Biases from omitted risk effects in standard gamble utilities.

Utilities elicited under the "probability equivalence" version of the standard gamble procedure systematically exaggerate the utility increments of expensive health interventions that bring the recipient towards full health. Switching to the alternative "certainty equivalence" version would introduce the reverse bias. The biases arise from anticipated pre- and post-decision risk effects that expected utility theory omits. To include these risk effects, partition the future by changes in knowledge into: (1) pre-choice periods; (2) the pre-outcome future before the chosen act's outcome becomes known; and (3) the more distant post-outcome future, when that outcome will have become known.