While I appreciate the suggestion, I prefer having the user find out earlier rather than later whether their solution is efficient enough. I'm also hesitant to make changes not suggested by the reviewers in the beta phase.

Fixed

Corrected, thanks!

Fixed.

fixed

done

Fixed.

No, this kata much easier because it doesn't care about draw.

Fixed, thanks for pointing that out

With the current JS setup, you can't even run multiple tests shuffled. You'd have to use a class instance to do that - there has to be state in the current ( or any ) setup ( unless you pass the complete history every time ). Passing -1 to reinitialise the game works well IMO, it makes for simpler programming than having to wrap the function in a class.

I think I got your point. There is no such problem, because all input is number of matches, output is single answer, no more interaction between tests and solution blocks. var log = []; does not uses in tests, it's a little tip (maybe not the best one) to use fields outside main function

well, my point is that users might keep track of state between these calls, because you are running each such test (multiple times back and forth) sequentially. If you run multiple such tests in order of breadth, you will deny users the advantage of keeping track of external state. I don't know how to explain this better. sorry :/