Ad
  • Default User Avatar

    ok, took a look at someone else's thread below and discovered it's a newline issue.

    This is a rank 7 kata. Lots of folks are gonna come across this who are just learning to code. You're setting a ton of people up for failure here with this trick.

    I can take being a dummy for a sec, I've been coding for a while. But for other folks, highly recommend adding a line at the end of this description like "hint: consider newlines" or better yet, just adding a case to the visible tests that includes several newlines at the end.

  • Default User Avatar

    Similar issue to others. "Expected: false, instead got: true" for "567890".

    If this case should return false, the description is broken, because full conditions are not explained.

    This test case is exactly six digits, and contains only numbers -- per the description, 'true' should be the answer.

  • Default User Avatar

    My understanding is that it used to be open source but isn't anymore, although there is a GitHub repo for tracking issues and ideas.

  • Default User Avatar

    I will try to suggest these options in their discord. Isn't codewars open-source? We can make a request there as well

  • Default User Avatar

    @groskowski Yes, an algorithm to identify outlier katas would be very helpful! Here's one possible way such a system could be implemented on Codewars:

    1. There's a period of time after a kata exits Beta where it's ranking is subject to change. Users are warned of this with a banner or some kind of marking next to the rank.
    2. Power users can be designated a "master of dojo", with a dojo corresponding to kyu, e.g. 7kyu. These power users would be responsible for finalizing ranks.
    3. If a master of dojo does not finalize the rank, the rank is finalized automatically after a certain fixed amount of time has passed.
    4. Algorithms continuously predict kata ranks and add warnings to katas that seem under or over ranked based on the preponderance of other katas in their category.

    That might be a reasonable place to start. Other variations are possible as well. There could also be "masters of dojo" for particular tags in a range of kyus, for example, like a "master of dojo, performance, 6-7kyu".

  • Default User Avatar

    @textninja - That is completely true. Maybe after the kata is approved they could have a dynamic ranking system that would adjust the rank based on total hours spent by all the users divided by completion count. This could even be weighted by the rank of the person who attempted to complete. If a 1kyu is spending 3 hours in a 6kyu kata, something should be wrong. Obviously it would require a lot of fine tuning and implementing a system to check if the user is actively working on the kata and estimate the amount of time spent would be a challenge.

  • Default User Avatar

    A weakness I'll note about assigning a rank based on the average of its votes is that the only people who can give a rating are people who have actually solved the kata, and for some people the key insight comes quicker than others; those who are used to easier kata are apt to give up in the face of the unknown. There is a signal in the number of attempts versus the number of completions but I think this is seldom used.

    I wonder if having specially designated ranking experts and gatekeepers for new katas of a particular level would be something worth implementing? The way I'm picturing it, rankings can be "sealed" or "unsealed", and a person can be designated a ranking expert on, say, 7kyus (only one level per person, but multiple people per level). Such people are given authority to seal whatever ranking was selected by the approver - that is, they give the ranking a final stamp of approval. That way, when ranking controversy occurs - which often happens AFTER a kata exits Beta and has been assigned a rank - there would be opportunity to adjust.

  • Default User Avatar

    Good point. I've added the tags and moved up the problem section

  • Custom User Avatar

    Good points. I think we should at least add the performance tag to this (and #2, #3) kata.

  • Default User Avatar

    Very cool kata, glad it on here -- haven't finished yet, but wanted to suggest for teaching purposes:

    Could add a clear note at the top that this is more about math & performance-thinking than about coding-as-language. Agree with another comment that 7kyu rating is steep. I understand that the mathiness/algorithm-generating nature of the problem is communicated through the last line "Your solution has to support 0≤n≤10^60 Brute-forcing will not work!" but since it's the very last line on the pg, I'd guess a lot of ppl miss/skip over it until they've already started, or might not understand at the 7kyu level what "brute-forcing will not work" means.