Jeez, glad I did not read most of this thread. But read page 7 where topseed wrote:
As a rule of mathematics, you don't use a very broad-based mean for a specific analysis with different factors as it will almost invariably give you a wrong answer. For example, the mean high temperature in Los Angeles on any given day - based on hundreds of readings - is 78 degrees Fahrenheit. Great.
The general mean high temperature means literally nothing in terms of the actual high temperature on February 15, or August 15. In fact, it is almost certainly wrong in predicting the actual high temperature on either day. The much better way to calculate is to derive the mean high temperatures for February 15 and August 15.
That fact matters because you are applying a general ratio - no matter how accurate, we can agree it's nothing more than a general statistic for an entire population - to a very, very discrete population that has almost no similarity to the population at large in terms of gender, age, height, weight, BMI, and on and on.
What effect do these differences have on the number of gay players in the NFL? I have no idea. Neither does topseed. But using the ratio for the general population simply makes no sense and almost certainly provides the wrong answer.