> Temperature is really meant for endgame analysis. > When the board situation is too fluid to be broken down into local > situations, even assigning move values becomes speculative. > What is important though for the whole game, is the concept of the last big > move. (http://senseis.xmp.net/?Tedomari) > The temperature may not drop monotonously, but there will be temperature > drops. > If one side got in one more move at the higher temperature than the other, > it profited by the margin of the drop. > One of the many differences between the pros and amateurs is that pros are > keenly aware of this concept. > (We amateurs have our hands full looking for the next big point.) > The decision between war and peace is often influenced by a looming > tedomari. > If the other side threatens to take such a point if the game proceeds > "normally", all hell will usually brake loose instead. > (The commentary usually says "fighting for sente". But the reason they are > fighting for sente is the tedomari.) > > Stefan >
I was trying to work out what concrete value to put on the tedomari, rather than just "tedomari counts double", which applies to the final move rather the last move before a drop in temperature, and came to the conclusion that agrees with you, i.e. it is worth the drop in temperature. I think the tedomari concept could be useful in building a model to relate the move value to the ambient temperature. During the game there will be many occasions where the move value is above the ambient temperature, i.e if a player tries to steal sente, or during a Ko threat or where there are an even number of equally large independant miai moves. If there is any such move not "in focus" then the definition of ambient temperature used above would mean that the ambient temperature has "spiked" upwards, or you have to treat all moves above ambient temperature as "in focus" which may be the intention of the ambient temperature definition but this is not clear. I think if you adjust the move values by the tedomari gain remaining to each player during the sequence of foreseeable moves at above ambient temperature you can avoid having to view the ambient temperature as having risen. The move values (not including change to tedomari gain) will spike upwards, but the move values (including change to tedomari gain) can remain at the ambient temperature. I was writing out examples, but the post got too long to be easily readable. _______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
