It's perfect until you try it with n > 300.

You seem to be implying this is wrong, but actually, it's correct up to n==300.
Not only that, but for all inputs in that range it features O(1) complexity
This is literaly the PERFECT solution

By doing the COBOL version, I realized my original function didn't handle properly cases like this one: a="0.2", b="5", a="0.2", b="0.005", a="0.1", b="0.001". I would suggest to add them to fixed tests for all languages.

Me lie un poco ... en fin.. a seguir aprendiendo.

Wow, that's some interesting fact you got there.

That's true for some languages, though each of these kata's have translations in specific languages not shared across this set of kata's.
I also think this is an artifact from the past and duplicates like this are less likely to get through nowadays.

It is a cute code enough.
There is an interesting model of variation.
To have a row variation and column variation instead of cells variety

?

Exectly))

This kata was authored by founder of the site, in 2013, probably as one of the first tasks ever available. Kata ranking used to be something totally different back then.

Thnaks man.

Your solution is cool btw.

yup.
(I rarely read those entirely, actually... :/ And I read yours some hours ago already... x)

and this is actually a wrong algorithm

spoiler flag! x/