That number seems not right no matter how I try to interpret it. Do you mean if you would have 1 affix you want already on your weapon and the other one not, you would have to re-roll an average of 64x with every roll being those two affixes as starting point? Even that non-real situation gives me 1/39 instead of 1/64. What did you do there to calculate it?
Edit: Oooh I see what you did there. You did a simple 9 x 8 and just removed 1 from the 9 to do 8 x 8. Because the one affix is not an option in the possibilities anymore. But that is the base of possible combinations with that rule as starting point. But after the first roll, that rule does not apply anymore. Also 8 x 8 is not right in the end anyway. I got almost 1/56 from the program which is 8 x 7. I think I can explain that from the fact that the first affix only has 8 options left where only 1 can be the correct one. The second affix has also 8 options left with only 1 being the correct one, BUT because you can’t have the same affixes twice the second affix is blocked 7 out of 8 possible affixes when it equals affix 1.
Thus, complicated way of saying 1/56 chance if you would roll only once, but you keep rolling and then you get the numbers of my program anyway on average. Which was around 1/25 not 1/64.
(anyway, thnx for the math puzzle , was fun)