A stock market speculation scheme proposed by Sidney S. Alexander was tested and was found to be more profitable than would be predicted on the assumption that stock prices perform a random walk with a Gaussian distribution and linear trend. The difference between empirical and theoretical gains was found to be statistically significant. By means of an F-ratio, it was demonstrated that the statistical estimate of the variance of the population of price fluctuations tends to increase as the size of sample is increased. This is consistent with the assumption that the population actually has an infinite variance, but is not consistent with the behavior which would be predicted by any variety of Gaussian random walk. The sophisticated quantity theory of money was found to be unable to account for any non-random movements in stock market price indices. / Business, Sauder School of / Graduate