Are you sure about that, or are you basing it on the word "chilling" in the title? My understanding is that Berlekamp's book is all about applying combinatorial game theory to go, and the "chilling" in question has very little to do with the "temperature-decrease" schedule in simulated annealing.
"Chilling" in combinatorial game theory means, in effect, making players pay a penalty for each move they make. The simplest games to analyse are ones in which it's always a disadvantage to have to move; these are equivalent in a certain sense to numbers; the "chilling" operation tends to turn games into simpler games. See e.g. http://senseis.xmp.net/?Chilling for a little more information.
The only connection between this and what happens in simulated annealing is that temperature happens to be a good metaphor for both. In combinatorial games, that's because the "value" of a position changes less wildly with each move in a cooled/chilled game. In simulated annealing, it's because the configuration you're looking at changes less wildly on each successful update as the temperature decreases.