On the efficiency of the pesäpallo betting market

The thesis is the first to study the (weak form) efficiency of pesäpallo betting market. Analysed data sample covers Superpesis (the top men pesäpallo league in Finland) seasons 2015-2020 (1394 matches) and related average odds obtained from multiple bookmakers. Following research questions were formulated: do probabilities of match outcomes derived from odds systematically differ from probabilities calculated afterwards using actual results and is it possible to develop a betting strategy producing consistently positive returns by analysing odds? The latter is a decisive criterion for the market efficiency. A difference between predicted and obtained probabilities was tested for a time period of at least a whole season with two different methods: a multinomial logistic regression and odds grouping. The former gives results on an aggregate level and latter makes possible to study wanted odds level. Results were consistent: some statistically significant deviation exists, but it cannot be categorized as a systematic (e.g. only for the one season, 2017, difference is shown to be significant by regression model). On the other hand, the betting strategy which produced positive returns each season (the obtained return over all seasons is statistically significant) when applied during a first half of season is presented. Interestingly, the strategy is not working during a second half. Thus, the main outcome of the thesis is that the pesäpallo betting market is not efficient. However, further research is needed to find the source of inefficiency. There are signs that clear favorites are yielding very well during a first half of seasons, but the issue has not been studied in details due to time limitations. The clear direction for further research would be to concentrate on analysing separately different time periods during a season