The wrong answers are where the work goes
Writing a multiple-choice question is easy. Writing three wrong answers a student who half-understands the topic would genuinely consider is the hard part, and it is what separates a test that measures something from a test that flatters you.
An obviously wrong option is not a distractor; it is padding. Three of them turn a four-option question into a one-option question — you score highly and learn nothing about what you do not know. That is the failure mode of almost every automatically generated MCQ set.
So the generator is pushed towards the mistakes people actually make: the commonly confused term, the value that is right for a neighbouring case, the definition that is nearly correct. When you pick one, the explanation names the confusion — and that confusion is the thing you came to find.