Our friend 1887 will know the answer if anyone does He opened my eyes to the sheer number of minor varieties that exist. Thank you for your vote of confidence. All you say is perfectly correct the Arabic 1 was used in the sets which would indicate that 797 were Arabic 1 but I have read that mintage of the proof double florin was, as with the other silver coins, 1084. This allows for up to 287 others. Recent work shows the proportion of Roman to Arabic double florins in a sample of 591 to be 200 Roman to 391 Arabic. This is using the reverse A and B as given in Davies. I have no way of knowing whether this proportion holds for the remaining 287 or if it applies at all, all I can say is, yes there are proof Roman I coins. You should also remember the specimen sets have a proportion of Roman I coins and people often mistake the high quality coins for proofs. Again the currency coins, that is those using the currency dies also have occasional proof or proof like strikings. As the years go by the best uncirculated coins, especially early strikings, take on an appearence of being proof but held up against the real thing it is easy to tell the difference. I would be happy to examine anything you think might be a Roman proof and send you photos of mine if you provide email. Obverse and reverse needed. As to value the last one I picked up for £140 not long ago. If I were pushed for a mintage number it would have to be less than 150 unless of course all the 287 are Roman.