I'll try either poker or the puzzle; I don't know how to play Risk, so it would make it doubly hard--I have to learn how to play, then I have to program the AI too. I'm pretty slow at this though...I'm still working on my Connect Four solution! But maybe my algorithms from that will come in useful...
The crows maintain that a single crow could destroy the heavens. Doubtless this is so. But it proves nothing against the heavens, for the heavens signify simply: the impossibility of crows.
By the way, which poker game are we talking about, 5 card draw, Hold 'em, etc.? And of course betting is involved I assume?
The crows maintain that a single crow could destroy the heavens. Doubtless this is so. But it proves nothing against the heavens, for the heavens signify simply: the impossibility of crows.
Originally Posted by PJYelton
You guys should post a minesweeper solution anyway. I still want more competition, and it would be as much fun for you to do it now even after the official contest is over.
I'm going to vote for poker too.
Yea I noticed that after I originally announced it, that's why I decided to offer some of the others I have in the works. The poker is Texas holdem and I think that's what I am gonna go with as it needs the least work to get ready.The sliding game AI is much easier. The trouble is, most people will probably roll out very similar solutions since the game is so well documented on the internet. I'm not sure how easy it will be to judge?
Awesome. I'm going to get a head start.The poker is Texas holdem and I think that's what I am gonna go with as it needs the least work to get ready.
The crows maintain that a single crow could destroy the heavens. Doubtless this is so. But it proves nothing against the heavens, for the heavens signify simply: the impossibility of crows.
OK I am almost done, took me a little time to figure a good way of keeping track of pot splitting for no-limit. My original framework had a simple betting system that used fixed bets with no raising.
As a final step, I wanted to get a little feedback as to what game info your AI would need?
Here's what methods I provide so far:
- list of player and their position
- best 5 card hand from 7 cards
- rank of hand, ie.. pair, flush, straight, etc.
- amount in pot
- amount player needs to bet to call
- amount in each players bank
- player acting as dealer (button position)
- your hand and community hand
- notification system - relates game play (bets, folds, wins, deal, etc)
let me know if there's something else you need (or want)
Blinds?
The Blinds get posted automatically (since you can't sit out a hand), part of the notification system will tell who post the large and small blind and the amount. However there is a function that gets the large and small blind amounts
It would probably be a good idea to start a new topic for the new contest.
I found the following site last week, it details a network protocol for poker bots. You can check for missing functionality in your interface by comarison against this one.
http://games.cs.ualberta.ca/webgames/poker/bots.html
The crows maintain that a single crow could destroy the heavens. Doubtless this is so. But it proves nothing against the heavens, for the heavens signify simply: the impossibility of crows.
suit?
I don't know if this is included in the your hand and community hand information, but if not, it would be nice to know suit for some of the higher hands which require spcific suits
a hand is basically a vector of Cards, which is a class that has a suit and a value member as well as functions for reading those values;
I will post what I have later when I get to work in a new thread as dante suggested that way people can start on their design. I was planning on starting a new thread once I had all the code ready to post but I think I will post what I have and ask if someone can make a simple test player. I must admit that when I do these projects I kinda burn myself out in making the framework and be so anxious to post code that I don't feel up to making test players then.
I'm assuming the blinds will go up after a certain number of hands? If so, then knowing the entire blind structure and the number of hands until the next increase will be useful.
Typically blinds double every round.
