[ QUOTE ]
I've got those. Selby comes the close to spelling it out, but there's some details he glosses over in his exposition.
I'm basically just trying to get off the ground using LP to "solve" some simple poker model games, like preflop all-in/fold as a starting point.
I've done this before using simulation and a whole slew of optimization approaches, but this is all very slow when you start adding anything non-trivial to the strategy space (like allowing multiple raise sizes or multiple betting rounds, even staying preflop.)
eastbay
[/ QUOTE ]
I plan on posting the basic HU NL PF LP problem to
A Slave to Variance fairly soon. I've got it worked out, but I'm in the middle of a move so I don't have good internet connectivity right now.
But an LP solution will not really allow you to deal with the complexity of adding multi round betting to your solution. The combinatorial explosion is too high to deal with general postflop play (you might be able to do a rank only solution though). The Selby formulation does allow for multiple bets, limit and pot limit, but doesn't allow for postflop betting at all. Strangely it doesn't do no limit directly, but you can use the code to generate reasonable solutions for jam/fold.
- Andrew
www.pokerstove.com