Bounded Rationality in Esports

This dissertation examines decision-making under bounded rationality in esports. The thesis contains five studies, one presenting the benefits of esports data for economics research, one testing the validity of the HLTV ranking system, and three investigating a manifestation of decision-making under bounded rationality, namely the hot hand, the honeymoon effect, and the incentive effect of ex-ante known heterogeneity in abilities. In the first study, I show that esports economics research is rare and how economics research can benefit from using esports data. In the second study, we validate the HLTV ranking system, widely adopted in Counter-Strike: Global Offensive (CS: GO), demonstrating its ability to order teams by their current strength. In the third study, we find no evidence of the hot hand in Kovaak’s aim trainer and mixed evidence in professional matches of CS: GO, suggesting that sufficiently high incentives may be required for the hot hand. The fourth study shows a clear honeymoon effect of a roster change on the performance of CS: GO teams. Team performance tends to decline gradually afterward without a recovery. Finally, we show that both favorites and underdogs exert less effort in CS: GO matches as the ex-ante known heterogeneity in abilities increases. We, therefore, find evidence for an incentive effect predicted by tournament theory. Methodologically, we differentiate between two approaches to define the hot hand and control for potential cooling effects of flow disruptors by incorporating the temporal dimension into permutation tests. We also demonstrate the utility of aim trainers for testing human behavior.