Alternation Bias and the Parameterization of Cumulative Prospect Theory

Several authors have recently addressed the question of whether cumulative prospect theory (CPT) resolves the St. Petersburg Paradox (Blavatskyy, 2005; Rieger & Wang, 2006). These authors show that direct application of CPT to the St. Petersburg gamble fails to resolve the paradox under most conventional CPT parameterizations. They also propose a number of remedial fixes to CPT, central among which is a constraint on the value function exponent to be smaller than the probability weighting function exponent (α < γ). As this constraint is violated by most experimentally determined CPT parameterizations,¹ the remedy amounts to a fundamental reparameterization of CPT.