[Disclaimer: This may make your head spin]
So I've been playing with game theory a bit and found some interesting results. Here's an example of a game theory strategy for NL.
Basic Concepts:
1) Get in pot cheaply
2) Massively overbet with some premium draws
3) Massively overbet with the nuts or the best hand
That's pretty much the idea, but it's not as easy to do as it sounds. You have to know what you're doing, cuz you play this the wrong way, and you're spewing chips like there's no tomorrow.
Example:
2 Players get to the flop and the pot is $500
Both players have $5,000 in their stacks
Player 2 on button has Ac Kd
Flop [Ah 7s 6s]
Game theory player moves all in.
Now he's on a draw a good 75% of the time. Here are hands covering the spectrum.
GT Hand, Bet/Call, GT Chances, GT Equity, Player Chances, Player Equity
1) Ts 8s $5,000 47.80% $5,019 52.20% $5,481
2) 8s 5h $5,000 37.00% $3,885 63.00% $6,615
3) 7h 7d $5,000 98.40% $10,332 1.60% $ 168
4) 5h 9s $5,000 23.70% $2,488 76.30% $8,012
Game Theory Equity: $21,724 Player Equity: $20,276
Now even knowing he's on a draw 75% of the time, that 25% costs so much that you cannot call profitably. In fact if you called his "stupid" overbet with tptk, he's going to win $1,448 every four hands played this way, or $362 a hand.
On top of this, when the game theory player shoves and is called by a better hand which gets drawn out on, that player may go on severe tilt, and blow off a ton more chips in the hands following the beat.
Finally, this player is going to pick up SO MANY chips from uncalled pots. It's very tough to have a hand you want to call that $5,000 bet with. This super pressure allows the GT player to win chips he never could from his hand's strength alone.
The biggest downside is going to be variance. Playing this hyper-aggressive style is going to yield some wild swings both for good and bad. If he shoves in ten times in a row with a 35% chance to win, 1.5% of the time he will miss all ten shoves. That's going to cost him $50,000. But make no mistake about it, the reward of this style is much higher than the risk when employed correctly.
Now that I touched on some basic game theory, my next post is my first attempt at creating my own game theory for uNL.
Friday, February 15, 2008
Subscribe to:
Post Comments (Atom)

No comments:
Post a Comment