There are three related failures of probability. One is Wigner's fallacy - the ultimate circularity of science. Another is Sheldrake's fallacy – the impossibility of the infinite tosser. A third emerges as the non-gambler's fallacy – which we may call the ten to one to what? The accuracy varies with the knowledge of the what. There are others, but three is sufficient. And knowledge is a probability.
Now while we may find takers for games of chance, fewer gamble on a coin landing as other than a coin, despite the rather larger number of other forms it could return as. To our knowledge there are too many tosses inherent in the coin as coin.
A probability is only ever with reference to another. The probability of logic=true is assumed to be 1. This gives us say QED very probably correct, yet it is certainly incorrect. It is incorrect by empiricism, since it defies unification with reality with greater and greater probability. But it is certainly incorrect since one can never measure spacetime within it. There is no without. Ultimate checks are uncalibrated. Logic too is ultimately an aspect of spacetime. The fabric of reality is beyond valid deduction.
It's lucky coins know all this. Quite the miracle, if you'd asked poor Wigner.