Few questions about greed

  1. Will deducing the rng seeds / hardcoding gems positions and times be treated as cheating?
  2. Some low-level solutions (like spawning the army without building any collectors) can sometimes beat the top-ranked solutions. Do you plan to use those rare low-level players for the final ranking?
  3. How do similar battle simulations influence the ranking? Sometimes i see 8 battles with the same player in a row. I guess this happens because fetching the battles from the queue is not synchronized.

  1. You get a separate random seed per match, and if you want to try to reverse-engineer your seed at runtime in order to predict gem placement, be my guest; that sounds pretty badass. (Although if you do it too easily, we may patch it.) If your opponent uses Math.random(), it’ll foil that attack.
  2. I’m guessing that the top-level players will all be able to harvest gold efficiently enough to always defeat the all-in rush, but if that’s not possible with some seeds, then maybe we’ll nerf that a bit.
  3. We were simulating extra matches on demand, and it would pick the nearest enemy opponent, but we just switched it to pick a random opponent instead. It’s also possible that a bug is causing multiple simulation chains to exist, but we really try to prevent that from happening. Let me know if you keep seeing this.


For #2, This isn’t too hard to code against and I agree it’d be cool to see a mix of those types of attacks in the final ranking.


About #3: i still see something like that, less often and less time repeated, but it still happens. And now i see some weird pairs of battles - first battle is a tie, subsequent ones are wins. When I open those ties with ‘defeat the ???’ link, i see that there’s no tie in the replay. Are those first simulations valid? Does the replay reproduce the seeds that were used in original simulation?


The replay does reproduce the seeds, but it’s possible that either the player resubmitted their code, or that there is a Simulator out there (like an old browser tab) which is maybe running an old version with an old bug that doesn’t function properly. Or, perhaps there is a current bug in the Simulator that is still producing disagreements. It will be helpful if you post the team and opponent session if you see a current match where the Simulator disagrees with reality.