Will computers dominate poker as they did with chess?

[Pirron]

Anwares the question in the topic. If you think the answare is yes than when and what conequences it will bring?

To be specific - domination - i mean beeing able to beat strongest human players and to be even more specific i mean texas hold'em deep stack no limit cash games shorthanded or heads up, because this suposed to be the most challenging form of poker for bots (as far as i know).

Thx for any input!

[infinity235]

Nope.

Chess = 100% skill
Poker = a mixture of skill and luck.

[jukofyork]

[ QUOTE ]
Nope.

Chess = 100% skill
Poker = a mixture of skill and luck.

[/ QUOTE ]
Backgammon = a mixture of skill and luck... Try playing against Snowie.

Go = 100% skill... Try reading the Computer Go wiki page.

Juk [img]/images/graemlins/smile.gif[/img]

[DavidC]

[ QUOTE ]
[ QUOTE ]
Nope.

Chess = 100% skill
Poker = a mixture of skill and luck.

[/ QUOTE ]
Backgammon = a mixture of skill and luck... Try playing against Snowie.

Go = 100% skill... Try reading the Computer Go wiki page.

Juk [img]/images/graemlins/smile.gif[/img]

[/ QUOTE ]

Awesome link.

Just curious, though...

Since they solved 5x5 Go (black wins the board, lolz), and for that matter, I think they solved checkers too recently, how feasible would it be for a crappy computer program to search through the solution to either checkers or go to play perfectly, even without knowledge of why it was working or any strategy, just following the pre-finished solution?

Would it be something that would be able to play against a human at reasonable speed running on a reasonable modern personal computer, or would you still need a high-end (super?) computer to run it at a decent speed?

--Dave.

[Devilboy666]

Dave

The search space for a poker hand is so large that all the computers in the world would take millions of years to go through it all. That's part of why this is such an interesting problem to AI researchers - it's impossible to solve using brute force.

[Devilboy666]

[ QUOTE ]
Dave

The search space for a poker hand is so large that all the computers in the world would take millions of years to go through it all. That's part of why this is such an interesting problem to AI researchers - it's impossible to solve using brute force.

[/ QUOTE ]

Sorry Dave I just noticed that I am a retard. You were asking about go / checkers...

This link:
http://erikvanderwerf.tengen.nl/5x5/5x5solved.html
Mentioned that the entire a/b search was completed on reasonable modern hardware in just 4 hours. (This is for the 5x5 Go)

Once the search is complete the resulting dataset can be used to play perfect go with no need for the computer to do any calculation / tree search and can thus act instantly.

[jogsxyz]

[ QUOTE ]
Dave

The search space for a poker hand is so large that all the computers in the world would take millions of years to go through it all. That's part of why this is such an interesting problem to AI researchers - it's impossible to solve using brute force.

[/ QUOTE ]

Chess and Go need brute force, poker doesn't. Look at pokerstove. It gives the user the option of brute force or monte carlo. The poker computer would choose monte carlo.

[indianaV8]

[ QUOTE ]
Nope.

Chess = 100% skill
Poker = a mixture of skill and luck.

[/ QUOTE ]

Yes.

Chess = 100% skill
Poker = 100% skill

[jogsxyz]

[ QUOTE ]


Chess = 100% skill
Poker = a mixture of skill and luck.

[/ QUOTE ]

It depends on the length of time. Even Chess isn't
100% skill. If it were against the same opponent
with white you would always win, always draw or
always lose.

With poker skill is a linear function of time.
Luck is a linear function of the square root of
time.

That actually true with chess also. Only in chess
skill can count more than luck within one game.

[indianaV8]

[ QUOTE ]
[ QUOTE ]


Chess = 100% skill
Poker = a mixture of skill and luck.

[/ QUOTE ]

Luck is a linear function of the square root of
time.



[/ QUOTE ]

it's not linear function.

otherwise great reply ...!