tyonuts said:
Sorry if I offended you shroomy, it was not my intention. But programming requires sacrifices (because any idiot can place two labels, one button and press F5). I didn't suggest spending money, because that ReadIris cannot be bought, not even by small and medium-sized companies. The Office thing is still the best way.
I suggested genetic algorhytms because i want him to know HOW it's done.
And what you ask of me could bring lots of money, but that's not done with genetic algorhytms, but with heuristics and Combinations (see Newton's Binomial equation for an example).
no worries, I dont get offended.
And to take this thread on a tangent.
combinations, permutations, expected value, normalization and other well defined algorithms apply to poker (suprisingly not binomial equations)
But none of them address the issue of prediction of opponents actions which is a huge importance for a strong player.
Typically the issue has been addressed (to modest success) with
expert systems (too hard to modify and test)
neural networks (too slow and costly to train)
Decision Trees (also difficult to train)
Genetic algorithms would be an interesting and different way to attack the issue.
btw to give you a extreemly simplified example....
Say you use combinations to figure out your chance of winning a hand to be 60% (though you can not even get to that 60% number with any certanty due to the nature of poker)
Now there are two more rounds of betting. What makes you the most money?
You could
bet and he calls
check and he bets
check and he bets and you raise
check and he checks (and gets an oportunity for a card to improve his hand)
bet and he raises (does this indicate he has a stronger hand that you predicted?)
.. and more combinations of the above
Now add the next round of betting where every thing changes due to whatever card came out.
etc ... and that is simplified by an order of magnitude.
but try to predict the outcome and choose the best course of action.
now add in the fact that the size of the pot makes a huge difference too, because it effects the odds you are getting for each bet. (if you have a 20% chance of winning do you call a $10 bet? Yes if the pot is over $60 no if it isnt)
