Dabbling in Game theory

[patrick dicaprio]

i have a math background (degree from MIT) and have some knowledge of game theory. I have been considering learning a lot more about it in the hopes that I canimprove my poker game, especially NL. has anyone else tried/done this and does anyone believe it is a waste of time? and if you have done it, any tips or advice? Thanks. I realize this is a very general question.

Pat

[Monty Cantsin]

I believe there is very little in the existing body of GT literature that will have a significant impact on improving your ability to win at Poker.

GT, in my limited understanding, is about analyzing optimal strategies between totally rational agents. Winning at Poker is about exploiting the errors of imperfect agents.

The classic example of applied GT in Poker is the development of optimal mixed strategies for deciding bluffing frequencies such that your opponent can't profit by always calling or always folding. With the right frequency, you put your opponent in a situation where it doesn't matter whether he calls or folds - thereby removing his ability to exploit your strategy with one of his own.

This is an incredibly interesting and important idea, but I'm pretty sure it doesn't come out of original GT texts - it's the result of guys Sklansky applying GT thinking to Poker. In other words, in terms of useful applications, there's more stuff coming out of Poker Theory then there is coming out of classic, traditional Game Theory.

But even moreso, how often does an optimal mixed strategy bluffing frequency come into play in your actual game? It's useful to understand the principle, but in actual play you are much more likely to be in a position to exploit your opponent's mistakes through theoreticaly sub-optimal mixed strategies (eg. bluffing more frequently against a tight/weak player).

There's one book that I know of that is a full-length attempt to establish a GT-based approach to Poker, and it's generally considered not that useful.

To summarize: some GT concepts are really useful, but I don't think reading primary GT texts will be that useful.

In my view, you will benefit so much more by poring over this site and its archives, reading things like this and this and this to choose three random examples.

There's no way I personally would spend a lot of time reading papers about zero-sum hyper-correlated regret equilibria when there's there are juicy conversations like those to participate in.

Actually, I guess what I'm suggesting is for you to go read those papers and then figure out yourself their applications to Poker...here, where I can benefit!

/mc

[Neil Stevens]

I don't understand why people on this forum keep saying game theory is limited to perfect play.

Even my introductory textbook showed how to model irrational play. And even if you don't do that, concepts like dominated strategies are applicable against rational or irrational opponents.

[Monty Cantsin]

You're right, my description of GT was too narrow. My overall point still stands, I think - if your goal is to improve your game, you're much better off reading good applied poker theory than primary academic GT texts./mc

[Monty Cantsin]

[ QUOTE ]
In my view, you will benefit so much more by poring over this site and its archives, reading things like this and this and this to choose three random examples.

[/ QUOTE ]

Btw, that first "this" was supposed to be a link to Sredni's Shania post, which has mysteriously disappeared from 2+2. Since it didn't work I am reproducing it here, without permission, as a public service.

/mc

[Iceman]

[ QUOTE ]
i have a math background (degree from MIT) and have some knowledge of game theory. I have been considering learning a lot more about it in the hopes that I canimprove my poker game, especially NL. has anyone else tried/done this and does anyone believe it is a waste of time? and if you have done it, any tips or advice? Thanks. I realize this is a very general question.

[/ QUOTE ]

Game theory is absolutely useless in actual poker games. It can be used to solve some "thought experiments" involving contrived extreme situations, but for real-life play you're much better off learning poker theory and strategies for specific games, and gaining experience applying those theories and strategies.

Take the situation where you have an obvious draw on the river against an obvious made hand, both you and your opponent know you win if you hit your draw and lose otherwise, and you want to know how often to bluff if you miss your draw. This would seem to be a situation where game theory would be useful. However, game theory will give you the same answer (of what fraction of the time to bluff) for every opponent. You'll do much better adjusting to the specific opponent you're facing - for someone who will fold even a moderate fraction of the time here to one bet in a large pot, you should bluff a lot, while some players will always call for that one final bet, and you should never try to bluff them. If you don't know your opponent at all, then you should go by what players in that game, limit, and cardroom are normally like.

[Kopefire]

[ QUOTE ]


Game theory is absolutely useless in actual poker games. It can be used to solve some "thought experiments" involving contrived extreme situations, but for real-life play you're much better off learning poker theory and strategies for specific games, and gaining experience applying those theories and strategies.

[/ QUOTE ]

I absolutely disagree with this as stated.

Game theory is as usefull in poker as something like electrical engineering is in computer programming.

The utility does not come from telling you how to make a particular play, but in understanding the mathematical reason behind a particular play being correct.

You won't use GT at the table to decide if you should fold, call or raise. But you will be able to evaluate the arguments for folding, calling or raising away from the table with much greater depth and precission with an understanding of GT than without.

GT provides access to the underlying mathematical resons plays are good or bad.

At the table, practical experience and advice will take you furhter, faster to start with. However, at some point, understanding the underlying reasons for plays being correct will enable one to advance further and faster at the high end of the skill curve (all other things being equal of course).

[burningyen]

My guess would be that GT has some important ramifications for tournament play, especially final table play.

[Tom22]

so how would one learn more about game theory....

I heard of the concept but know for little about it. Would game theory dictate when you should fold, call, raise, and amount to bet based on how your opponent would react.