If you come up with any good ideas on how to do that, please share :) While 
existing engines don't differentiate between the quality of a win, there are 
some ways that additional information is extracted from playouts. Off the top 
of my head:
1. RAVE/AMAF: Identify effective moves inside playouts. This can turn up 
surprising moves that don't fit expected patterns.
2. Last good reply - Finds counter moves to certain attacks for use inside 
playouts
3. Criticality - I may have the term wrong, but I think Remi correlated 
ownership of certain board areas with winning the game
4. Score - Most engines discard this (improves strength), but mild use of this 
can yield more natural games. Fuego has an option to do this.

On Apr 1, 2013, at 3:08 PM, "Gabriel .Santos" <[email protected]> wrote:

> Thanks Aja.
> 
> I was just wondering if it is possible to get more relevant information from 
> the playouts and take advantage of that. 
> =D.
> 
> 
> On Mon, Apr 1, 2013 at 3:58 PM, Aja Huang <[email protected]> wrote:
>> 
>> 2013/4/1 Gabriel .Santos <[email protected]>
>>> 3 - I guess you misunderstood my idea of "think", or perhaps i just choose 
>>> the wrong word. By the question that I raise, i mean, how anyone say that 
>>> one state is better than another one if they have the same winrate in MC 
>>> methods. How could the machine determine this ?
>> 
>> If one of the states has more trials than the other, then the machine 
>> determines it is better with higher confidence.
>> 
>> Aja
>> 
>> _______________________________________________
>> Computer-go mailing list
>> [email protected]
>> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
> 
> _______________________________________________
> Computer-go mailing list
> [email protected]
> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go

Reply via email to