This is one of those exceedingly rare chances where I can actualy post something on/. while at work and consider it work related:-)
I'm the lead programmer for a Poker training package called Poki's Poker Academy.
Poki started as an Artificial Intelligence research project at the University of Alberta, where I did my MSc. We've recently commercialized the AI into the above product.
The UofA
research page is an excellent resource for geeks interested in poker. Our publications look at the math and algorithmics behind writing sophisticated poker AI.
Poker is an incredibly geek friendly game. There is a lot of reward in being able to play analytically. The 'reading' of people is a much smaller part of the game than most folks think -- at the highest levels, the best players simply don't have any easy tells, so there is no point looking for any.
Texas Hold'em actually has a much higher skill-to-luck ratio than most poker variants. Since all the players use the public board cards, you can make much stronger inferences about the types of hands your opponents have, and where your hands stands amongst them. This is why it is the game played for the world poker championships.
What you should do is write a spyder-bot that crawls through Amazon.com's site, and collect a huge amount of book information. If you can do it surreptitiously enough, you might not get sued:-D
Uh... getting a computer to bluff is very easy. My MSc. thesis topic is computer poker AI. You can play against my poker bots with my online applet
Getting it to play *well* is the hard part.
Getting a bot to bluff is actually quite easy.
The formulas can be quite simple. If the opponent will fold 20% of the time to your bluff, and the ratio of the cost of your bet to the size of the pot is 0.20 then it is profitable to bluff.
It all boils down to math. Who cares what's on the other player's faces? If they are good players, it won't tell you anything, anyways.
Slashdot Mirror
I, for one, welcome our new butterfly overlords.
100 of which are legit, with the remaining 69900 being computer generated google-rank link-farms....
Q: How do you spot an extroverted computer geek?
A: He stares at *your* shoes when talking to you.
Did you hear the one about the computer scientist that died in the shower? He read the shampoo instructions: Lather, Rince, Repeat.
This is one of those exceedingly rare chances where I can actualy post something on /. while at work and consider it work related :-)
I'm the lead programmer for a Poker training package called Poki's Poker Academy.
Poki started as an Artificial Intelligence research project at the University of Alberta, where I did my MSc. We've recently commercialized the AI into the above product.
The UofA
research page is an excellent resource for geeks interested in poker. Our publications look at the math and algorithmics behind writing sophisticated poker AI.
Poker is an incredibly geek friendly game. There is a lot of reward in being able to play analytically. The 'reading' of people is a much smaller part of the game than most folks think -- at the highest levels, the best players simply don't have any easy tells, so there is no point looking for any.
Texas Hold'em actually has a much higher skill-to-luck ratio than most poker variants. Since all the players use the public board cards, you can make much stronger inferences about the types of hands your opponents have, and where your hands stands amongst them. This is why it is the game played for the world poker championships.
What you should do is write a spyder-bot that crawls through Amazon.com's site, and collect a huge amount of book information. If you can do it surreptitiously enough, you might not get sued :-D
Uh... getting a computer to bluff is very easy. My MSc. thesis topic is computer poker AI. You can play against my poker bots with my online applet
Getting it to play *well* is the hard part.
Getting a bot to bluff is actually quite easy.
The formulas can be quite simple. If the opponent will fold 20% of the time to your bluff, and the ratio of the cost of your bet to the size of the pot is 0.20 then it is profitable to bluff.
It all boils down to math. Who cares what's on the other player's faces? If they are good players, it won't tell you anything, anyways.
------