I think the simpler solution to the apparent disconnect between the regression theorem and cryptocurrencies is that cryptocurrencies break the regression theorem. I like Mises and all, but really? People are trying to apply the economics of the 1940s to the monetary system of today...
The regression theorem is account of how money arise from a barter economy. The theorem is irrelevant to bitcoin.