Implication and statistical correlation can be the same when what's implied is a likelihood instead of a certainty.

I can't tell if I disagree with you in a substantive way or just in your word usage (i.e., semantics). Can you please translate this assertion into math?

The vast majority of intelligent people - educated, knowledgeable people - were once theists of one sort of another. The fact that significantly more of them were atheistic/antitheistic than the general population does not change that choosing one at random was still grossly unlikely to produce an AT/AnT.

So, in other words, you mean precisely what Cyan and I had assumed you meant, but you refuse to acknowledge that the word "correlation" has an unambiguous and universal meaning that differs greatly from your usage of it; if you persist in this, you will misinterpret correlation to mean implication where it does not.

For example, smoking is correlated with lung cancer, but a randomly chosen smoker probably does not have lung cancer.

I don't know what else to say on this topic, other than that this is not a case of you being contrarian: you are simply wrong, and you should do yourself the favor of admitting it.

ETA: I'm going to leave this thread now, as the delicious irony of catching Annoyance in a tangential error is not a worthy feeling for a rationalist to pursue.