The A.R.B Connect 4 (8x8) Computer Program Challenge 2021

Started by Anthony.R.Brown, October 16, 2021, 19:51:25

Previous topic - Next topic

Anthony.R.Brown

Hi

I would like to Challenge all Programmers using any Programming language,to try and beat my Connect 4 program below...

THEARBCONNECT4V8X8PROGRAM {The Strongest Program in the World!} A.R.B :)

http://www.petesqbsite.com/phpBB3/viewtopic.php?f=3&t=14879


Xerra

I'm guessing modesty isn't one of your strong points judging by how you describe your connect four program on the forums you've linked to. Not to mention posting this thread twice :)

I'll tell you what. Put some kind of at least basic graphics display on your connect four game, rather than a hard to decipher console window text character output and maybe you'll get some more interest. Connect four is circles on a 8 * 8 playing field. it's not loads of text chars and log files.

Also maybe refer to the classic 80's film war-games for how to basically see Connect 4 as a pointless game when you get two reasonably skilled players. It's just noughts and crosses with a bigger playing field and red/yellow tokens.
M2 Pro Mac mini - 16GB 512 SSD
ACER Nitro 5 15.6" Gaming Laptop - Intel® Core™ i7, RTX 3050, 1 TB SSD
Vic 20 - 3.5k 1mhz 6502

Latest game - https://xerra.itch.io/Gridrunner
Blog: http://xerra.co.uk
Itch.IO: https://xerra.itch.io/

3DzForMe

Quote.Also maybe refer to the classic 80's film war-games for how to basically see Connect 4 as a pointless game when you get two reasonably skilled players. It's just noughts and crosses with a bigger playing field and red/yellow tokens.   

The fun part was the pastime with your siblings.... Happy Dayz.

+1 for Xerras feedback. I was a connect 4 fan back in the day.....
BLitz3D, IDEal, AGK Studio, BMax, Java Code, Cerberus
Recent Hardware: Dell Laptop
Oldest Hardware: Commodore Amiga 1200 with 1084S Monitor & Blitz Basic 2.1

DaiHard

Since Connect 4 is solved, and is a win for the first player, presumably anything approaching a "strongest player in the world" will draw 5-all in your contest?

Arguably more interesting is to make an attractive version for humans to play against - perhaps deliberately making it sometimes beatable (for example, by playing first in an outer column, or by limiting it's look-ahead, and forbidding lookup tables)?

Best wishes,

D

Derron

Just make it interesting by having 2-3 extras for each player.
- bomb which explodes 3x3 grids
- some diagonally thrown spears destroying \ / ...
- some horizontal and vertical destroyers (-- and | )
- collapse the "coins" afterwards - potentially leading to a "4 in a row"
- if both players then have a "4 in a row" at the same time, "destroy"/eliminate both's "4 in a row" coins without a "winner". If there is more "4 in a row" for one of the players in a "collapse", this one wins.

This way it becomes like a match-4 :-)


bye
Ron

Anthony.R.Brown

#5
Quote from: Xerra on October 16, 2021, 22:44:17
I'm guessing modesty isn't one of your strong points judging by how you describe your connect four program on the forums you've linked to. Not to mention posting this thread twice :)

I'll tell you what. Put some kind of at least basic graphics display on your connect four game, rather than a hard to decipher console window text character output and maybe you'll get some more interest. Connect four is circles on a 8 * 8 playing field. it's not loads of text chars and log files.

Also maybe refer to the classic 80's film war-games for how to basically see Connect 4 as a pointless game when you get two reasonably skilled players. It's just noughts and crosses with a bigger playing field and red/yellow tokens.

It always amazes me how people can read something quick ? and make so many mistakes concerning what they read ? :(

First of all modesty has nothing to do with a programmers ability ?
Second I have only post once each of my two types of Challenges! (take a deeper look they are different!) ?
Thirdly the Challenge is about the strength of play of the programs and nothing to do with fancy Graphics etc. if you want to play those type of games ? then get a Playstation :( ?
Fourthly another mistake by you is in the film WarGames (1983) is about how pointless it is to play Tic-Tac-Toe because it's a Draw! Connect 4 on an 8x8 board is a much different game! so show me a normal program running on a home PC that can play the game perfectly like Tic-Tac-Toe and I for one will be amazed! there is a hint that it has been solved ? but that was if it's true played under perfect conditions using a lab Mini Super Computer,with a vast opening book,put in by Humans (so it's kind of cheating!) This Challenge is for as I said a Normal size program people can download and run on their own Normal PC!




A.R.B :)
Below is the Wargames link...

https://www.youtube.com/watch?v=F7qOV8xonfY

bplus

8 x 8 solved for player 2?
https://www.littlegolem.net/jsp/forum/topic2.jsp?forum=80&topic=77

(I found a link to playing any board size by math guy at Berkeley but it was not working, hmm the link or the theory? Have yet to find solid link to 8x8 Solution except that link above. ARB did you litter at that forum too, LOL)

My theory that if you played "perfectly" it would end in a tie (like Tic Tac Toe) appears to be wrong.
Yet, as I say haven't found solid link.

PS the guy who originally put the game out as 7 X 6 board already knew there had to be a winner in a "perfect" game because the board is not symmetric.
1 person likes this

Anthony.R.Brown

Quote from: bplus on October 17, 2021, 15:49:25
8 x 8 solved for player 2?
https://www.littlegolem.net/jsp/forum/topic2.jsp?forum=80&topic=77

(I found a link to playing any board size by math guy at Berkeley but it was not working, hmm the link or the theory? Have yet to find solid link to 8x8 Solution except that link above. ARB did you litter at that forum too, LOL)

My theory that if you played "perfectly" it would end in a tie (like Tic Tac Toe) appears to be wrong.
Yet, as I say haven't found solid link.

PS the guy who originally put the game out as 7 X 6 board already knew there had to be a winner in a "perfect" game because the board is not symmetric.

Hi bplus

Yes it seems like all talk to me about a solution to the 8x8 version of Connect 4,as no one has post a working version of a program that can solve it ?

By the way to show just how complicated it is,I will be posting my 3rd update to the AI for my program after a lot of work trying to solve it! look out for version 20/10/2021 it will be my final Challenge version for 2021,if no one can beat it in a 10 game match during 2021,then my program will go down in History as being the best for that year! :)

bplus

Ha well as you (@ARB) already proved my AI is fallible, thus might be fun for humans to play, at least until they find the pattern of weakness and get tired of exploiting. Sorry not getting back to you at Steve's forum but dang ProBoards, mixed up passwords, I did visit 3 and now I am down to one. Too bad you didn't visit 3 years earlier when I was building my AI for Connect 4 for 8x8 boards.

I do like Derron's game variation suggestions, one reminded me of old game Click-O-Mania, I think was the name. I had to smile imaging mixing those 2 games together.
1 person likes this

DaiHard

Hi Guys,

Yes, I missed the significance of the 8x8!   :-[ I was only thinking about the standard 7x6. It's not clear that's solved for 8x8, though bplus's post is suggestive - albeit only with a big opening book.

Best wishes,

D

DaiHard

This link:
https://tromp.github.io/c4/c4.html
suggests he has proven 8x8 to be a player 2 win with optimal play, but I can't see a link to the proof...

Anthony.R.Brown

Quote from: DaiHard on October 18, 2021, 09:38:22
This link:
https://tromp.github.io/c4/c4.html
suggests he has proven 8x8 to be a player 2 win with optimal play, but I can't see a link to the proof...

Yes! DaiHard ? :( all a mystery and even worse no one has seen a so called Solved Connect 4 Version (8x8) working program ???




A.R.B :)

Anthony.R.Brown

Quote from: bplus on October 17, 2021, 18:48:45
Ha well as you (@ARB) already proved my AI is fallible, thus might be fun for humans to play, at least until they find the pattern of weakness and get tired of exploiting. Sorry not getting back to you at Steve's forum but dang ProBoards, mixed up passwords, I did visit 3 and now I am down to one. Too bad you didn't visit 3 years earlier when I was building my AI for Connect 4 for 8x8 boards.

I do like Derron's game variation suggestions, one reminded me of old game Click-O-Mania, I think was the name. I had to smile imaging mixing those 2 games together.

I think I understand what you are saying bplus ? can you not reset your passwords via email with ProBoards ?

By the way your Connect 4 AI is excellent it was only my programs learning ability which enabled me to find winning ways! :)

And for anyone who thinks they can play well ? below is bplus's program download...

https://www.qb64.org/forum/index.php?topic=3365.0

Anthony.R.Brown

Below are some Screen shots of my Connect 4 playing a game with the Move scores output file...


bplus

Quote from: DaiHard on October 18, 2021, 09:38:22
This link:
https://tromp.github.io/c4/c4.html
suggests he has proven 8x8 to be a player 2 win with optimal play, but I can't see a link to the proof...

@DaiHard Thanks for the link! I think I saw something similar in the past and did not bookmark.
I just tried the link to the proofs there and surprise! Once again the proof is no longer available but "maybe we can Help" and they show a screen with a bunch of price tags, yeah a sucker is born every second.

So I don't know, is it possible to copyright the proof and charge people to see it, say at games forums or is the proof still out there waiting to be properly discovered?

This is turning into a sort of game, looking for solid proof of an 8 x 8 solution, the rumors and legends and talk are certainly out there but show me the proof! :)

ARB posted a link to my very average AI game for Connect 4 8X8 board. On the off chance someone from here actually is interested, I will leave a zip with compiled program for Windows and source at top of the thread there, ASAP.
https://www.qb64.org/forum/index.php?topic=3365.0
(OK done, only 286 LOC! .exe compiled for Windows)

Again ARB's program has beat the AI (so he claims with screen shots at Steve's forum and so too some QB64 forum members), I'd like to see his source code for proof but he is very closed handed with it so... there you go another difficulty in the search for truth ;-)) It's true if you have something really good why not keep it secret? On the other hand, con men act the same way!
1 person likes this