Bots and the future of online poker - Page 12

RIIT · #**111** 10-03-2007, 10:34 AM

[ QUOTE ]
[ QUOTE ]
because you cannot play perfect poker. there's just no such thing.

[/ QUOTE ]

most people dont actually think this, right?

[/ QUOTE ]

Poker involves uncertain information and as such the bottom line is that the best guesser wins more.

"perfect" is a relative term that depends on whether or not you're maximizing the information available to you at any given moment in the hand/session.

I think most players believe that the good players have the ability to better interpret the available information and thus are playing more "perfect" poker.

There's really no such thing as "most perfect" poker as long as uncertain information is involved.

RIIT

CallMeIshmael · #**112** 10-03-2007, 10:47 AM

well, it depends on your definition of perfect play, but assuming the nash strategy, against which it is impossible to win, and the best you can do is exactly split (before the rake) can be called 'perfect', it certainly exists.

Now, it wont win the most in (nearly) any game, but it will win in ALL games, and that would be pretty scrary to have in bot form.

RIIT · #**113** 10-03-2007, 11:07 AM

[ QUOTE ]
well, it depends on your definition of perfect play, but assuming the nash strategy, against which it is impossible to win, and the best you can do is exactly split (before the rake) can be called 'perfect', it certainly exists.

Now, it wont win the most in (nearly) any game, but it will win in ALL games, and that would be pretty scrary to have in bot form.

[/ QUOTE ]

There is no absolute/static NE for hold'em (headsup or otherwise). Focusing on HU for a moment, this means that if you publish an absolute/static HU hold'em strategy then there is at least one strategy that trumps your published strat. This also means that the there is at least one trumping strat for the strat that trumped you and so on.

This then means that the best anyone can do in HU is a mixed (dynamic) strat. and this agrees with the current accepted NE theory.

google for trumpbot.whf and read the associated article.

A mixed strat. involves "guesses" which means that the best guesser wins more ... same as rock/paper/scissors

RIIT

davidlong14 · #**114** 10-03-2007, 11:12 AM

a lotta folks better start to work on their live games

Paxinor · #**115** 10-03-2007, 03:33 PM

riit seriously you have no idea what you talking about.

there is an optimal way to play poker just accept it. just because there is no pure strategy nash equilibrium doesn't mean its possible to play in the nash equilibirium which will of course contain mixed strategys. since you don't know what that means really:

roshambo the nash equilibrium is [1/3, 1/3, 1/3]

a computer who plays perfect roshambo will play every choice 33%. WITH NO PATTERN meaning random. (generate a random number between zero and 1, if its smaller than 0.33 pick rock if between 0.33 und 0.66 pick rock and so on.

if you play against him you can not gain positive EV.

i know its easy to reach zero EV against a roshambo bot picking one choice after another.

in poker a mixed strategy means that you play a specific hand on a specific flop not always the same but different. meaning 20% raise, 30% call, 50% checkraise for example.

this is allready done by good pokerplayers and nothing really new!!

you just have to do it randomly

this has nothing to do with "no stable" nashequilibrium because there is one as long as collusion is not interfiering.

it also has NOTHING to do with guesses because you cannot guess random numbers!! there will be no pattern, thats the hole point of it.

you obviously never really studied gametheory so just believe it., because you misinterpret everything!!

you try to turn the facts in a way that it fits your opinion, this is not correct!

if anybody wants to read more deeply into that topic read the discussion about pokerbots in pokertheory forum

RIIT · #**116** 10-03-2007, 03:45 PM

[ QUOTE ]
riit seriously you have no idea what you talking about.

there is an optimal way to play poker just accept it. just because there is no pure strategy nash equilibrium doesn't mean its possible to play in the nash equilibirium which will of course contain mixed strategys. since you don't know what that means really:

roshambo the nash equilibrium is [1/3, 1/3, 1/3]

a computer who plays perfect roshambo will play every choice 33%. WITH NO PATTERN meaning random. (generate a random number between zero and 1, if its smaller than 0.33 pick rock if between 0.33 und 0.66 pick rock and so on.

if you play against him you can not gain positive EV.

i know its easy to reach zero EV against a roshambo bot picking one choice after another.

in poker a mixed strategy means that you play a specific hand on a specific flop not always the same but different. meaning 20% raise, 30% call, 50% checkraise for example.

this is allready done by good pokerplayers and nothing really new!!

you just have to do it randomly

this has nothing to do with "no stable" nashequilibrium because there is one as long as collusion is not interfiering.

it also has NOTHING to do with guesses because you cannot guess random numbers!! there will be no pattern, thats the hole point of it.

you obviously never really studied gametheory so just believe it., because you misinterpret everything!!

you try to turn the facts in a way that it fits your opinion, this is not correct!

if anybody wants to read more deeply into that topic read the discussion about pokerbots in pokertheory forum

[/ QUOTE ]

Paxinor,

You've said nothing here that contradicts any of my previous statements. Not exactly sure why you're indicting my understanding of game theory but there is an apparent contradiction that exists only in your mind.

RIIT

CallMeIshmael · #**117** 10-03-2007, 03:57 PM

RIIT,

A mixed strategy doesnt involve 'guesses', its randomization.

and, obv there is not going to be a pure strategy nash, it will be mixed. but, that doenst mean there isnt a strategy that is unbeatable (dont confuse strategy with pure strategy)

also, just to note: discussing RPS and who 'wins more' moves the discussion to the irrelevant (for this matter) area of results. If I am playing 1/3, 1/3, 1/3, then everyone has an EV agaisnt me of exactly 0. That I won or lost a 10 round match isnt really relevant to the fact that my strategy is unbeatbale in the long run.

Paxinor · #**118** 10-03-2007, 04:38 PM

oh geez...
[ QUOTE ]

There is no absolute/static NE for hold'em (headsup or otherwise). Focusing on HU for a moment, this means that if you publish an absolute/static HU hold'em strategy then there is at least one strategy that trumps your published strat. This also means that the there is at least one trumping strat for the strat that trumped you and so on.

[/ QUOTE ]

this here is WRONG. there is a strategy that is not exploitable this is the whole point of it.

static and pure doesn't mean the same. static means that you can have a strategy that doesn't have to change.

pure means that you can always play one situation the same.

you can play same situations different but still be in a static nash equilibrium, where your opponent can do WHAT he wants and he will lose..

you say the better guesser wins, but this is also wrong. you cannot guess! because you don't have any information.

you should really know that if you'd know anything about this topic

your point is: there is no strategy that can not be exploited, you have to change strategys. this is wrong. you can stick with ONE strategy and you won't be exploitable.

your point is: the better guesser wins. which is also not true, because it's not about guessing.

even though the strategy is mixed, you will still not be exploitable, and the strategy doesn't change just because you raise 50% of a specific hand and and 50% you just call...

you assume that you have to play one particular hand always the same way which is of course not true...

HP · #**119** 10-03-2007, 04:58 PM

[ QUOTE ]
There is no absolute/static NE for hold'em (headsup or otherwise). Focusing on HU for a moment, this means that if you publish an absolute/static HU hold'em strategy then there is at least one strategy that trumps your published strat.

[/ QUOTE ]
yeah, um, no.

RIIT · #**120** 10-03-2007, 08:26 PM

[ QUOTE ]
[ QUOTE ]
There is no absolute/static NE for hold'em (headsup or otherwise). Focusing on HU for a moment, this means that if you publish an absolute/static HU hold'em strategy then there is at least one strategy that trumps your published strat.

[/ QUOTE ]
yeah, um, no.

[/ QUOTE ]

Guys (Pax,Indi) my bad. I should have brushed up on my game theory terms before stating "absolute/static strategy" when what I really meant to say was the proper game theory term "pure strategy"

My apologies for the confusion.

RIIT