Re: computer-go: new strategy game with $10,000 prize

[Måns Ullerstam <mans@xxxxxxxxxxxx>]

I agree to all mails regarding this new game Arimaa. But I think there is
one interesting aspect of developing software that can beat the best human
in this game. The interesting part is that there are still no humans that
are good in this game, no common opening theories or common ways to decide
what is good and bad.

For AI purposes this means that we have to develop software that is good
at a task where we don't know what is good or bad ourselves. There is no
way we can watch a game of Arimaa played by our software and decide if it
is playing good or bad, we can only watch the results of its play.

I think this makes it an interesting AI exercise, since our minds are not
clogged with prior knowledge of (what we think is) good and bad strategy.

/ Mans

> Hi all,
>
> well, I don't know about this game, but there isn't really any big
> problem to create "new" games that will be harder to solve than Chess or
> Go. Arimaa is basically just trying to defeat the brute-force approach
> used in Chess.
>
> Just let's imagine Go played on a 25x25 board. Even if 19x19 Go would be
> "solved" just the sheer additional space to be searched would add enough
> problems for computers. If 25x25 is not enough, then 35x35 or whatever.
> That is, if the main power of a Go playing program was raw CPU power...
>
> I think humans will still be able to play the game reasonably well,
> because the local game would be the same, just strategy would differ
> (simplified). If a "human level" Go program would use the human-like way
> of thinking, then it should be able to adapt to the new complexity.
>
> For a new game however, just as Don said, the only feasible way for a
> program to play is by raw brute-force search, because there is no theory
> to use for higher-level analysis yet.
>
> My 2 cents of ramblings :-)
> Vlad
>
> ----- Original Message -----
> From: "Don Dailey" <drd@xxxxxxxxxxxxxxxxx>
> To: <computer-go@xxxxxxxxxxxxxxxxx>
> Cc: <computer-go@xxxxxxxxxxxxxxxxx>
> Sent: Monday, January 20, 2003 5:21 AM
> Subject: Re: computer-go: new strategy game with $10,000 prize
>
>
>>
>> I took  a look at this.   I'm not sure  it's as complex as  he claims.
>> Even though the legal move count  at any given node is quite high, the
>> game is very  threat oriented (tactical).  Also, for  what it's worth,
>> there have not been hundreds of  years of human experience at the game
>> which should help some!
>>
>> So if a program can be designed that is very good at seeing short term
>> tactics, it might be very hard for a human to beat.
>>
>> $10,000 isn't worth the effort  by itself, but it seems like something
>> that would certainly be worth the fun of trying.
>>
>>
>> Don
>>
>>
>>
>>
>>    X-Sender: fotland%smart-games.com@xxxxxxxxxxxxxxxxx
>>    Date: Sat, 18 Jan 2003 15:58:57 -0800
>>    From: David Fotland <fotland@xxxxxxxxxxxxxxxxx>
>>    Content-Type: text/plain; charset="us-ascii"; format=flowed
>>    Sender: owner-computer-go@xxxxxxxxxxxxxxxxx
>>    Precedence: bulk
>>    Reply-To: computer-go@xxxxxxxxxxxxxxxxx
>>
>>
>>    http://arimaa.com/arimaa/
>>
>>    He thinks his game is 100x more complex than go, and offers $10,000
>> for
> a
>>    program
>>    that can beat a human by 2020, with $1000 for the best program at
>> the
> end
>>    of 2003.
>>
>>    I think this prize can be won this year, but I don't have time to
>> work
> on it :(
>>
>>    Any takers?
>>
>>    David