I have seen this problem with the following interpretation.
But both Eliz and Dennis may be right?
If we can repeat a digit in numerator & denominator of the ratio, then can we think of these pairs
(1,1), (1,2), (1,4), (1,8)
(2,1), (2,2), (2,4), (2,8)
(4,1), (4,2), (4,4), (4,8)
(8,1), (8,2), (8,4), (8,8)
as (numerator, denominator) of the ratios?
Also, this is sampling with replacement. The two ratios are formed independently.
How many ratios equal to 1/8? Is it one?
What is the probability of two random ratios will both equal 1/8? Is it (1/16)(1/16)?
How many ratios equal to 2? Is it three?
What is the probability of two random ratios will both equal 2? Is it (3/16)(3/16)?
There are seven possible values for the ratios to equal. What are they?
Find the probability associated with each value.
What is the probability of two random ratios will both equal each one of those values? Then add them up.
If we cannot repeat remove the diagonal.
So P(1/8)=1/12 & P(2)=3/12 as opposed to 1/16 & 3/16.
Instead of seven possible values we now have six.