Testing effeciency in Gambling markets: a comment

Publisher Website
Google Scholar
Abstract
In a recent article in this journal, Even and noble point out that the regression-based test for unbiasedness in point spread betting markets is not valid test when the mean and median of the market's forecast errors are different. In the case where skewness is present, Even and noble propose a likelihood ratio statistic to test for market efficiency. In this comment, three concerns are explored: (i) whether distributions of market forecast errors in these markets are consistently and significantly skewed; (ii)whether the test proposed by Even and Noble suffers from the same problem of low power as do other statistical tests in these markets;(iii)whether the Even and Noble test statistic is actually a new test or simply the equivalent of very simple, well-known statistical test.