OP solved it, closing

OP solved it, closing ~~

because it is! ;)

thirt(n)
  -> thirt(total1)
    -> thirt(total2)
      -> thirt(total3)
        -> return total4

here is what you're doing right now (try to firgure out what's the problem ;)

hard (if not impossible) to tell as long as you don't format your code correctly so that it's "readable python" ;p (see here)

edit: I believe I got it. It's because you use reccursion, but "incompletely".

that just means you don't put your "return" at the correct place in your code.

Not a kata issue.

Please read the last bullet point in this paragraph of FAQ.

You're not returning anything.

@ktroy try it without numpy.

@ktroy: Can you paste here your solution and mark it with the spoiler flag. The reference solution uses a very straightforward approach and the kata has more than 1000 Python solutions. Therefore, I don't think that the tests are too strict. Also to come up with some more complex comparison mechanism in random test is IMO an overkill for a 7kyu kata. Regards, suic

only the first results are the same as expected

Well, we don't know your approach, but there's definitely another one.

>>> 49.400181665024846 % 16.46672722167495
16.466727221674944
>>> 49.40018166502485 % 16.46672722167495
0.0

So the returned result really is incorrect.

Edit: You're right. For python, random tests are too strict.

Maybe change your browser, I have no problem running your code and it passes.
PS: OK, now I see you posted above.
Good.