Question on answer choice D

Hi, so I am having a hard time working through answer choice D. I understand what others have said about it being correct but one scenario keeps me from understanding it fully. What if the subjects had chosen the bottom every time instead of the top, believing that the "pattern" made it likely the bottom was correct every time. Then they would have chosen every bottom image correctly and still been less than 50% accurate because the images appeared mostly at the top? Or is my scenario of them choosing the bottom every time incompatible with the stimulus in the first place?