how come theres no go threads? even though it is the most perfect of virgin autist pastimes
why not shogi then?
>even though it is the most perfect of virgin autist pastimes
You are an imbecile.
If you really think that go fits the board, why are you making a shitposting thread instead of a go one?
Machines have already beaten go, it's pointless now. Soon chess and poker will join among others.
>>17530792
But chess was the first to go.
>>17530792
>poker
I mean, considering that the entire game revolves around chance, I don't really see how an AI could perfect it like Chess or Go.
>>17531052
playing the odds is an important part of poker
playing the players is even more important though, so skynet is probably a good while off on winning this one
>>17531216
Nah. Heads up limit hold 'em is solved, heads up no limit AI wins significantly vs pros.
With more players there's more complexity though, games with more people aren't beaten yet I think. Not sure how much effort there has been at it, though.
>>17530975
I think they mean "solving" as in "finding every possible permutation of piece positions"
>>17531216
But that doesn't eliminate the worst case scenario where the AI is dealt losing hands. I'm sure the AIs can have an above average win rate but never 100% across a large number of games.
>>17531389
An AI only really has to average a higher win rate than contemporary players in world-class tournaments.
>>17531466
Within the context of Poker sure, but it's not nearly as impressive compared to the likes of Chess or Go, don't you think? Both of those games are far more complex with all variables known (minus what your opponent's move will be).
>>17531547
I just see a lower-than-100% win rate as a testament to humans not sucking so much.
If you coded a perfect go AI and let two of them go against each other, one black and one white, you'll get a win rate slightly different from 50/50, reflecting the effective first move advantage (because the compensation for the white player is not dead-on accurate).
If the same AI played a human, every single win it would score over the rate reflecting the first move advantage rate (or 50%, if they alternated black and white) would be owed to a mistake made by the human, and not to a flaw in the AI (becasue it's perfect and has none)
>>17531216
>playing the odds
>playing the players
That's exactly why computers will suck at poker for a good long while. Their risks and bluffing are too predictable.
Another game computers suck at is backgammon. That one is largely about accounting for probability while deciding when to take risks.
>>17531987
>That one is largely about accounting for probability while deciding when to take risks.
that should be perfectly solvable. has no one bothered to write a decent bot?
>>17531389
perfect play and 100% win rate are completely different things. cmon dude.
>>17531994
It seems like it should be solvable, but I've never played a computer program that was any good at judging the bell curve of possible rolls and deciding when and how often to take risks. I guess there was a program designed to master it back in the '90s, but I don't know how consistently good it was.
>>17531997
In a game like Chess and Go, perfect play and 100% win rate can absolutely be the same thing.
>>17532129
we were talking about poker bots, fuckhead
>>17532129
Perfect play in poker is to maximise profits in winning situations and to minimise losses in losing situations.
>>17530230
I'm pretty sure person with brain is more favored than a person who only have muscles, at least in ancient time china where go was invented.
>>17530230
go is not fun
>>17534432
that's because you don't know how to play
also, that aesthetic... nnnnngh
http://www.cosumi.net/en/replay/?b=You&w=COSUMI&k=0&r=b6&bs=7&gr=ddcecdbddecfdfbccccbdbdgefegfgcgffbbdacadctttt
>>17539531
no go is just not fun