Odds are not favorable to a sign-value notation system as opposed to a positional value notation system thinking to use 2d10 to come up with '42'. I can't recall if they were sign or positional, but I would guess in roman numerals a large value with ten sided dice would be more likely to be created by thinking to paint two dice with a I, II, III, IV, V, X, L, C, D, and M each and adding them up, which wouldn't give you an even spread from smallest to highest value. To recognize how superfluous a d100 is you'd have to have a positional notation system, the base, 10 or otherwise, wouldn't matter. They could do it anyway but it would be a hell of a lot less obvious and sufficiently weird and counter-intuitive that it wouldn't likely show up in casual gaming. If base 6 is your system, you probably would have the 2d10 for 00-99 concept in the games rather than 2-12 bell curve on 7 we use, for instance. If your number system is sign or tally you'd like as not just roll more dice for bigger numbers and shrug off random distribution.
From what I found at Wikipedia just now, evidently ancient Egypt had:
1) A hieroglyphic number set for engraving, with ONLY symbols for 0, 1/2, 3/4 and the powers of 10 through 1,000,000 arranged as nonpositional tallies,
2) hieratic set for papyri, with symbols for each ordinal number AND their multiples by 10, 100 and 1000, and
3) a Demotic set developed in the middle of the last millenium BC, on which I have seen little detail.
The die in the article looks to be inscribed in Greek, which further complicates things because it seems there was both Demotic (Egyptian) and demotic (Greek which has descended as the official language of modern Greece.) To top it all off, I found one journal publication available online claiming Golden Age Greece probably borrowed its number system from Demotic.
http://en.wikipedia.org/wiki/Egyptian_numerals
http://www.math.buffalo.edu/mad/Ancient-Africa/egyptTOgreek.html
In the first two systems d100 would still be superfluous, but could not be replaced by two identical d10; each die would need a wholly different symbol set (unless they just said, "screw it: This other d10 is a d10-100 now.") It looks very likely, however, that the die in the article uses the third (assuming it is meant to depict numbers, not letters,) and I have no idea how that one worked, so maybe Tom can weigh in and enlighten us.
Yeah, it is that speculation that is so intriguing (and I did consider the short road from divination to game, or vice versa.) They could be naming cubes for all we know. I am reasonably confident without checking there is not a "valid" Egyptian name for all combinations of Egyptian hyroglyphs, but then again the Bengals used to have a receiver named Chad Ochocinco, so.... shrugs
Well it is too much for me to ponder, I wouldn't mind knowing which system they were using but once language gets into things my brain shuts down. Gabriel's a classic's guy so he might be able to tell
- Albert Einstein
King of Cairhien 20-7-2
Chancellor of the Landsraad, Archduke of Is'Mod