I was thinking about the problem some more.
from 24 affix groups, there are C(24,6) possible 6 affix items. In the 24 affix groups, there are 4 single stat affixes and one double stat group. So, the chance of a double stat equals the chance of at least one single stat * a matching double stat / C(24,6).
Having "at least" (ie greater than or equal to) and a "matching" double stat complicates the math and it's above my head right now.
if someone could solve this problem that would be helpful.