Occasionally, I might hit a multiple choice answer option unintentionally and get it right through sheer luck; it would be nice to have [an option to display] a "Redo question" button (as it currently works on wrong answers: the mastery change is reverted, the term remains available for review, and the question is repeated at the end of the current quiz) also on initially-correct answers.
My impression is that "Mark as wrong" moves you to a state equivalent to having given the wrong answer (i.e. vector mastery drops by 2, question is repeated at the end of the quiz and tomorrow), while I'd want to be in a state equivalent to the question not having been asked at all (the use case being accidental input without me actually considering the question). It's similar to the difference between "Mark as correct" and "Redo question" following a wrong answer.
That is correct - it will affect the data negatively in terms of progress.
While I can see this feature being of use to some people, I do not believe I will be adding this - it's a fine line I have draw between requests and adding *too* many settings, as renshuu's complexity (both on the main pages as well as in the settings) are by far the biggest complaint that I receive from current and former users.
Yeah, settings complexity is a familiar pain. One way to avoid adding more settings for this would be to change the already-existing "Show redo button after wrong answer" setting into simply "Show redo button after answer" (regardless of whether it was wrong or not), as I think the intended use case is practically identical. Doesn't save you from having to write more undo code, though, so I don't mind if it doesn't happen (I can just keep doing my current "would I have gotten this right? => take the free point / mark as wrong" assessment).
That is a partial solution, but the quiz state doesn't handle this at the moment - an already answered question is either right or wrong, and "wrong" ones come back up in review. So there is not a good third option yet for "right, but still review it", and that business logic is not centralized enough so that a simple, third state could be added. So even if the visual complexity is resolved, there would be tech debt made by trying to jam this into the core quizzing flow.
