The test cases are unacceptable: the input range is mostly 1 <= len(enemies) <= 3 and len(enemies) + 2 <= len(spells) <= len(enemies) + 4 for all random tests, and when array sizes are big they're always in the form [x, x, x, x, ..., x, y] where x <<< y.

In fact author solution (or any of the existing solutions) can't even handle

test.assert_equals(overkill([100], [*range(1, 21)]), 19)

. So this kata is asking for something unsolvable and should be sent back to draft.

If input array is mutated, random test failure message shows the mutated value instead of original value, which is misleading. A copy of the array should be passed in the user function instead.

Dear thquinn,
I have been suffering for a long time, even though your task is difficult but interesting, but I hate to watch spoilers or refuse the task, I hope that you can tell me a little about the last test.
All tests are fine except for the last one, I get 736.1488211652747 and I can't get 750.440997 points (my combination is {'kickflip': 1, 'pop shove it': 6, '360 flip': 3, '50-50 grind': 3, 'noseslide': 1, 'manual': 9}).
It is quite possible that I am missing something and not taking into account something, but as I understand from the comments, 750.440997 points can be obtained only by sorting through all possible combinations of tricks?
Isn't there a prettier way to get this answer?

Is it about doing as few tricks as possible with the best score?

Could someone please explain why is the correct volume 1 in the below test (I got it as one of the random tests, Python translation):

14, 0, 0, 12, 12

-5, 14, 10, 9, 9

-5, 2, 7, 7, 14

1, 5, 11, 13, 10

9, 0, 1, 1, 7

Thanks in advance!