A Sequential Game Model of Sports Championship Series: Theory and Estimation

Abstract

Ferrall acknowledges support from the Social Sciences and Humanities Research Council of Canada. Helpful comments on an earlier draft were provided by Dan Bernhardt, John Ham, Michael Veall, and seminar participants at McMaster and the Carnegie Mellon-Pittsburgh Applied Micro Workshop.

Leading-Effect vs. Risk-Taking in Dynamic Tournaments: Evidence from a Real-Life Randomized Experiment

“…In all speci…cations, the order of the advantage is never signi…cant. Our …ndings using team sport data are consistent with Ferrall and Smith (1999). In contrast, papers using data on individuals as Malueg and Yates (2010) and Genakos and Pagliero (2012) …nd evidence of strategic e¤ects.…”

“…In contrast, papers using data on individuals as Malueg and Yates (2010) and Genakos and Pagliero (2012) …nd evidence of strategic e¤ects. A possible explanation for this di¤erence between teams and individuals is provided by Ferrall and Smith (1999) who argue that incentives within teams may attenuate incentives between teams. Our results suggest that if the setting is known by participants and individuals are competing in teams the timing of symmetric advantages seems irrelevant.…”

“…We bene…t from a randomized natural experiment in soccer knock-out competitions with two games in which each team is randomly drawn to play either the …rst or the second game at home. Given the robustly identi…ed result that home teams have an advantage (Clarke and Norman, 1995;Ferrall and Smith, 1999;Neave and Wolfson, 2003;Pollard, 1986), we investigate whether the random allocation of this advantage in either the …rst or second game has an impact on the probability of winning the knock-out. The setting of a knock-out allows us to go beyond what can usually be done in empirical work on selection e¢ ciency in corporate tournaments where many variables are not observable and the data is not available.…”

“…Regarding the leading-e¤ect, Malueg and Yates (2010) and Ferrall and Smith (1999) study the existence of strategic e¤ects in dynamic tournaments using individual and team level sports data, respectively. Malueg and Yates (2010) …nd that players in best-of-three tennis tournaments strategically adjust e¤orts across sets conditional on the intermediate score.…”

“…Consistent with empirical evidence, the winner of early districts is endogenously more likely to win later districts than the loser. A possible explanation for this mixed evidence is that incentives within teams may attenuate incentive e¤ects across games which might explain why strategic e¤ects are present in settings where individuals rather than teams compete (Ferrall and Smith, 1999). In addition to e¤ort choices, in many situations agents may also choose risk to in ‡uence their performance.…”