A new perspective on weak form efficiency: empirical evidence from the UK bookmaker based betting market

This paper looks for evidence of weak form inefficiency in the British racetrack betting market, in particular the bookmaker-based market. In addition, it seeks to identify market segments, based on the degree of expertise of participants (races run during holiday v.s. non holiday periods), which exhibit a greater degree of market inefficiency. The expected returns show that the favourite-longshot bias exists in the market as a whole and for each market segment. However, the degree of the favourite-longshot bias is significantly higher in the markets dominated by those with less expertise (holiday periods). A conditional logit model incorporating a single predictor, log of normalised starting prices probability, is developed to determine the probability of a horses’ winning for the whole market and for each market segment. A positive return (2.58%) is obtained by employing the model generated probabilities (using data for the whole market) using a Kelly wagering strategy. The positive return can be further increased by limiting the maximum fraction to bet on horses with short odds. Overall these results suggest that the U.K. bookmaker based horserace betting market is weak form inefficient. When the dataset was further split into two market segments based on whether races are held in holiday periods, there is evidence that a much higher positive return (19.21%) can be earned for races held in holiday periods. This offers empirical support for the ‘holiday effect’ and confirms that markets where casual bettors are most likely to dominate demonstrate a greater degree of weak form inefficiency.