Wisconsin9 wrote:
Look at the air. Do you see ice cream? If the answer is no, then it is not. Look at the show. Do you see roughly equal numbers of males as there are females? If the answer is no, then there are not. This is basic. If X does not equal Y, then X does not equal Y.
Well, I mean, this math isn't really relevant to the question I think. Since we have never seen the full count of populations x and populations y, what we have are only samples z and c, say.
The question then isn't; "Is population x equal to population y?" but; "Given samples z and c what is the probability that population x is greater than population y?" Of course, this question, while more accurate, stil fails to address the question of biased sample distribution (such as would be the result of non random sample selection for example, see mane 6).
Even more problematicaly, we can not establish the size of our samples as it is impossible to verify with any accuracy which background ponies are the same pony. As a result, the probability of counting one pony more than once is very high. One might plausibly correct for this by counting only ponies in a single frame as it should be a reasonable assumption that no pony can be in two places at the same time. However, this will decrease the sample size and increase the likelyhood of selection bias (though selection is of course already non-random).
Then there is of course the problem of ponies that demonstrably can be in several places at once, namely Pinkie Pie.

Cute.
Those were some really weird pictures.







