In a series of preliminary tests, we examined the pattern of mean daily returns throughout the lunar month, including visual inspections of return histograms. This examination reveals one interesting regularity. We observe that high returns tend to cluster around the new moon date, while low returns tend to cluster around the full moon date. Following this observation, we structure our returns tests to reflect the possible difference between new moon and full moon periods. Specifically, most of our tests are simple comparisons of mean daily returns for various return windows centered on the new moon and.
