Incredulity test
After the fact sampling is basically Ploppy methodology. It makes me skeptical when people argue for a form of AP using ploppy methodology. Not only is the methodology suspect, but they become suspect for using it.
Agree completely-but it is a little unfair to label the dice control crowd as ploppies just because of this. Both Wong and Thorp (probably others also, just examples that spring to mind) use the "bets I have made-sessions I have won" methodology on occassion. It enlivens dry theory and explains the practical reallities of the theory. I completely avoid using this methodology for the reasons you state, and then people say "so you have never been in a casino?" or "so you never won any money"?
The point that you and John made about the quality of play decreases with higher stakes is an interesting one. But I am not sure that is true if you increase your bets gradually. When I was a red-chipper, losing $1000 was a big deal. Now it is trivial. Losing 50K would be a big deal for me, but if I had doubled my bets several times, I wouldn�t feel that way. You get used to the fluctuations as the betting increases. This is true, but it illustrates a broader point. Consider your outlook on this before you were knowledgeable about card counting. If I had told you would making 1k+ bets and shrugging off thousand double-digit loses, using a 40-year old method to fleece casinos under their noses in plain view of thousands of surveillance cameras, completely legally, I suspect you would have said something like "And this is after my job at the fairground flying small children to the moon on the back of flying elephants?". Many credible things seem incredible initially.
That said, I think this is a half-empty, half-full question. I too am sceptical, but my scepticism manifests itself in the form of wanting to encourage a more sophisticated methodology for proving/disproving the concept, rather than simply dismissing it.