At present, the Two Choice quizzes all have a wrong choice that is entirely randomly chosen (the only constraint being it's not the same as the correct answer).
It would be better to choose the wrong choice (or wrong choices in the case of polychotomous quizzes) based on likely confusions. The could be based on global confusion or it could be a bayesian average.
(copied from https://github.com/jtauber/learning-greek/issues/7)
At present, the Two Choice quizzes all have a wrong choice that is entirely randomly chosen (the only constraint being it's not the same as the correct answer).
It would be better to choose the wrong choice (or wrong choices in the case of polychotomous quizzes) based on likely confusions. The could be based on global confusion or it could be a bayesian average.