Arimaa Forum - Print Page


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

        Arimaa >> General Discussion >> variant idea "arimaa22"
        
(Message started by: Hollerschorle on May 18^th, 2016, 7:42am)

Title: variant idea "arimaa22"
Post by Hollerschorle on May 18^th, 2016, 7:42am

http://i649.photobucket.com/albums/uu214/sebheindl/Arimaa22a_zps4y0q7lzz.jpg

Hi folks,
when thinking about a variant to make the game harder for bots I had the idea of creating a wider board with more pieces. As you can see there is a new piece (please welcome the Rhino!) as second strongest piece. And yes, 11 rabbits and a third cat. And two more traps! But maybe it's not a good idea to still have a single second strongest piece because a rhino hostage would probably be too devastating. Your own rhino could roam rather freely in the opposite wing with the other elephant being really far away (compared to a camel hostage in Arimaa).
Anyway, there could be a few new interesting strategic options. What do you think?

(Of course, increasing the moves per turn to 5 or 6 would be a lot worse for bot's search depth but also quite confusing for humans. And it would change the nature of the game...)

Title: Re: variant idea "arimaa22"
Post by ChrisB on May 18^th, 2016, 1:45pm

Looks interesting! Although I tend to dislike making changes to an already great game, expanding arimaa to make it even more challenging for bot programmers could be very interesting.

on 05/18/16 at 07:42:53, Hollerschorle wrote:

But maybe it's not a good idea to still have a single second strongest piece because a rhino hostage would probably be too devastating. Your own rhino could roam rather freely in the opposite wing with the other elephant being really far away (compared to a camel hostage in Arimaa).

I'm guessing that a Rhino hostage wouldn't be overly devastating (compared to the camel hostage in current arimaa). It would take a lot of time for the hostage taker to secure the hostage at the edge of the board. During that time the hostage giver could start swarming the hostage area to hopefully free its elephant. So, I'm guessing that a Rhino hostage could be an interesting and potentially risky strategy for both the taker and the giver.

on 05/18/16 at 07:42:53, Hollerschorle wrote:

(Of course, increasing the moves per turn to 5 or 6 would be a lot worse for bot's search depth but also quite confusing for humans. And it would change the nature of the game...)

Increasing the number of steps per turn seems like a great way to make the game more challenging for bot programmers. Maybe that could work, without drastically changing the nature of the game, by having the board larger in both directions, the traps further apart, and a higher piece to trap ratio. (Maybe still four traps but more pieces.) These additions would hopefully offset the ease that the additional steps would provide for capturing or goaling. But, I guess the risk in adding steps is that the bots may overwhelm the humans in tactics for the first few plys. Humans would need to find good long term strategies to offset that near term advantage for the bots.