An attempt by Amos Tversky and Daniel Kahneman to make decision theory less rational, but more realistic as a description of humans, than ordinary expected utility.
An agent based on prospect theory:
- is loss-averse, and so weighs losses (relative to some reference point) more strongly than gains
- is risk-averse with respect to gains, but risk-seeking with respect to losses
- uses probability weighting: small probabilities count more and larger probabilities count less than they would under expected utility.
See also