What the heck...I'll take a crack at it.
First off, I'm assuming that the answer choice is randomly generating (i.e. there is always a 1/5 chance that any random answer picked is correct).
Now that the assumptions are out of the way:
The chance of getting the 1st 61 questions right and the last 39 questions wrong is (1/5)^61 * (4/5)^39 = 3.831 * 10^-47.
Next: figure out the number of combinations that you can get 61 questions right and 39 questions wrong.
100!/61!39! = 9.014*10^27
We want to find p(combo1) OR p(combo2) OR p(combo3)...OR P(combo 9*10^27)
Since all combos have the same liklihood of occuring, the probability of scoring 61% on a 100 question test using random guesses (guessing ACDC without looking at the question should have the same probability of guessing all As or Bs) would be 3.831*10^-47 * 9.014 *10^27 = 3.453 *10^-19.
In other words there would be a 0.00000000000000003453% of getting 61 right on a 100 question test where the right answer choices are randomly and all answers are bubbled in randomly.
Anyone care to share their ideas?