Optimal and Exploitive Strategies

[Liv Tyler]

Reading Matt Matros' article in Cardplayer on game theory and poker got me thinking about strategies in a general sense (meta-strategy, if you will). Quick summary of the relevant parts: The optimal strategy is a game plan such that, even if your opponent knows what you are doing, leaves him unable to change your EV in the game by adjusting his play. The exploitive strategy utilizes knowledge of the opponent's play in order to maximize EV, producing a better result than the optimal strategy.

As Matros points out, optimal strategy in poker isn't easy to figure out. That's what poker literature is basically about though: teaching "optimal" strategy. We have a whole body of work to draw on for optimal strategy advice! Different authors/players have somewhat different ideas about what constitutes the best strategy, and they may equally approximate an optimal strategy (although they are sometimes very different.)

In the absense of "reads", optimal strategy is clearly the best, since there is no way to employ exploitive strategies. No matter how they play, the optimal strategy figures to have the best EV. This means that there can be (most) right plays in specific situations, even if the other players are complete unknowns. Once more information about the opponent is known, exploitive strategies can be employed, but until then, one should "stick to the basics".

Interestingly, since strategy in this sense means the collection of ALL plays used by a player, the correct play in certain situations depends not only on how the opponent plays (exploitive strategy), but also on how YOU play (Shania). Not only are you reading them, they're reading you, and you want to make this as difficult as possible. This implies that certain plays become right (in an optimal sense) because you also make the same play with other holdings.

That's what I've been thinking about. Just wanted to share.