Just recently I posted this king-of-the-hill question. It was a rock paper scissors contest, but opponents had access to each other's source codes. It wasn't very well received - it had some downvotes and will likely soon be closed. I would like to know why it was't well recieved, to keep note of for future posts. Here are some reasons that I could think of:

  • The submissions were limited to javascript only, which puts limitations on how people can submit
  • Related questions had been asked before, like this one and this one
    • However, there were differences between my question and the ones that were brought up as related. One was a simple rock paper scissors competition, but it didn't have any things like access to other people's codes
    • The other one was a similar premise, but for a different game (prisoner's dilemma)
  • Flawed premise - if I were to guess this would be the main reason
    • eval() calls were banned to prevent an infinite loop of code calling each other
    • This possibly led to a less interesting challenge

Personally I find king-of-the-hill competitions a pretty interesting concept - being able to interact with other peoples code looks like it would add more depth to a challenge

  • 4
    \$\begingroup\$ I'm not going to write a full answer, but my bet is on "flawed premise". King of the hill challenges are hard to write and the previous two don't seem like they are really huge problems to me. \$\endgroup\$
    – Wheat Witch Mod
    Aug 9 '20 at 0:54
  • 3
    \$\begingroup\$ I'll note that the earlier RPS KotH was from 2014, and I doubt it would be so well-received nowadays. It seems like the RPS premise runs into the general issue that later answers can be tailored against existing ones. \$\endgroup\$
    – xnor
    Aug 9 '20 at 3:56
  • 4
    \$\begingroup\$ When you say "being able to interact with other peoples code looks like it would add more depth to a challenge", what are you imagining? I'm not seeing how this would work without eval'ing code, unless the posted code is really simple. One could take already-posted answers, fingerprint their code, and if it matches do the strategy that always beats it, but I think that's not what you're going for. \$\endgroup\$
    – xnor
    Aug 9 '20 at 8:00
  • 3
    \$\begingroup\$ In general, you should ask for feedback like this in the sandbox instead of as a separate meta question. For this case, I generally agree with the other comments and would add that RPS is kind of inherently uninteresting in that no strategy can do better than 1/3 win rate against a truly random opponent. Doing "smart" things can only beat people who are less "smart." \$\endgroup\$ Aug 9 '20 at 19:58
  • \$\begingroup\$ All three of my popular KoTHs have been Javascript only. I think it's a perfectly valid requirement for a KoTH challenge, esp. one with source code reading involved. \$\endgroup\$ Aug 11 '20 at 19:58
  • \$\begingroup\$ One thing you didn't think of is, RPS is pretty much dead. The optimal solution is going to be a lookup table. \$\endgroup\$
    – null
    Aug 20 '20 at 9:36

You must log in to answer this question.

Browse other questions tagged .