Probability forecasts of binary events are often gathered from multiple models and averaged to provide inputs regarding uncertainty in important decision-making problems. Averages of well calibrated probabilities are underconfident, and methods have been proposed to make them more extreme. To...