• yes you are right about python. Better to remove this statement.

  • But why say you may return a string if the tests mark it wrong?

  • This many tests really slows down most browsers 😬 and is a nightmare to debug...

  • I added 20000 random tests, for both languages, which includes edge cases too.

  • One of them is induction and the other one is deduction. So dont mix concepts.

    • You may return the result as string => is an alternative way for the languages exceed the size.
    • "090" => And this one is away from the description of the reverse number.
  • You may return the result as string

    I don't think this is the case in python => just enforce numbers only to keep the desc language neutral.
    Or if you want to keep the shallow equality then put a ~~~if:javascript in the desc.
    However allowing strings introduces ambiguities like "090" etc...

  • It's okay to have two languages with different ranges. The important thing is that if two solutions (in two different languages) have the same complexity, they should take similar times to run in codewars.

  • As for the testing, it's a bit better but 10^3 is still a drop in the ocean compared to 10^11.
    I actually wouldn't worry about this (unless someone else has the same type of approach as me). My edge case is when n is very close to a power of 10 so maybe try a section of edge cases that generate numbers within 1% of a power of ten? This would be better than just having 1000 random tests

  • Part of this is because I solved v1 with the approach you needed for v2...
    I don't see why you can't increase these tests up to 10^100 or even more.

    The problem here is if you have both of them approved, say one at 6kyu and one at 5kyu, then, when a user solves the easy one, he can copy/adapt the code from all the solutions of v1 and then use that to solve v2. I'm saying that it's a bad idea to ever have two katas whose intersection of solutions is non-empty. This is not the case for the katas you linked; although there are other kata pairs but that's not something that we should aim to repeat ;)

    As I say I think v2 is good (I haven't upvoted yet but that's because I'm concerned about v1). As for ranking, maybe 6 is harsh but these things are hella subjective 🙄

  • Bro isn't having the same rank absurd, at least they have different ranges. At least give 5 for the extreme one :D

    • "Also, the larger the range, the less the chance of wierd edge cases." => i increased random test to 1000
  • Yes you are right, your v1 solution passes the v2 too. That is not same for everyone, Even tough they only have different ranges, they need different approaches(yes one's unit tests contain the other's tests). But you may consider these katas as 2 versions of the same question. I have seen many katas like these, like https://www.codewars.com/kata/56a73d2194505c29f600002d and https://www.codewars.com/kata/5993c1d917bc97d05d000068

  • Also, the larger the range, the less the chance of wierd edge cases.

    This can be solved with edge cases in fixed tests and multiple sets of random tests, one with small numbers and one with large numbers.

    I agree that one kata is sufficient. With large numbers is better (it's possible to exec 1000 tests with number as large as 10**2000 in about 1 second).

  • I don't really see how this is much harder than the other one. Yes the numbers are larger, but is there an algorithm that works for the last kata but is too slow for this one? Yes some algorithms might need to be tweaked or extended, but I don't see how this one tests the user in a way the other one doesn't?

    Also, the larger the range, the less the chance of wierd edge cases. My alogrithm isn't perfect but faired much better on these tests than the other kata simply because the numbers were so large :/

    I think one of these katas would be a good addition to codewars but I can't see a difference between the two (I ranked both at 6).

    You're in danger of duplicating yourself!

  • Loading more items...