Programming C, C++, Java, PHP, Ruby, Turing, VB
Computer Science Canada 
Programming C, C++, Java, PHP, Ruby, Turing, VB  

Username:   Password: 
 RegisterRegister   
 Programming ideas.
Index -> General Programming
Goto page Previous  1, 2, 3
View previous topic Printable versionDownload TopicSubscribe to this topicPrivate MessagesRefresh page View next topic
Author Message
Hikaru79




PostPosted: Tue Aug 02, 2005 7:26 pm   Post subject: (No subject)

[Gandalf] wrote:
The problem is, tic-tac-toe isn't really fun when you can just program all the possibilities by hand in a few lines - then (if the person has any brains at all) the game will always be a tie.

buzpodder suggested Connect-4, not tic-tac-toe. Connect-4 is a much more interesting and strategic game than Tic-tac-toe! Our comp sci class spent more time playing it than doing much of anything else Wink
Sponsor
Sponsor
Sponsor
sponsor
bugzpodder




PostPosted: Tue Aug 02, 2005 8:39 pm   Post subject: (No subject)

[Gandalf] wrote:
The problem is, tic-tac-toe isn't really fun when you can just program all the possibilities by hand in a few lines - then (if the person has any brains at all) the game will always be a tie.


try my 4x4x4 tic tac toe game... pretty fast for a turing program and still undefeated also Wink

http://www.geocities.com/bugz_podder/ttt.html

connect 4 is pretty easy too... if you want something harder, try five-in-a-row... thats also solved (first player win)
and so is checkers i believe
Hikaru79




PostPosted: Tue Aug 02, 2005 8:42 pm   Post subject: (No subject)

bugzpodder wrote:
[Gandalf] wrote:
The problem is, tic-tac-toe isn't really fun when you can just program all the possibilities by hand in a few lines - then (if the person has any brains at all) the game will always be a tie.


try my 4x4x4 tic tac toe game... pretty fast for a turing program and still undefeated also Wink

http://www.geocities.com/bugz_podder/ttt.html

connect 4 is pretty easy too... if you want something harder, try five-in-a-row... thats also solved (first player win)
and so is checkers i believe

Hehehe. Undefeated with first move, sounds hard Rolling Eyes

Checkers is not solved -- only one of the openings is, and even that is considered tentative. The project details of Chinook can be found here: http://www.cs.ualberta.ca/~chinook/

And, connect-5? On one of those 19x19 boards (Gomoku)? I doubt that's been solved, the branching factor is absolutely *huge* (almost as large as Go's) and they still have world championships and stuff for it.
bugzpodder




PostPosted: Tue Aug 02, 2005 8:51 pm   Post subject: (No subject)

five-in-a-row = gomoku
it is solved (at least its claimed to be solved)

the same person who solved 4x4x4 tic tac toe (Where i first got the idea from)

i thought chinook solved checkers. i might have recalled something wrong
[Gandalf]




PostPosted: Tue Aug 02, 2005 8:57 pm   Post subject: (No subject)

4x4x4 tictactoe is the same thing. Only the first player can possibly win, but that will not happen if the other player thinks at all. The second player can never win if the player is decent. Never. It's like a 'law' of tictactoe.
Sure, it gets harder to control the bigger the 'board', but the basics are still the same.
bugzpodder




PostPosted: Tue Aug 02, 2005 9:04 pm   Post subject: (No subject)

I dont understand what you are trying to say.

try play against my engine as the first player then and see how many times you can beat it. i am willing to bet that without help you'll lose every game, even though technically its a first player win. Feel free to use the undo option as many times as you like Wink
[Gandalf]




PostPosted: Tue Aug 02, 2005 9:18 pm   Post subject: (No subject)

Well...ok, but what I am saying is that I know that it is humanly harder to follow everything with 4x4x4, but the rules are the same. Your program hasn't lost when being first because it can't lose.

*edit* Dang, it's 4:34am, and I am killing myself over my chess game. I guess it is hard...
Sleeping
bugzpodder




PostPosted: Wed Aug 03, 2005 8:59 am   Post subject: (No subject)

meh... its not like I have built the entire game tree for the game. it's just some ad-hoc heuristics applied that seemed to work well. But I am more interested in making something that creates an entire game tree for connect-4 or this game... this is imperative for a game of chess... you want to make sure you dont re-calculate the same game tree twice in a game (so store it). plus its really difficult to give a weighting to a board unless pieces were traded... even then... (while it is relatively easy for connect-4).
and chessmaster is so good because it has a huge set of opening/closing moves database. we dont have that.
Sponsor
Sponsor
Sponsor
sponsor
[Gandalf]




PostPosted: Wed Aug 03, 2005 2:52 pm   Post subject: (No subject)

Oh, but I do. I have a move database 1gb big. Optimize that, and you have a nice and quick 100mb's of good openings.
Many chess programs are good because, yes, they have fairly large transposition tables and databases, but that is not the only reason. I don't know how big programs like Chessmaster do it, but that's a massively optimised professional program.

Anyways, yeah, starting with something simpler is a pretty good idea. We can see how much input we get on it and if there's enough, later we can switch to Chess.

Quote:
plus its really difficult to give a weighting to a board unless pieces were traded... even then...

Well, if I understand you correctly, for those moves when pieces aren't traded, thats when it uses the opening book. Once that is finished, it looks ahead and takes into consideration, what pieces will be taken, but also things like how much control of the board you gain, how much you have opened up your pieces, king safety, etc.
bugzpodder




PostPosted: Wed Aug 03, 2005 2:55 pm   Post subject: (No subject)

Quote:
Well, if I understand you correctly, for those moves when pieces aren't traded, thats when it uses the opening book. Once that is finished, it looks ahead and takes into consideration, what pieces will be taken, but also things like how much control of the board you gain, how much you have opened up your pieces, king safety, etc.

exactly why the weighting is difficult
Display posts from previous:   
   Index -> General Programming
View previous topic Tell A FriendPrintable versionDownload TopicSubscribe to this topicPrivate MessagesRefresh page View next topic

Page 3 of 3  [ 40 Posts ]
Goto page Previous  1, 2, 3
Jump to:   


Style:  
Search: