[computer-go] Hydra theory (was Hybrid theory)

[David Doshay]

I looked up borda voting on Wikipedia. I did not know this was called
Borda voting, and it might be called a zeroth-order version of what I
am thinking. Rather than just take rank order from each, I intended to
try to include other metrics, for example, some measure of distance
from top. One engine may evaluate that there is one really great move
with all others considered very bad. That is different than many nearly
equal good moves.


Cheers,
David



On 1, Feb 2008, at 2:41 PM, Don Dailey wrote:

> I'm not expert on decision theory,  but it's my understanding that  
> borda
> counting or voting is excellent way to integrate different decision
> making agents.    Of course this depends a lot on the nature of the
> decision to be made, but if you have N choices and several agents that
> are capable of ranking those choices, the whole is greater than the  
> sum
> of the parts.
>
> One of my first primitive MC programs evolved moves using genetic
> algorithms.  I discovered it worked surprisingly better to evolve a
> handful of players and borda vote the best choice.   It was  
> surprisingly
> the best use of resources I could find, based on a simple evolution
> strategy that is.
>
> I don't really understand why it worked so well.   I think it is  
> because
> any particular playing strategy is pretty brittle.   The nature of the
> evolved individuals was such that were probably full of
> intransitives.    They could beat particular strategies easily, but  
> were
> susceptible to other strategies and with borda voting you tended to  
> find
> a move that was reasonable with many strategies instead of super-tuned
> for just a few.
>
> There are many papers on making decisions using borda voting,  and  
> some
> of these papers are  not just about voting theory or sociology but
> computer based decisions too.
>
> Like I say, I don't know much about this and perhaps you do, but I
> thought I would present it just in case.    I think it's very
> interesting figuring out how to integrate knowledge based on "experts"
> or agents that have wildly varying strengths and weaknesses.
>
> - Don