poker vs chess wich is easier??

[silly_monkey]

[ QUOTE ]
well jeffnc is the answer to your question:



"Chess is theoretically solvable, but as you say probably not possible now. I wonder if it will be solved in my lifetime. I'd be interested to read the results of that :-) Do you have an opinion on what the result would be?

answer:

first a program now named chinook KNOWS ALL the moves possibilitites in checkers ,500 billions of billion(a lot huh?)and cannot loose but the best game ends up in ties...


[/ QUOTE ]
This isn't entirely true. They haven't worked out every possible combination of moves in the game. What they have done is worked out the important lines and shown that they lead to draws, and proven that all of the other lines can't be any better, without actually expanding them all. In game theory jargon, they have "weakly" solved the game. In the paper in Science, they estimate that they needed to evaluate 10^14 positions for their proof, or roughly 1/10,000th of the search tree.

[ QUOTE ]

now as for chess with the brute force of multiple computer we have now,to be able to solve all the probabilities:

arcticle quote:

"Il n'est pas encore question de s'attaquer à toutes les combinaisons du jeu d'échecs dont le nombre astronomique occuperait les ordinateurs d'aujourd'hui les plus puissants pendant plusieurs siècles, selon les scientifiques."

wich basically means....even with the most powerfull computer today,because of the astromomical high number of chess possibilities(seems like the number of checkers moves possible in a game is a joke??),would still take a couples of century S (dont forget the s and it more then 2...)

[/ QUOTE ]

This is certianly true. In the paper they estimate that chess has a search space on the order of 10^40, meaning the calculations would take something like a billion billion times longer.

Incedentally, the this is also the same research group that is working on most of the poker AI research, including the HU matches with Phil Laak. See www.cs.ualberta.ca/~games . Very cool stuff.

[pacecar86]

the body of knowledge for optimal chess strat > poker because there is no optimal poker strat for multi-player/multi-street games

[Montrealcorp]

well silly monkey i think you are wrong for the checkers game..


the guy who start that program in 1989 left it after 1997 when is computer program couldnt get beat(so your statement is true up untilt then), he restart it it in 2001 with better computer with up to 200 computer wroking at the same time..

so is quote is saying in this today article:

«Percer toutes les combinaisons du jeu de dames a été mon obsession pendant près de vingt ans et il est très satisfaisant d'y parvenir», a relevé l'informaticien.

meaning:

solving all the possibilities of combination in checkers as been my primary goal for over 20 years,feels good to finally acheieved it....

seems to me his done it???

[silly_monkey]

[ QUOTE ]
well silly monkey i think you are wrong for the checkers game..


the guy who start that program in 1989 left it after 1997 when is computer program couldnt get beat(so your statement is true up untilt then), he restart it it in 2001 with better computer with up to 200 computer wroking at the same time..

so is quote is saying in this today article:

«Percer toutes les combinaisons du jeu de dames a été mon obsession pendant près de vingt ans et il est très satisfaisant d'y parvenir», a relevé l'informaticien.

meaning:

solving all the possibilities of combination in checkers as been my primary goal for over 20 years,feels good to finally acheieved it....

seems to me his done it???

[/ QUOTE ]

Go read the actual _scientific_ article at www.sciencemag.org. It explains in detail how they solved the game. They DID NOT solve every position in the game. However, they were able to prove that the game is a draw with perfect play. Read the original scientific article rather than one from a newspaper.

Also, I've been following Dr. Shafferr's research for quite a while. He works at the other major research university in Alberta. I also use computers extensively in my research and we use many of the same facilities and resources.

[jeffnc]

[ QUOTE ]
wich basically means....even with the most powerfull computer today,because of the astromomical high number of chess possibilities(seems like the number of checkers moves possible in a game is a joke??),would still take a couples of century S (dont forget the s and it more then 2...)

[/ QUOTE ]

Right, with most powerful computers TODAY. But 2 things might happen. They might develop new technology that would be able to computer faster than normal computers. They might develop some math algorithms that allow them to prune the move tree massively.

[davidlong14]

Poker is VASTLY more complex (i.e. less "solvable.")

[MicroBob]

[ QUOTE ]
I don't know but poker must be more of a mental test, or do chess players go on super mega monkey tilt too?

[/ QUOTE ]


How much one is prone to tilt doesn't necessarily fit in with my definition of 'mental test'.

Competitve chess (even at the lower levels where I used to participate) is far far far more exhausting and mentally taxing than poker imo.
This might be different with the pressures of the final-table at a major tournament or something.

but overall a 6 hour chess match is far more draining than 6 hours of poker could ever be.
I've taken up to 45 minutes on a single chess-move before trying to find the right combination or find some way to squeeze out of a bad position and I don't think poker situations really compare.

[silly_monkey]

[ QUOTE ]
Poker is VASTLY more complex (i.e. less "solvable.")

[/ QUOTE ]

I disagree. I _think_ the game tree for poker is much smaller than for chess, which means it would be easier for a program to exhaustively search all options (including future cards). However, I think the tree is still far to big to search fully given modern computer hardware.

However, it is worth noting that optimal poker (in the game theory sense) doesn't require you to know anything about your opponent. You just assume that your opponent plays perfectly as well, and choose your best option accordingly. What is more interesting to me would be a maximal player, that tries to model its opponents mistakes in order to capitalize on them.

IMO chess might be the easier game to learn to play at an intermediate level. There is much more good literature available on chess play, and relatively little for poker. Chess also doesn't have the brutal partial reinforcement learning that poker does. If you win or lose a chess match its because of how you played; luck can play a large factor in winning or losing a poker hand, which makes learning correct play more difficult.

[seemorenuts]

I didn't read the entire thread, but I refer you to an article published in Scientific American in the late 70s stating that though chess may be solvable, it is intractable. (Hey, I was a kid then, so I can't remember the quote exactly.) There isn't enough time in the universe to solve chess, so relax. In theory, it's either a win for white or a draw, but humans will never be able to determine which it is.

As for poker, what does a computer care about losing his roll? Ha ha... [img]/images/graemlins/shocked.gif[/img]

[Heisenb3rg]

A few things:

Computing the game theoreticaly correct optimal play for a multi-player agent in an environment with imperfect information and random elements is extremly diffiult. My guess is MUCH more difficult than creating AI for 2 player chess.

The simple nature of adding more players has a HUGE effect on the number of states.

However, despite this, I would call chess more complex because:

The toughest decisions in poker are
a) rare (19/20 decisions are automatic easy)
b) often in very difficult decisions, two lines run extremly close in expectation.. So you are nearly indifferent between two actions anyways..

Each chess move effects future chess moves much more so than is true for poker. Sure metagame is important, but each hand is not nearly as strategically coupled like it is in chess.