andersx table is correct (btw is there any better function to count this than BINOMDIST in excel adnersx? cause when using this one I have to sum up all probability of all k succeses in n trials) assuming all affixes have the same chance to roll. Also keep in mind people that it's only probability to roll certain affix without a certain value attached to it. So trying to improve something like 9.5 crit chance or 45%+ crit dmg is generally not worth it.
Would be cool to see similar chart with chance to roll, say, 10% crit chance, or 9% or greater! *nudge nudge*