outcome or rank.
• The transform — a function mapping p to decision weight π(p).
• Non-identity — decision weight generally differs from raw probability.
• Curvature and elevation — shape determines which probability regions are amplified or compressed.
• Boundary behavior — changes near zero and one can differ qualitatively from equal changes in the interior.
• Separation from outcome value — probability weighting and the value function remain distinct components.

What It Is Not

Probability weighting is not simply risk aversion. Concave utility can produce risk-averse choice while probability remains linear; nonlinear weighting can produce risk-seeking or risk-averse patterns depending on the probability region and outcomes. It is not loss aversion, which compares gains and losses around a reference point. It is not ambiguity aversion, which concerns uncertainty about the probabilities themselves.

It is also not calibration. Calibration compares stated confidence with long-run frequencies. A well-calibrated person can still assign nonlinear decision weight in choice, and a poorly calibrated forecast does not by itself identify the weighting function.

Scope of Application

• Lotteries and gambling: small chances of large gains can receive disproportionate decision weight.
• Insurance: small probabilities of severe loss can motivate premiums above expected loss.
• Medical decisions: rare side effects or treatment successes influence choice nonlinearly.
• Safety and public policy: low-probability catastrophic hazards compete with common moderate harms.
• Consumer warranties: a small failure probability can receive more choice weight than its frequency alone predicts.
• Experimental choice: common-ratio and common-consequence designs identify departures from linear probability weighting.

Clarity

The abstraction decomposes a risky choice into two transformations: how outcomes are valued and how probabilities are weighted. Without that distinction, any rejection of an expected-value gamble can be called “risk aversion,” hiding whether the behavior came from outcome curvature, loss asymmetry, probability distortion, or their interaction.

Manages Complexity

A single function organizes a family of observations across probability regions. Its curvature captures sensitivity to changes in p; its elevation captures generally optimistic or pessimistic weighting; its boundary treatment captures possibility and certainty effects. Researchers can compare functional forms and parameters instead of treating every choice reversal as an isolated bias.

Abstract Reasoning

Replacing p with π(p) changes marginal probability value. The increase from 0 to 0.01 may receive more effective weight than the increase from 0.40 to 0.41, and the increase from 0.99 to 1 may receive a special certainty premium. Equal objective increments therefore need not be equal decision increments.

The function also supports separation tests. Hold the probability pattern fixed while varying outcomes to identify value curvature; hold outcomes fixed while comparing probability ratios or boundary moves to identify probability weighting. The model is useful precisely because it prevents one component from absorbing every departure from expected utility.

Knowledge Transfer

Within judgment and decision making, the same probability-space analysis transfers across gambling, insurance, medicine, consumer behavior, and public risk. A designer can ask whether a response reflects the size and sign of outcomes or the region of probability space in which the choice is framed. The named abstraction remains human-choice-specific even though nonlinear transformations are mathematically portable.

Examples

Small-probability weighting

A person buys a lottery ticket with negative expected value and also purchases insurance priced above expected loss. The outcomes differ in sign, but both decisions can be supported by high decision weight on a small probability, interacting with gain and loss values.

Certainty boundary

A chooser prefers a sure outcome to a slightly larger outcome with very high probability, yet reverses the preference when both probabilities are scaled down proportionally. The disproportionate value of eliminating the final uncertainty is the Certainty Effect species.

Structural Tensions

• Belief versus preference: observed choice weights do not uniquely identify subjective probability beliefs.
• Functional-form dependence: different weighting functions can fit the same limited choice set.
• Outcome interaction: value curvature and probability curvature can trade off in estimation.
• Rank dependence: weights in cumulative prospect theory depend on outcome rank and cumulative probabilities, not isolated event probabilities alone.
• Description and experience: probabilities learned from experience can produce patterns different from explicitly described risks.

Structural–Framed Character

The function has a structural mathematical form, but the named abstraction is domain-specific because its roles are subjective decision weights in human choice under risk and its evidence comes from behavioral choice patterns.

Structural Core vs. Domain Accent

Remove lotteries, choosers, gains, and losses, and a generic nonlinear mapping remains. That mapping alone does not preserve the abstraction's explanatory role. The domain accent—probability representation, decision weighting, and separability from outcome value—is constitutive.

Instantiates / Related Primes

• Expected Utility is the linear-probability benchmark against which the transform is identified.
• Certainty Effect is the strict p=1 boundary species.
• Risk Aversion concerns outcome-value curvature and must be estimated separately.
• Loss Aversion concerns reference-dependent asymmetry over outcome value.
• Calibration concerns forecast-frequency correspondence, not choice weights.

Relationships to Other Abstractions

Current abstraction Probability Weighting Function Domain-specific

Parents (3) — more general patterns this builds on

• Probability Weighting Function is a kind of Function (Mapping) Prime

  A probability-weighting function is a function mapping specialized to probability inputs and subjective decision-weight outputs.

• Probability Weighting Function presupposes Probability Prime

  Probability weighting presupposes numerical probabilities as the inputs whose decision impact it transforms.

• Probability Weighting Function is a decomposition of Nonlinearity Prime

  Removing human-choice and decision-weight framing leaves a nonlinear input-output mapping whose increments have probability-region-dependent effects.

Children (2) — more specific cases that build on this

• Certainty Effect Domain-specific is a kind of Probability Weighting Function

  Certainty Effect is the p=1-boundary species in which the subjective transform assigns a disproportionate increment to eliminating the final probability of failure.

• Possibility Effect Domain-specific presupposes Probability Weighting Function

  The Possibility Effect is defined as the small-probability branch of a Probability Weighting Function and cannot be stated without the objective-probability to decision-weight map.

Hierarchy paths (4) — routes to 4 parentless roots

• Probability Weighting Function → Function (Mapping)

Show alternative paths (3)

Not to Be Confused With

• Expected Utility uses raw probabilities as weights.
• Risk Aversion is curvature in outcome utility.
• Loss Aversion is unequal sensitivity to losses and gains around a reference point.
• Ambiguity Aversion concerns missing, imprecise, or second-order probability information.
• Calibration evaluates probability judgments against frequencies.

References

<!– TODO: Claude re-authoring pass must add and verify primary prospect-theory and probability-weighting sources. –>

Notes

Created from workspace/mixed_dag_2026/missing_node_adjudications/probability_weighting_function.yaml. Identity and hierarchy are adjudicated; final voice and citations remain editorial tasks.