I do like my randomization, which is why i suggested this.
however i also mentioned that the epic was the best not the only option. having a scale directly proportional to the chance of finding it seems wrong, because then there is not much reason for the random chance is there? if there is a 1:1000 chance of finding it, assuming you had the worst of luck and only gotten common, you are always ensured to get one per thousand try, due to that. however remember i mentioned 4 rarities, uncommon and rare would also drop, uncommon reducing the common ore needed by 10 and rare by 100. meaning we would get a way smaller number. which is why i suggest a slightly higher rate, after all, its an epic material, should be rare, like really rare, i like the idea of being able to sacrifice a huge amount of common materials to get better materials, but if they are the exact same rate then there isnt much fun to the randomness is there?
doing the math, statistically after 1000 tries you would have
1 epic
10 rare
100 uncommon
899 common
if you have a one to one ratio, this would mean you would statistically have
3 epic and 8 rare and 99 common
However the droprate should show the chance of getting it, meaning that it should be through luck and not just brute force that you achieve it.
The numbers might be a bit high ofc, depending on how you view the material in question and how many would be needed to craft something, if we use it as glyphs which requires 2 rare and can be placed on any tier weapon, then i dont think its too much, if its only useable on the low tier and you might need 5 or 10 ore, then yeah it might be too much.
The other answers made me realize that plain dmg isnt necessarily the best choice, i just like the idea of having rarer materials which would be worth more, rather than it being like ESO where materials were just flooding the markets (also depends on whether or not mining profession will be a thing or everybody can collect every material)