Quickselect has O(n**2) in the worst case though. Just iterating over the list and setting the two biggest like in my solution would be optimal in runtime

Does it get faster if you put the x right into the place of self? Or is the 'self' function optimized away?

This is awesome! My solution is O(n) too but yours is abolutely concise and elegant!

#cathax

Nice catch, I just got rid of it.

It's the last of the fixed tests, not the random ones: it has an extra space at the end of the second line.

Thanks, I'm glad someone found this kata useful :D

I just took a look at the random tests, and the rows are all working properly. Hmm...

grouping is pretty usual to "spell" the numbers in pairs

This one was hard because I think I'm not quite grasping the math involved. I got something I'm happy with visually - but I can't quite understand what the shorter, better solutions are doing.

For some reason this passed the tests but shouldn't have: it started taking the fibs at 0, 1, 1, 2... instead of 1, 1, 2, 3...

Of the non-valid examples, why is "(56) 84 6 552" not valid?