I think that you might be confusing this with alpha-beta search for other games.
There are not a lot of reasonable "functions" for computing the value of a board. There are unambiguous situations, and the playouts are intended to run until one such unambiguous situation occurs, after which evaluating the score is straightforward -- under chinese rules, any small child could do it -- there is basically no interpretation necessary, and no vagueness about life and death, seki, etc. this is slightly oversimplified, and there are special cases, but under chinese counting, it's basically unambiguous. s. On Fri, Mar 29, 2013 at 12:14 PM, Chun Sun <[email protected]> wrote: > Hi Nick, > > Thanks for replying! > > So what decides who has won? Let's say, the search is to the "end" (which, > I agree, is a vague definition) and both players all pass. In this > situation, I'm given a board with stones, how do I decide who has won? > > w.r.t. search.... My intuitive thinking (again, I have not done any work > before) suggests that most people here grows a search tree, At the tip of > the search tree, the leaf, is a board with some stones. I assume we need to > score this leaf to decide black or white is in favor, right? So my original > question is actually: how do you decide a board with some stones is in > favor for black or for white, do you give +1, -1, or a number within a > range? > > Sorry to repeat myself, hopefully this is a little bit clear? > > The intention to my questions is to getting some reading materials to get > me started. > > Thank you, > Chun > > > On Fri, Mar 29, 2013 at 3:03 PM, Nick Wedd <[email protected]> wrote: > >> On 29/03/2013 18:08, Chun Sun wrote: >> >>> Hi all, >>> >>> I'm very new to this area. Actually, just subscribed. >>> >>> I'm seeing a lot of Monte-Carlo search related algorithms being >>> discussed here and previously and this is all great! However, I can't >>> help wondering what evaluation function do you use in the end? >>> >> >> Unless I have totally understood "Monte-Carlo" ... "in the end" means >> "after both players have passed", and the evaluation function is who >> has won. >> >> Nick >> >> >>> All I know available is gnugo's score estimate. >>> >>> Do we have anything else? Any research papers on this topic? >>> >>> Thank you! >>> Chun >>> >>> >>> ______________________________**_________________ >>> Computer-go mailing list >>> [email protected] >>> http://dvandva.org/cgi-bin/**mailman/listinfo/computer-go<http://dvandva.org/cgi-bin/mailman/listinfo/computer-go> >>> >>> >> >> -- >> Nick Wedd >> [email protected] >> ______________________________**_________________ >> Computer-go mailing list >> [email protected] >> http://dvandva.org/cgi-bin/**mailman/listinfo/computer-go<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
