Scratch

roijac wrote:

don't really like chatbots,

Why?! 

what do you say about infinite tic-tac-toe board where you need to make a row of 5?

Infinite board=infinite possibilities=infinite computations

Last edited by Hardmath123 (2012-05-28 07:48:20)

blob8108 wrote:

Hardmath123 wrote:

roijac wrote:

what do you say about infinite tic-tac-toe board where you need to make a row of 5?

Infinite board=infinite possibilities=infinite computations

Not necessarily! Couldn't you just compute part of the infinite grid? Surely that's the trick to it...

Yeah, I guess. I still think infinite boards are problematic. How would I render it? What if a program chooses to put a piece at the 90000000000000000000000000000000th cell?

MoreGamesNow wrote:

Hardmath123 wrote:

blob8108 wrote:

Not necessarily! Couldn't you just compute part of the infinite grid? Surely that's the trick to it...

Yeah, I guess. I still think infinite boards are problematic. How would I render it? What if a program chooses to put a piece at the 90000000000000000000000000000000th cell?

You could say that you can only place a piece next to an already-placed piece.  The problem I see is what if a "race to the edge" scenario happens and both computers just keep moving in one direction forever 

Ouch. 

So why don't you guys (amcerbu and roijac) like chatbots?

amcerbu wrote:

I'm not sure.  It just doesn't seem within our possibilities.  Most competent natural language programs take a lot of time building their "vocabularies" and talking with real people.  A crucial part of a chatbot is its list of examples, exceptions, and rules to understanding English.  Cleverbot can't even carry on a conversation with itself (I opened two tabs and manually entered the answers from one into the speech of the other).

this^^

sparks wrote:

You wanted something harder than TTT and easier than hex and came up with a chatbot? 

xD

amcerbu wrote:

Cleverbot can't even carry on a conversation with itself (I opened two tabs and manually entered the answers from one into the speech of the other).

You mean like this? 

sparks wrote:

You wanted something harder than TTT and easier than hex and came up with a chatbot? 

It sounds like a fun project, but much to complex for a small competition. Consider that hundreds of pounds in cash prizes have been offered for years to a coder who can create a chatbot that can defeat the Turing test and no one has yet won.

*sigh* I hate having things put in perspective by someone smarter than me... but we don't need a super-fluent Turing-test-worthy bot; it should be able to understand basically what you say and respond, right?

jji7skyline wrote:

We could have a chatbot argument 

First chatbot to convince the other chatbot of something wins :3

XD

Or maybe an AI quiz where we ask it questions like "who was the 73rd curator" and it had to respond. Like Watson, except simpler.

I'm still open to suggestions, so if anyone comes across a simple game which needs very little framework and AI coding, please let me know! 

amcerbu wrote:

@Hardmath123- I found the Wikipedia list of abstract strategy games to be particularly useful, and I picked out a few that seemed doable/fun.  They're not in any particular order.  I've also avoided Chess and its variants; that's just too hard. 

If I had to choose my favorite ones, I would say Go and Reversi.  They're both pretty cool games, and require a lot of strategy.  Diabillik and Abalone are unusual, but really cool.  But please take the time to look a little at all of them (a few are really neat).  Sorry for the massive list. 

• Connect Four
• Dots and Boxes
• Checkers
• Breakthrough (never heard of this one, but it looks cool)
• Go (this one would be a lot of fun)
• Goats and Tigers (I used to play this with my dad)
• Abalone (haven't heard of it, but it looks cool)
• Chinese Checkers
• Diaballik (again, never heard of it, but it looks really fun)
• Reversi/Othello (this one could be interesting)

Okay, that's all I've got.  I went through everything on that Wikipedia list.  Wow. 


Card games might also be interesting.  Blackjack, or 21, for example, don't take much framework programming (just a data structure for holding an ordered collection of cards).  But I'll leave that up to the next person to think about.

Woah, you really thought this through—I appreciate that.    But I think if we're gonna settle on a board game, we'll go with Hex. I found a Tic-Tac-Toe variant which might be cool too: "Super" Tic-Tac-Toe, where you play on a sudoku-board-like-thing made up of 9 ttt boards. You play on all simultaneously, and everytime someone wins on a board, it is marked as that player's piece on the "big" board. But I see lots of tie opportunities in that. 

Hardmath123 wrote:

Woah, you really thought this through—I appreciate that.    But I think if we're gonna settle on a board game, we'll go with Hex. I found a Tic-Tac-Toe variant which might be cool too: "Super" Tic-Tac-Toe, where you play on a sudoku-board-like-thing made up of 9 ttt boards. You play on all simultaneously, and everytime someone wins on a board, it is marked as that player's piece on the "big" board. But I see lots of tie opportunities in that. 

I found two people at my school playing a different variant that I thought was really cool.  The board was the same as yours, but, when each player went on a square, the opponent had to make a move on a corresponding game.  Er, that didn't make sense, diagram time:

1  2  3
4  5  6
7  8  9

If someone puts an "X" in square 3 of game 1, the opponent (who goes next) has to make their next move in game 3.  If they play 6:3, the original player has to make their next move in game 6.  This means that, when you make a move, you have to look at what options that opens for your opponent.  It also means that, if your opponent can force a win on one board, you can still win/tie the game by simply avoiding that board.

amcerbu wrote:

I like Tic-Tac-Toe (and its variants) but it just seems too simple.  Is everyone familiar with the rules of Go?  I think it's complicated enough that a simple "check all possible moves" strategy wouldn't do the trick.  Please check out the Wikipedia page; I think you'll find it very interesting.

A repeat: if we do a board game, we will probably to Hex. But what do you think of PacMan? I sorta want to divert your thoughts of AI from Game Tree: they're different things. You can have an AI without a Game Tree, and I think it would be cool to try a game which needs a different kind of programming, some creativity. Pac-Man is great for that; it needs pathfinding and "sport" algorithms, and maybe a short game tree to predict how a ghost would move. 

Also, we need to postpone this again, because we obviously can't start tomorrow. Unless you agree on chatbots... 

Last edited by Hardmath123 (2012-05-31 05:53:39)

http://scratch.mit.edu/forums/viewtopic … 5#p1256185
http://scratch.mit.edu/forums/viewtopic … 8#p1256328
Some people don't seem to agree (probably because they have to code it, but I'd love to code a chatbot if someone convinced me to; it sounds like so much fun!).

blob8108 wrote:

Hardmath123 wrote:

Some people don't seem to agree (probably because they have to code it, but I'd love to code a chatbot if someone convinced me to; it sounds like so much fun!).

Go on then, code a chatbot! (You can do it after the pacman/Go/board game competition, if you must.) 

Okay! I'll make it in Objective-C for speed. 

"Till then, can we please decide on a game, preferably PacMan/Chatbot/Anything easy to code for both me and the competitors?

Uno?