Motivation: In some situations, forecasters are rewarded based on their performance relative to other forecasters. For example, the forecaster who attains the lowest mean squared error may receive a reward. While this setup seems natural, it is more complicated than the traditional case where a forecaster simply aims to optimize their own expected performance.

Goal of the thesis: Analyze the above situation, incorporating the existing literature on the topic. The thesis likely involves a mix of (decision and/or statistical) theory, simulations and empirical analysis.

Literature (starting point):
Pfeifer, P.E., Grushka-Cockayne, Y., Lichtendahl Jr., K.C. (2014): The Promise of Prediction Contests. American Statistician 68, 264-270, 2104. Link