Arimaa Forum - Print Page


    
      
        Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> General Discussion >> Average number of rabbits forward
        
(Message started by: Fritzlein on Oct 21^st, 2005, 3:13pm)

Title: Average number of rabbits forward
Post by Fritzlein on Oct 21^st, 2005, 3:13pm

I just posted a new graph http://arimaa.com/arimaa/twiki/bin/view/Arimaa/GraphicalStats to the Wiki which shows the historical average number of rabbits that humans are putting in the front row of their piece setups. (Scroll to the bottom of the page.) As I expected, the number initially dipped close to zero and has since rebounded, but there has been more variation than I expected. Also the average has never risen higher than 1.8 rabbits forward and currently stands near 1.4. I had thought that with blue22 routinely putting four rabbits forward the average might be slightly over 2.0, but I guess there are still enough games with no rabbits forward to pull the average back down. I am so persuaded by the 99of9 setup that I everyone else would be by now as well. ;-)

Title: Re: Average number of rabbits forward
Post by omar on Oct 21^st, 2005, 5:08pm

In some recent games against BombFast I've been using the Claude Chaunier setup with 3 rabbits in the front row and most of the rabbits on one side. I found some ways to win against BombFast this way, but I haven't tried it with the slower version or the other bots. I don't think this setup has ever been used in HH games.

Maybe we should create a Wiki page devoted to initial setups to document the various popular setups, when they were introduced and by whom. Maybe we can eventually also have a program that catagorizes the games based on setups; that would allow us to generate some stats about the winning ratios of the various setups.

Title: Re: Average number of rabbits forward
Post by Fritzlein on Oct 21^st, 2005, 6:01pm

on 10/21/05 at 17:08:35, omar wrote:

Maybe we should create a Wiki page devoted to initial setups to document the various popular setups, when they were introduced and by whom. Maybe we can eventually also have a program that catagorizes the games based on setups; that would allow us to generate some stats about the winning ratios of the various setups.

I know mouse was very interested in this type of discussion. Too bad he isn't here now. But anyway I still think it is a bit premature discuss opening theory when we don't yet understand the elephant-horse attack. If we base our opening theory on the premise that the advantage goes to whoever gets in an elephant horse attack first, then everything we say about openings will become wrong if we discover that the elephant-horse attack is unsound. And so on.

I keep being reminded that it is early days in Arimaa theory when I look at games by blue22 and others with a totally messy board. I simply don't know what to do in such positions. I'm getting better at playing when the position is "clean", but even then correct play depends on being able to punish the other fellow for wanting to complicate the position. And if complicating the position gives the other guy the advantage, even in theory, than what I know about sedate maneuvering isn't that important anyway!

Title: Re: Average number of rabbits forward
Post by omar on Oct 25^th, 2005, 6:29am

on 10/21/05 at 18:01:25, Fritzlein wrote:

... then everything we say about openings will become wrong ...

True, but it will leave a historical record of how our knowledge of the opening setups evolved over time. I think it might be good to preserve that.