Ray wrote,

"The real problem I see so far is Martin's associating probabilistic counting rules with proposed groupings of zodiac signs. With a deck of cards, it's unassailable--but can we really assume that every possible combination of every zodiac feature in every position is equally likely? If not, as you argue, the odds are less.

The deeper concern I have as a research designer is to stall the statistical analysis short of the established measuring tools: correlations, regressions, and direct comparisons between means."

Ray, does what you are saying here have to do with sample size? Because if it does, this is what I'm thinking is going to be an inherent limitation of any similar analysis at GT. We have little to work with.