*It appears that it could be used to measure coins, but old Roman coins in that time period were not very round. They were hammered on a die, not cast, so they weren’t a perfect circle. Weight and material constitution was more important to the value of an ancient coin, not a perfectly round shape (especially since the coins were debased of their silver so much during the 3rd century). In addition, some dods had triangular and oval holes.
The Roman-Gallo dodecahedron was created for the Roman troop arms merchants (Negotiatores) & Auxiliary troop arms officers. It was used to measure, procure & maintain the Roman Auxiliary troop polearm weapon shafts; to ensure standardized specs for the varying Auxiliary troops (they were allowed to use their own weapons which varied by troop origin). This would include the diameter - mid shaft, the taper of the shaft just outside/away from the joint at the weapon head socket & the taper of the wood shaft, just inside the head socket. Seach Quora under my name (Richard Allday) for more info.
I thought the same, that it could be used to either mint new coins or test existing coins were the correct dimensions.
My next thought was that each side held a face-plate, maybe painted wood, that could be pressed in to the face, the knobs would help there. The holes underneath the panels could have been used to just make it weigh less, which could explain the differences in hole sizes found.
>which could explain the differences in hole sizes found
There were multiple denominations of Roman coins, and I assume some variance in size at each mint over the centuries. That could explain why each hole is a different size, and why the sizes aren't uniform between each one found.
It seems so easy to test that I feel the idea must have some readily apparent flaw.