26 replies to this topic

[Ranteur]

I'd like to request a cellblock autoplayer to be made. Yes I realize that in the past, one was being worked on, but was discontinued or delayed due to the unknown mechanics of the game. Instead, rather than having a AI that does the cellblock for me, just have one that does the same pattern over and over, because I've found one that works 90+% of the time.

Constantly having to click through this is a pain, and I'm aiming for the gold medal, I got through the first tournament in like 5 minutes of using the same pattern over and over, but yeah, an autoplayer would be great.

Just have it do this
Made by: Dave

[Scot]

Infamous Joe cellblock AP

>_>

[iargue]

We could do that. That would be easy. What we want to do is make it a complete algorithm so it beats the last boss and everything.

If people really want it to just be like this, I'll make the program and release it in a day or two.

[Boggart]

yeah honestly, it's a more expensive way (since you'll lose every once in a while) but using a cellblock AP I got the gold trophy in a day

Of course you have to be able to play tournament 11 by yourself because no AP can predict rocks.

[Ranteur]

Make it so that it auto stops on tournament 11. 200 np is a very small price to pay.

[jcrdude]

Use a Macro program if you want it to click the same locations over and over again?

To compensate for random events, press "Page Down" on every page load.

[Ranteur]

Well, I would, but I have no experiences with creating macros and programs. Haha.

[jcrdude]

Well, I would, but I have no experiences with creating macros and programs. Haha.

Creating macros: Easy and free

http://www.spacetornado.com/DoItAgain/

It maps your clicks and button presses then can repeat them ad nauseum

[weke]

that is just nifty

[Noitidart]

We could do that. That would be easy. What we want to do is make it a complete algorithm so it beats the last boss and everything.

If people really want it to just be like this, I'll make the program and release it in a day or two.

I think that would be awesome. At least have an option for it to repeat this. When we get our algorithm done we can add it in as another option.

[iargue]

Beta is ready.

Yeah....

Using that algorithim straight up....

Win rate sucks.

[07:56:47 AM] T[1] We didnt win!. Lets try again
[07:56:19 AM] [1] We won that!
[07:55:52 AM] [1] We won that!
[07:55:25 AM] T[1] We didnt win!. Lets try again
[07:54:58 AM] [1] We won that!
[07:54:28 AM] [1] We won that!
[07:53:59 AM] T[1] We didnt win!. Lets try again
[07:53:31 AM] T[1] We didnt win!. Lets try again
[07:53:02 AM] [1] We won that!
[07:52:24 AM] [1] We moved to the next level.
[07:52:24 AM] [1] We won that!
[07:51:56 AM] [1] We won that!
[07:51:27 AM] [1] We won that!
[07:50:57 AM] [1] We moved to the next level.
[07:50:57 AM] [1] We won that!
[07:50:30 AM] [1] We won that!
[07:50:02 AM] T[1] We didnt win!. Lets try again
[07:48:36 AM] T[1] We didnt win!. Lets try again
[07:48:10 AM] [1] We won that!
[07:47:42 AM] [1] We moved to the next level.
[07:47:42 AM] [1] We won that!
[07:47:16 AM] [1] We won that!
[07:46:48 AM] T[1] We didnt win!. Lets try again
[07:46:17 AM] T[1] We didnt win!. Lets try again
[07:45:54 AM] Playing Cellblock...

Thats not the 90%+ win rate decided.


The AI just kinda ignores me.




Maybe I'll add in a block detection.

[Waser Lave]

Beta is ready.

Yeah....

Using that algorithim straight up....

Win rate sucks.



Thats not the 90%+ win rate decided.


The AI just kinda ignores me.




Maybe I'll add in a block detection.

You could always do it with proper AI, minimax style.

[iargue]

You could always do it with proper AI.

I cant get the proper AI to work out. Every time I get it to play decent, I try and fix it and it totally dies.

Ask Kweh. The program seems to ignore alot of our filters.

If you want to try and help develop a proper AI. I would appreciate the help.

[Waser Lave]

I cant get the proper AI to work out. Every time I get it to play decent, I try and fix it and it totally dies.

Ask Kweh. The program seems to ignore alot of our filters.

If you want to try and help develop a proper AI. I would appreciate the help.

What have you done so far with it? What kind of algorithm or logic are you using?

[iargue]

What have you done so far with it? What kind of algorithm or logic are you using?

We tried MinMaxing using a point system. He had the grid in an array, and we tried to assign points to moves based on whats near it. We kept having issues getting it to put the right priority on it. I had it working once, and it won almost straight through Tourney 1 and 2, but made some mistakes, so I tried to correct those, and broke it again. The biggest problem I see is that it kept playing on blocked rows. (Like, the enemy plays three, you block on end, and the next move it plays the square after the block), and no matter what rules we set up, it wouldnt listen to them. Never broke on them.

I can send you our most recent test algorithm with small explanations if you want.

Also. Things picked up :|

[08:22:08 AM] [2] We won that!
[08:21:41 AM] T[2] We moved to the next level.
[08:21:41 AM] [2] We won that!
[08:21:15 AM] [2] We won that!
[08:20:46 AM] [1] We won that!
[08:20:18 AM] [1] We won that!
[08:19:37 AM] T[1] We didnt win!. Lets try again
[08:18:16 AM] [1] We won that!
[08:17:46 AM] [1] We won that!
[08:17:18 AM] [1] We won that!
[08:16:49 AM] [1] We won that!
[08:16:20 AM] T[1] We moved to the next level.
[08:16:20 AM] [1] We won that!
[08:15:51 AM] [1] We won that!
[08:15:25 AM] [1] We won that!
[08:14:56 AM] [1] We won that!
[08:14:28 AM] [1] We won that!
[08:14:00 AM] T[1] We moved to the next level.
[08:14:00 AM] [1] We won that!
[08:13:31 AM] [1] We won that!
[08:12:59 AM] [1] We won that!
[08:12:30 AM] T[1] We didnt win!. Lets try again
[08:12:04 AM] [1] We won that!
[08:11:34 AM] T[1] We moved to the next level.
[08:11:34 AM] [1] We won that!
[08:11:04 AM] [1] We won that!
[08:10:33 AM] [1] We won that!

[Waser Lave]

Do we have a thread on this in the programming section? I'm sure with more people looking at it it could be improved.

[iargue]

Do we have a thread on this in the programming section? I'm sure with more people looking at it it could be improved.

Er. No

SL. Me, Kweh, and supposed JCR have all looked at this :|.

I'll add a thread. Maybe someone will take a look into it

[fxckyouguys]

good luck, i've been meaning to play cellblock for a while but i've been too lazy. ;P

[Tsunade]

I'm very interested in an intelligent one that can play the 10+ tournaments with the stupid rocks

[Waser Lave]

I'm very interested in an intelligent one that can play the 10+ tournaments with the stupid rocks

We're looking at it, it's not simple though.

[Tsunade]

We're looking at it, it's not simple though.

Brilliant, because I've played through 10-12 myself and I'm struggling hard... I just want that stupid avatar

[Tsunade]

I've had this plauging my mind the last few days.

Cellblock is just a simple m,n,k game and I know there are those people who work out perfect play for these games, Gomoku is solved but it is a 15,15,5 game compared to 10,10,5 of cellblock.

This obviously doesn't take into account how the neo AI plays but I did find this paper and hope it helps....

Paper

[Waser Lave]

Already read that. It's one of the methods we're trying out.

[Tsunade]

Already read that. It's one of the methods we're trying out.

Damn, this is going to be awesome when you get it working.

I'm guessing you have read this as well, it makes little sense to me though Wikilinx

[Waser Lave]

Damn, this is going to be awesome when you get it working.

I'm guessing you have read this as well, it makes little sense to me though Wikilinx

We don't need to measure game complexity, we just need to make the best moves.