Erasmus university rotterdam
Download 1.62 Mb.
|
- Navigate this page:
- The early evidence
- Favorite longshot bias in other sports
Literature reviewThis section contains a literature overview which is divided in four parts. The first part goes back to the early literature about the favorite longshot bias. After the history, the literature about the favorite longshot bias in other sports is discussed. The third part contains literature of the favorite longshot bias in tennis. As the favorite longshot bias is not found by everyone, in the last part the contradicting literature is discussed. To have a clear overview of the articles which provide insight for the research, the tables 1 to 4 are used. This tables give a summary of the articles by writing the author(s), date, title, research question, time period and outcomes of the article. In some papers there are more outcomes than explained in the tables, but those outcomes are not directly relevant for this study. This section goes back to the history of the favorite longshot bias which is mainly based on the horse racing sport. The very first who came up with the favorite longshot bias was Griffith (1949). He researched if bettors at horse race tracks were able to predict the exact probabilities of a horse winning the race. The data he used were horse race tracks in the United Kingdom in 1947. He found out that the bettors tend to overvalue the longshot horses and they undervalue the favorite horses. In 1956 McGlothlin did a similar research like Griffith, which was also based on the behavior during horse racing. He looked at the behavior of bettors during race tracking, especially during the day. The same favorite longshot bias as Griffith (1949) was found. The results of his research also show that betting on favorites becomes more valuable during the day. Almost 20 years later, in 1977 Ali also investigated the favorite longshot bias, but he used far more data than Griffith (1949) and McGlothlin (1956). Ali’s dataset was based on races from 1970 till 1974 which contains 20.507 races in total. There was also another advantage comparing to Griffith (1949). Griffith worked divided the groups based on odds, which means that there was a chance that two competing horses were in the same group. In the research of Ali there was no possibility that two competing horses were in the same group. The results of the investigation of Ali shows a favorite longshot bias in every year (1970-1974). A few years after in 1982 Arsch and Malkiel also did research on horse racing. The research question they wanted to answer was: Do all horses go off at odds that reflect their true probability of winning or is there a systematic tendency to over bet longshots and under bet favorites? And they came to the same conclusion as their predecessors: the favorite longshot bias exists. The odds are a good indication of how horses will finish. Table 1: Literature summary ‘The early evidence’. The table below summarizes the academic articles that are discussed in 2.1 The early evidence.
After all the research which was done on the favorite longshot bias in horse race tracking, from the year 2000 there came more research on the favorite longshot bias in other sports. In this section those articles are discussed. The articles based on the tennis sport are discussed in section 2.3. In 2000 Cain et al. did research about the favorite longshot bias in UK football. Therefore the research question of this study was: Does the favorite longshot bias appear in UK football? The study shows that the same favorite longshot bias is found as in horse racing. The bias appears in betting on the results of the game (home win, away win or a draw), but also In betting on specific scores. Shmanske (2005) did examination of the extent to which the odds predict the outcome. He studied the golf-betting market, but did that with a small dataset. The data only contains the PGA TOUR events for the 2002 season. There was evidence found for a positive favorite longshot bias; betting on favorites will pay the highest return to bettors, but because of the small dataset size, the results are doubtful. Vlastakis, Dotsis & Markellos (2008) did an assessment of the international efficiency of the European football betting market by examining the forecast ability of match outcomes. The data contains the European football league games from 2002 to 2004. In the results of the study a favorite longshot bias was found. Three years later, Andrikogiannopoulou & Papakonstantinou (2008) also investigate football games. They examine whether the prices set by bookmakers on large number of soccer events are efficient. The dataset they used contains 10.000 soccer matches. Besides the match results, they also looked at the behavior of bettors. A favorite longshot bias was found for the matches in the years from 2005 to 2009. They also found a favorite longshot bias looking at the outcomes of the soccer matches. For example the winning of home and away games of favorites were underbetted. Looking at the results of the behavior of bettors, they found out that only 2% of the bettors constantly bet on the favorite, while 6% of the bettor bet on the underdog. Table 2: Literature summary ‘Favorite longshot bias in other sports’. The table below summarizes the academic articles that are discussed in 2.2 Favorite longshot bias in other sport.
|