Computer Science Canada :: View topic - Compsci.ca's own contest, would you like to have one?

Computer Science Canada

Compsci.ca's own contest, would you like to have one?

Author:	MysticVegeta [ Thu Jul 14, 2005 8:20 pm ]
Post subject:	Compsci.ca's own contest, would you like to have one?
Ok here i go, I was lying down and I was thinking about school's wonderfull days when we used to have Math, Programming contests (programming mainly). So many of them CCC, Dwite, ECOO, Pascal, Fermat, etc. Then i thought, well i could still participate in topcoder.com but found out that an average person there had a X100 more better problem solving capability than mine. Futhermore, it used C++ (which i happen to know very little of) (dont laugh) using some public stucts and classes and whatnot (stuff that goes over my head, weeeee) Then it thought, wouldn't it be nice to compete our own friends? click compsci.ca Then, I talked to Tony about it, he told me he would think about it if the contest followed a better format than CCC/Dwite's old boring way. A format that would not only attract people from compsci but also the outside people. So, if you have a good idea as to how to improve the contest format in any way, to make it more creative, please feel free to post it here! because to a grand success... - "It only takes an idea" Oh, please vote in the polls, and any type of comments/suggestions are welcome!! Lets make this contest a reality!!!! W00T!!

Author:	Notoroge [ Thu Jul 14, 2005 8:28 pm ]
Post subject:
... Lay off the sugar man.

Author:	Delos [ Thu Jul 14, 2005 9:20 pm ]
Post subject:
Lay off the blueberries.

Author:	[Gandalf] [ Thu Jul 14, 2005 9:49 pm ]
Post subject:
Lay off the psychedelic br0wnies. It seems like a pretty good idea, but it would take some work. Let's not have too much math in it though (I've never been too good )

Author:	Andy [ Fri Jul 15, 2005 4:25 am ]
Post subject:
this was actually thought of about an year ago, but think about it, who has the time and money to start one, and how many schools will actually participate?

Author:

MysticVegeta [ Fri Jul 15, 2005 5:02 pm ]

Post subject:

Quote:

money

eh...? why money...? We have lotsa geniuses here to make questions, and it shouldn't be like a formal school vs school contest, individual contest!

besides the judges will mark it over a few days period depending on how mnay ppl participate. we already have 50% saying yes. I think it will rise soon.

code:

sugar

Whats that supposed to mean?

code:

bluberries

was that a counter to his post or mine?

Author:	Hikaru79 [ Fri Jul 15, 2005 5:31 pm ]
Post subject:
I think it's a neat idea. I was planning on posting something similar where I challenge everyone here to write a decent Chess AI and then at the end of two months or whatever, we play them off and crown a winner. Andy, money? Nah, we do it for the prestiege And no need to get schools involved, just have it be inter-forum. MysticVegeta, if you want help organizing, writing, etc, this contest, PM me. I think it's a good idea, if everyone will stop being so cynical about it

Author:	Tony [ Fri Jul 15, 2005 6:40 pm ]
Post subject:
MysticVegeta wrote: we already have 50% saying yes. And only 10 users bothered to answer the poll. I'm not putting together a contest for 5 people. Sorry. I think you need a more solid idea then just "lets have a contest". As I said before - put together a plan. Who's writing the questions? What kind of question will there be? Who's marking the answers? How are they marking the answers? What do participants get out of this?

Author:	Dan [ Fri Jul 15, 2005 8:03 pm ]
Post subject:
And who is paying my laywers bill's when i get sewed form this some how?

Author:	Delos [ Fri Jul 15, 2005 8:06 pm ]
Post subject:
MysticVegeta wrote: was that a counter to his post or mine? Neither.

Author:	MysticVegeta [ Sat Jul 16, 2005 8:54 am ]
Post subject:
Hikaru : Woot thanks a lot, if the contest bill is passed i will sure as hell talk to you about it. Chess A.I is a pretty neat idea since it hasnt been made in Turing (none that i know) But there is only one problem, to make chess A.I, one has to be good in chess, so people who are not good/dont play might not participate. Tony wrote: MysticVegeta wrote: we already have 50% saying yes. And only 10 users bothered to answer the poll. I'm not putting together a contest for 5 people. Sorry. I think you need a more solid idea then just "lets have a contest". As I said before - put together a plan. Who's writing the questions? What kind of question will there be? Who's marking the answers? How are they marking the answers? What do participants get out of this? yea, all valid questions, i think we should wait for more polls meanwhile i will prepare a list of ppl who will answer your questions, and if the poll reaches a successful point, i will ask them to do a favour.. kindly lol

Author:	[Gandalf] [ Sat Jul 16, 2005 5:12 pm ]
Post subject:
Quote: Hikaru : Woot thanks a lot, if the contest bill is passed i will sure as hell talk to you about it. Very Happy Chess A.I is a pretty neat idea since it hasnt been made in Turing (none that i know) But there is only one problem, to make chess A.I, one has to be good in chess, so people who are not good/dont play might not participate. Well, the reason it hasn't been done is because it would be hard - really hard to make. You don't really have to be good at chess to make an ai, you just have to know the rules and know how chess values work. Then you have to make an amazing move search algorithm and you still have some stuff to do... Optimize how long it takes to move, have some opening move databases, etc... Well, good luck getting ideas for the contest.

Author:	MihaiG [ Sat Jul 16, 2005 5:32 pm ]
Post subject:
i will volunteer martin to pay youire bills dan

Author:	MysticVegeta [ Sat Jul 16, 2005 6:15 pm ]
Post subject:
[Gandalf] wrote: You don't really have to be good at chess to make an ai, you just have to know the rules and know how chess values work. Then you have to make an amazing move search algorithm.. Hmm, move search algorithm depends on where you want your program to think about making a move and that requires a basic strategy of killing the king, which in turn requires some good knowledge of chess. Ex: If queen is on c6, and bishop on e5 then queens next move will be to kill the pawn to checkmate. Well, in that case, thats a very common strategy, so the person who has programmed the AI has to make a function to avoid that move, which means he has to be experienced (or should play other ppl a lil) Or maybe we should ask alikhan about this, he made an AI chess in java, he might know something.

Author:	lyam_kaskade [ Sat Jul 16, 2005 9:12 pm ]
Post subject:
Maybe we could make a chess AI that learns from each game that it plays. And why stop at chess? Imagine the possibilities of having a program that can rewrite itself... Seriously though, a Chess AI sounds like a lofty goal. I better get to work.

Author:	Notoroge [ Sat Jul 16, 2005 10:20 pm ]
Post subject:
Didn't I see some crazy Tic-Tac-Toe AI on this site before, made in Turing? Something like, it calculate 100 possible different possibilities and weant with a scenario that would make it never able to lose?

Author:	[Gandalf] [ Sat Jul 16, 2005 10:42 pm ]
Post subject:
Yeah, but I was reading about writing chess ai's for years, that's how I first started programming. I was learning about making chess ai's in C++ before I knew anything about it . Of course it would help to know how to play pretty good, but to make a good ai you only need to know the basic moves, piece values, and checkmate conditions. Then just make the program search all the moves, make a 'priority' database, and store all the calculations there for further learning and use. It's not a perfect method, but it works if you know just the basics. I'm not saying it's easy in any sense, no, not at all . edit yep, for those kinds of games, you could make a perfect ai. The reason why a chess one is so hard is because there are 110*somehugenumber possibilities.

Author:	md [ Sat Jul 16, 2005 11:28 pm ]
Post subject:
I wrote a checkers AI that did that, looked like 8 moves ahead before it got too slow... the best part was that somehow I managed to get it to cheat and make moves that you weren't supposed to be able to take. It was lots of fun to make it play itself... very interesting...

Author:	Dan [ Sun Jul 17, 2005 2:12 am ]
Post subject:
Notoroge wrote: Didn't I see some crazy Tic-Tac-Toe AI on this site before, made in Turing? Something like, it calculate 100 possible different possibilities and weant with a scenario that would make it never able to lose? Tic-Tac-Toe Ai is very easy to write, i my self have made unbeatable (but not untieable) Ai for tic-tac-toe in just few 100 lines and that included the code for playing and drawing the game.

Author:	Cervantes [ Sun Jul 17, 2005 7:54 am ]
Post subject:
Notoroge wrote: Didn't I see some crazy Tic-Tac-Toe AI on this site before, made in Turing? Something like, it calculate 100 possible different possibilities and weant with a scenario that would make it never able to lose? You mean bugzpodder's 4x4x4 tic tac toe?

Author:

Tony [ Sun Jul 17, 2005 3:00 pm ]

Post subject:

lyam_kaskade wrote:

Seriously though, a Chess AI sounds like a lofty goal. I better get to work.

Turing:

put Sys.Exec("[url=http://en.wikipedia.org/wiki/Deep_Blue]deep_blue.exe[/url]")

Your move

Author:	Hikaru79 [ Sun Jul 17, 2005 3:01 pm ]
Post subject:
Wow, my Chess AI idea seems to be catching on Honestly, I just threw that in as an example -- it was a challenge between me and alikhan and one I was planning on extending to the entire compsci.ca community, and I'm glad that people like the idea. To clarify, no, the program does NOT have to be in Turing. If we enforce that, then we'll end up with a large collection of Turing programs that all essentially do the same thing -- some minmax searching. For this to be an interesting, educational challenge, it should be open to any language I can reasonably obtain a compiler/interpreter for (to mark it). In fact, participants should be ENCOURAGED to step outside their comfort zone. Winning is really not going to be of much practical value here, while proficiency in a new language certainly will be. Sorry for hijacking your thread a bit, MysticVegeta, but why not use the Chess AI challenge as the contest? 1) People seem to like the idea 2) Doesn't take much explaining 3) It will be very easy to mark, and there won't be much room for complaints about objectivity (we play them off. You win, good. You lose, bad) The only difficulty I can foresee is deciding to what extent we should allow database searching as a valid method. A lot of great programs do it -- but then that depends a bit too much on how large a collection of .pgn's the participant can gather. Anyway, we can deliberate that point later. And for those who think that you need to be a good chess player for this, wrong. It might help for programming heuristics or whatever, but a good knowledge of the rules should be enough. Look at http://moyogo.com -- one of the most useful and advanced Go software written, and the author himself admits that he is not a Go player at all. And Go is unarguably harder to program than Chess. (Yes, I know MoyoGo is not a Go-playing program per se, but if you play it's 'a-ranked' moves against GnuGo it will coincidentally win more often than not.) So, how about it?

Author:	MysticVegeta [ Sun Jul 17, 2005 3:39 pm ]
Post subject:
we need more votes to support "Yes" then we have a chance. Waiting for polls

You can syndicate this boards posts using the file backend.php or view the topic map using sitemap.php.

Terms of Use | Privacy Policy