Subject:
|
Re: Connect Four game issues
|
Newsgroups:
|
lugnet.org.ca.rtltoronto
|
Date:
|
Tue, 16 Sep 2003 16:46:36 GMT
|
Viewed:
|
667 times
|
| |
|
|
In lugnet.org.ca.rtltoronto, Bruce Sheridan wrote:
> > I've done some searching as well. turns out Derek was right. this game is
> > a proven first player wins game.
>
> True, but, I read that the guy had an opening book of 60,000 moves requiring
> approx 1mb. Guess what nobody's going to be able to implement.
[butting in]
By my calculations there are 261,827 distinct states the board can be at by the
8th move (184,868 board combos for turn 8 itself). I don't have numbers for
further states, cuz, well, it was taking too long for my program to calculate.
As a board being stored at 3 bits per 2 spaces, that's nearly 2 megs of space
just for the 1st 8 moves. Assuming you create an algorithm that does a
look-ahead in order to win, I'm thinking that it's just plain impossible to fit
it into an RCX. So, it would still appear that unless there's a non-look-ahead
algorithmic way to win (that I'm sure some game theorist might be able to come
up with), chances are people's algorithms will have at least a chance of losing
since they have to fit into less than (say) 20 RCX's.
Further, the "correct" solution only works if you go first. Assumedly applying
the algorithm to go 2nd would yield all possible moves resulting in a loss; so
you've *still* got to design your algorithm such that it tries NOT to lose, even
if, by the winning algorithm, you're guaranteed to lose.
Here's a puzzler for anyone who's up on the game theory bit (I'm not):
Lego's "BrickGame" is a version of Connect 4, however it's got a 7x7 board
instead of the traditional 7x6. Why'd they do that? The voice in my head wants
to theorize that Lego made it a 7x7 board so that there wasn't a "guaranteed
win"; IE that a 7x7 board makes the game solution a tie rather than a victory.
Or it could just be that they just thought 7x7 looked nicer, or were avoiding
lawsuits from Milton Bradley. Anyway. I've always been curious. So, if anyone
comes up with the "correct"algorithm, I'd be curious to know how it performs on
a 7x7...
DaveE
|
|
Message has 1 Reply:
 | | Re: Connect Four game issues
|
| (...) Speaking of fitting stuff into 20 RCXs, is there a limit in this game for how many RCXs can be used? If not, doesn't that kill the idea that we can put the RCXs fact to face, to communicate with the opponent? Dang. I liked that idea better (...) (21 years ago, 16-Sep-03, to lugnet.org.ca.rtltoronto)
|
Message is in Reply To:
| | Re: Connect Four game issues
|
| (...) 'Cause you're not doing it right? I've never used any kind of protocol or error checking and I've used this in all of my multi RCX robots (and no comments about the last one...messaging was not the problem) (...) True, but, I read that the guy (...) (21 years ago, 16-Sep-03, to lugnet.org.ca.rtltoronto)
|
72 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
