4 ms·
> Is RPS a game of skill? If you do a 2/3, or 3/5, etc, then maybe. Numberphile did an episode on how you're more likely to win by following a particular strat
by throw0101a 7y ago
> Is RPS a game of skill?
If you do a 2/3, or 3/5, etc, then maybe. Numberphile did an episode on how you're more likely to win by following a particular strategy:
* https://www.youtube.com/watch?v=rudzYPHuewc https://www.youtube.com/watch?v=rudzYPHuewc
Presumably this works best if the other person is not following the same strategy. The paper that they reference (PDF):
* https://arxiv.org/pdf/1404.5199v1.pdf https://arxiv.org/pdf/1404.5199v1.pdf
- petemc_ 7y agoBut then you play someone using similar strategy and you meet the nash equilibrium.
- ulucs 7y agoNo pure Nash Equilibria means they won't ever converge, as humans are great at seeing patterns that aren't there and cannot make random choices without outside help.
- quickthrower2 7y agoI just watched that video. Fascinating. Of course it means if the other person is playing the strategy, then you want to play the strategy to beat that strategy, which is just as simple. But if they realise you are doing that they might switch to the one that beats the one that beats the one. I guess it gets a bit like poker where you are trying to appear "loose" then suddenly play tight or whatever. I was wondering how a human can generate random numbers while playing. One way could be if you can hear your heart beat, the number of beats between plays mod 3 would allow you to play randomly enough to throw off your opponents strategy. You could also memorise a recently generate sequence of random numbers!
- cesarb 7y ago> Of course it means if the other person is playing the strategy, then you want to play the strategy to beat that strategy, which is just as simple. But if they realise you are doing that they might switch to the one that beats the one that beats the one. Reminds me of Iocaine Powder: https://web.archive.org/web/20100131080653/http://ofb.net/~egnor/iocaine.html https://web.archive.org/web/20100131080653/http://ofb.net/~e...
- kentonv 7y agoMy favorite entry in that competition was Fork Bot. The contestants were written in C, you see, and all rounds of play executed in the same process. Obviously there's all kinds of ways you could cheat in this environment but Fork Bot struck me as especially clever. Fork Bot, on its turn, would fork() twice and return a different move in each of the three resulting processes. On its next turn, it would look back to the previous turn and see if it won. If not, it would exit(). So only the process where it won every round would survive... Though, it ran into a problem with one of the other cheaters: "Unfortunately, since all three moves lost to the Psychic Friends Network after the first turn, the program exited and the remainder of that match was declared forfeited." http://webdocs.cs.ualberta.ca/~darse/rsb-results2.html http://webdocs.cs.ualberta.ca/~darse/rsb-results2.html
- throw0101a 7y ago> You could also memorise a recently generate sequence of random numbers! 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 …
- nabla9 7y ago20 years ago my friend made simple RPS computer game that always won iterated games against human player. It was relatively simple Bayesian algorithm (not naive Bayes exactly). Humans really suck at producing randomness.