Empirical Study of Information Efficiency in Betting Markets: Evidence from Men's Professional Tennis

Kenneth Johansen

Student thesis: Master thesis


This paper investigates the degree of information efficiency in the sports betting market on men’s professional tennis based on a sample of quoted odds on 45,813 matches played across 18 years. More specifically, it will analyze whether it is possible to obtain abnormal profits based on inefficiencies from prior studies.
Initially, the paper presents an overview of three biases identified in the previous literature. The first is the favorite-longshot bias which implies systematically betting on favorites yield higher profits than betting on longshots. The second is arbitrage opportunities where bettors can take advantage of quotes from multiple bookmakers on the same match to place bets that yield an almost risk-free profit. The third is the theory on quasi-arbitrage opportunities where differences in bookmakers’ quoted odds are used to quantify the disagreement between bookmakers and identify a group of more profitable bets. In theory, bets on outcomes with higher levels of disagreement should yield higher profits than bets with less disagreement.
Thereafter, the empirical analysis determines whether it is possible to formulate simple betting strategies that are able to generate abnormal profits based on the biases presented earlier. The analysis finds that the favorite-longshot bias is present in the sample and that it can be utilized to obtain a statistically significant abnormal profit. Furthermore, several instances of opportunities for arbitrage profits are found, but it fails to identify a profitable group of bets based on the theory on quasi-arbitrages.
The ability to formulate betting strategies that are able to generate abnormal profits solely based on historical prices and profits are then concluded to be a clear violation of the conditions necessary for a market to be classified as weak-form efficient.

EducationsMSc in Applied Economics and Finance, (Graduate Programme) Final Thesis
Publication date2020
Number of pages76