What you say here is pretty much the rationale I arrived at by comparing online listings of Greek numbers to the characters visible on the die: It could not display all 27 Greek numerical glyphs, but could display Greek numerals through 20. Unfortunately I only have the two pics from the CNet article, which only shows ten of the dies faces. I can identify all but one or two of those shown (depending on whether the possible digamma IS the digamma.) Both kappa and rho are present, so my current working theory is that it may have been a quasi-d100 rolled twice and added to return a result 0-200.

The digamma MAY be present, I cannot be sure; if so, based on the orientation of the other numbers it is turned on its side. It could also be the hieratic glyph for 10, which would further complicate things because an adjacent face has an I. Also, if I understood Wikipedia correctly, any die prior to the Medieval era would, as you said, represent the digamma as F, not ϛ, which would raise the possiblity this is a Byzantine die found in Ptolemaic ruins (or, y'know, an archeologists practical joke/fraud.) There is also an additional glyph I do not recognize between B, E and Θ; I thought it might be an uppercase qoppa or sanpi, but it does not look like any examples I have seen of those.

I see the Greek glyphs for 2, 3, 4, 5, 7, 8, 9, 10, 20, 100 and MAYBE 6, then another I do not recognize.

EDIT: Based on your exchange with Isaac, I am taking you to mean the "other" is some version of alpha with which I am unfamiliar. So if the other one is the digamma, maybe that identifies all those wee can see, but there are still six hidden faces whose glyphs we can only guess.