My question stems from answering questions (on a test) whose answer order was randomly chosen (not by a human, by a computer). I’m skeptical of something I just read:
Your overall score will improve if you stick to choosing the same letter again and again when making blind guesses. This is because people are never truly random. If you’re making "random" blind guesses in a spread, you will almost certainly reduce your overall guesses odds, because your best attempt at random guessing cannot replicate computer-generated randomness.
This was found here which is a blog and may not be all that great of a source anyways, but that’s not my question.
They then went on to do a little example scenario in which they “randomly thought of a bunch of answers” to guess on the first 15 questions on an ACT from years ago, and also chose to answer C consistently on the same 15 questions. They then recorded how many they got correct while randomly guessing and how many they got correct while guessing C only. They got 2 correct with the “random” guessing and 4 correct while sticking to C.
Obviously running 1 example does not prove anything in statistics, after all anything could happen right?
My initial thought was that it wouldn’t matter if I chose different answers or the same one, I should always have a 25% chance of guessing correctly if there were 4 possible answers since the answer was chosen at random right?
If a multiple choice test’s answer order is randomly generated (I mean randomly generated as in not by a human), how could it be possible that I have a less chance of getting correct answers over many questions while coming up with my own guesses rather than picking a letter and sticking to it?