Motivation

The best known stylised fact to emerge from the area of research reflected in this volume is that, in horse and dog betting, financially superior returns (i.e. smaller losses) accrue to a strategy of wagering on short-odds rather than long-odds runners. This bias is sufficiently pronounced that expected returns may even be positive where one bets only on extreme favourites. The evidence for the existence of this bias in the odds dimension is impressively voluminous and has accumulated over more than five decades.

The existence and persistence of such ‘longshot bias’ represents an anomaly when viewed within the tradition of treating wagering markets as examples of financial markets. According to the efficient markets hypothesis, prices (odds) should reflect all available information relevant to the outcome of a race. With all agents risk-neutral, expected returns, in equilibrium, would then be the same for all possible bets and so invariant with respect to odds. Further, it is more common in economics to assume risk aversion rather than risk-neutrality and in this case one would predict that bets at short odds should yield lower returns, on average, than bets at long odds. Certainly a premium to risk rather than its reverse is what is observed in most financial markets other than betting ones.

Early attempts to explain the ‘longshot anomaly’ within betting were typically based on attributing particular tastes to a representative bettor such that he would be willing to accept worse financial returns when betting on outsiders than when betting on favourites.