Game Theory and Poker - Part 1

  • Other
  • Other
  • other gametypes
(5 Votes) 9323


Free membership

Join now


Enjoy the second part of the Game Theory series with DonTabamsey. In this series, Tabamsey who has recently completed his Ph.D in Economics in relation to game theory, discusses theoretical approaches to poker and games in general. In this video he discusses and presents general GTO play. Enjoy the video!


series thematic video Theory Video

Comments (22)

newest first
  • EuanM


    Enjoy the first Video from DonTabamsey!

    Please leave any thoughts or Feedback you may have for the new Coach!
  • Borisian


    Next time I watch this video I'll try to remember not to play FreeCell at the same time. I know I was warned but...
  • pokerjan


    Is it dubbed from Ger?
  • EuanM


    This is the English version, dubbed by the producer himself.
  • koppesh


    please translate to russians
  • w34z3l


    what the hell is a prize war?

    Is that meant to be "thread" or "threat" because it sounds like threat when he says it.

    I'm sure this is a great video I'll watch the other half later.
  • DonTabamsey


    @w34: a price war describes a situation in which all competitors drastically reduce their prices and do not make any more profit; clearly, a single firm could not avoid such a war, because when sticking to its high prices, it would lose all clients; sorry for the typo

    and yes, it is indeed meant to be threat..
  • YohanN7


    Nice. I am interested in the theoretical question of wether FL Holde'm is a solvable (in any precise sense) game or not, and if so, to what extent. You mentioned in passing that it is too complex.

    [I certainly don't expect an answer here, that would probably require more PhD theses. But a Part 2 of the video wold be nice.]
  • YohanN7


    To put #8 a little different way. If we know villains hand range in every situation, I'd not be surprised if the answer is "yes, it's solvable but not solved". One of the problems is that villains range will be a function of all previous hands against us. The task is then to figure out that function. Not easy, even heads up!

  • Jim9137


    While my knowledge of game theory is very limited, I think it has been proven that multiplayer (more than two players) games are unsolvable. I think Mathematics of Poker elaborates on this point. All solutions to these games are then just approximations.
  • Boomer2k10


    Heads Up LHE is certainly solvable. There are a finite number of situations and in a 2-player game that generally means there is an optimum strategy even if it is complex

    I've got a couple of vids in the workings based on this but suffice it to say that HUHU is certainly solvable in's just incredibly hard and more than a supercomputer can do

    Multiplayer games however are infinitely more complex and, as I've been told, there was a reason John Forbes Nash won the Nobel Prize for it!!
  • Boomer2k10


    Also one thin that is fun is that when the University of Alberta had a bot trying to play GTO some of its plays looked absolutely insane (i.e. Limping pre, callin down unbelievably light etc) which is why most of the plays you see in LHE claimin to be around balance and GTO really are just approximation using pure strategies which are "close enough" since almost no human being will be able to tell the difference
  • DonTabamsey


    @Yohan and Boomer: What Nash tells us is that there exists a solution to Limit Holdem (which is far from trivial). So yes, it is solvable.

    But Boomer is probably right that within our lifetime the exact solution will not be discovered. In a practical sense, the term GTO in poker should be used for a strategy against which the scope for exploitation is very limited.
  • w34z3l


    #7 Ah yes, I know what would be meant by price-war. Just wasn't sure if "priZe war" was some type of game theory lingo. :)

    Nice vid, the typos did confuse me though =D
  • YohanN7


    It is very difficult for a layman like me to se the logical reason that a two-player game (with a finite number of situations) is provably solvable and a three player game is provably unsolvable. [I believe you both #10 and #11, it's not that...]

    I define a game as an infinite number of hands between two (or more) players. The number of possible outcomes in one hand is a finite number (HU, SH or FR) of course. The number of outcomes in an infinite session is two to the power of aleph-naught. For comparison, a chess game is finite, and a real poker session is always finite.

    [The man who convincingly computes precisely what two to the power of aleph-naught is will recieve a very very substantial welcome bonus in the mathematical community.]

    Any good links?

    For an advanced intro to complexity of the topic, I can recommend this:

    The article presupposes some (quite a bit, but don't be put off) set theoretical knowledge. The whole question is really "Does a certain set exist?"

    I am sorry that I write too long comments. Won't write for a week. Promise.

  • BanMartins


    An translation to Portuguese or sub's are so wellcome.
  • DonTabamsey


    @Yohan: You are correct in that both the two-player and n-player game of flhu is solvable in the sense that we know an equilibrium exists (basically Nash's result). However, it is much more complicated to compute the solution if you add more players.

    The definition of a game in the case of poker could be one hand or any finite number of hands.

    Differing from the question of determinacy you raised, here the question is only if an equilibrium exists, not who wins in equilibrium. But Nash's argument why it exists also boils down to some non-trivial math using fixed point arguments.
  • Solovirs


    Video bored the hell out of me...
  • YohanN7


    #18 Thanks for the info DonTabamsey.

    I have, of course, more questions. If we have a Nash equilibrium, then we are playing optimally if our opponents play Nash too. But if our opponents play in any other way, then we aren't playing optimally, or are we? We are unbeatable and unexploitable, yes, but optimal?

    I guess what I'm trying to say is that I might have objections againt the terminology some times. What I'd prefer to call an optimal strategy is a strategy that exploits every player type to the maximum.

    A simpler related problem is the all-in end game in NL tourneys where one or more players aren't playing Nash, but are too loose or too tight.

    I think an article is called for. At least, we could have a forum thread. I'll start one if I manage to formulate a first post;)

  • DonTabamsey


    the content of your post is completely correct. but as you say, it is just a question of terminology, which is just a convention.
  • nineshine


    Just finished a semester of game theory and you explain it very well. Look forward to viewing more of your videos
  • z0fman


    nice video but seemed too basic... will watch the next parts to decide