Poker for college
- Thread starter Walkerie
- Start date
Soldato
- Joined
- 18 Oct 2002
- Posts
- 5,600
- Location
- Surrey
I'd allow about a week for me to do it, if you need to learn java at the same time them probably more like a month or more. It would be a good size for a final year project with the AI components and the UI layer.
I'd want to be able to set it's win / loss ratio by being able to control how optomistic it was. There already MILLIONS of poker playing bots out there btw and it's a classic avenue for AI researchers.
Paul
I'd want to be able to set it's win / loss ratio by being able to control how optomistic it was. There already MILLIONS of poker playing bots out there btw and it's a classic avenue for AI researchers.
Paul
Soldato
- Joined
- 18 Oct 2002
- Posts
- 5,600
- Location
- Surrey
WalkeruK said:
how good is your java? and MORE importantly how good are your software design / AI skills.
it would be fairly easy to write a prog that would let two or more people play poker over the web (web services for the input / output to the client) the next step would be to extend thais to write a client programme that would take the part of one of the players to bulk up the numbers. Start with the easy bits, see how you get on and then look at the bot aspect of it.
HT
Considering the problem myself and the approach i would take would be to get it working for 2 or more human players. You can get all the rules of poker working, such as who has the highest hand and everything, and this is fairly straightforward if you take time to work out the games logic and how it translates to rules within the program.
The computer player will have to make choices based on rules too, and its this thats going to provide more of a challenge, and in my eyes the interesting part in the project. The computer will have to choose to keep cards, turn them in, and work out what the hell its hoping to turn up, and this will be based on maths, and hopefully some random logic that will make it more interesting. Eventually you could look to make the computer bluff, etc.
Anyway i know this is a pretty useless post, but still its done now!
The computer player will have to make choices based on rules too, and its this thats going to provide more of a challenge, and in my eyes the interesting part in the project. The computer will have to choose to keep cards, turn them in, and work out what the hell its hoping to turn up, and this will be based on maths, and hopefully some random logic that will make it more interesting. Eventually you could look to make the computer bluff, etc.
Anyway i know this is a pretty useless post, but still its done now!
Hi,
What exactly are you trying to make - a poker bot which works with one of the online sites, or just a bot where you manually type in the cards and it makes a decision.
If you want it to work with online sites you will need to first write some software to spy of the GUI of the poker software, which will take you at least a few weeks to get working properly.
How long it would take to actually make the bot depends entirely on how good you want to make it. You could make a trivial bot that just goes all in with the top 10 hands and folds otherwise in a few hours. To make a player that actually has a chance of making money would take months, and would almost certainly be impossible unless you have a lot of advanced knowledge about AI.
To put it into perspective no-one has yet made a poker bot capable of making a profit in no-limit tournament poker. Low limit ring games is a little bit easier, but its a very complex problem.
What exactly are you trying to make - a poker bot which works with one of the online sites, or just a bot where you manually type in the cards and it makes a decision.
If you want it to work with online sites you will need to first write some software to spy of the GUI of the poker software, which will take you at least a few weeks to get working properly.
How long it would take to actually make the bot depends entirely on how good you want to make it. You could make a trivial bot that just goes all in with the top 10 hands and folds otherwise in a few hours. To make a player that actually has a chance of making money would take months, and would almost certainly be impossible unless you have a lot of advanced knowledge about AI.
To put it into perspective no-one has yet made a poker bot capable of making a profit in no-limit tournament poker. Low limit ring games is a little bit easier, but its a very complex problem.
