Temperature

Path: CGTPath   · Prev: Minimax   · Next: Environment
  Difficulty: Intermediate   Keywords: Strategy

Temperature is a term imported from combinatorial game theory that refers to the urgency of a play. Hotter plays are more urgent[1] than cooler plays. Temperature is related to several common Go terms.

Generally, the temperature of the whole board, or simply, the temperature, corresponds to the size of the largest play. As the game progresses, the temperature tends to drop. The local temperature corresponds to the size of the largest play in a region of the board. Generally speaking, a gote play lowers the local temperature and a sente play raises it, while if the local temperature stays the same, the play is ambiguous.

The ambient temperature refers to the temperature of the rest of the board besides the local region. In general it is the value of tenuki. As a rule it is time to make a local play when the ambient temperature drops below the local temperature. Sometimes the choice of a local play depends on the ambient temperature. A gote play, which lowers the local temperature, may be played with sente if its gote response is hotter than the ambient temperature. When both a local play by either player and its gote response are hotter than the ambient temperature, the play is double sente.

If two gote plays have the same size (in other words, if they are miai), then their combined temperature is lower than the local temperature of either one. If a play is the last one before a significant drop in the temperature of the whole board, it is tedomari. Tedomari is worth fighting for, as a rule.

Note: It is often useful to think of a game temperature that never rises during the game. Even if a hot battle erupts, as a rule it heats up a local region, while the ambient temperature remains the same.

-- Bill Spight


See also Temperature and Terminology Discussion.


Kirk: I think it would be fascinating to have the local and global temperatures spelled out (estimates, at least) for a (small) game. It would start at the end, of course, where things can be calculated exactly, and move toward the beginning becoming somewhat vague no doubt as the endgame stage is left. Calculating the temperatures would point out likely errors in play. I am particularly interested in the value of non-territorial plays, and an exercise like this would help me to make some of these ideas more concrete. Has this already been done?

Bill: See [ext] http://www.msri.org/publications/video/index0.html and [ext] http://www.msri.org/publications/books/Book42/files/spight.pdf for an analysis of the final stages of the Jiang-Rui Environmental Go Game. However, they do not discuss the size of non-territorial plays. Ishida does so in one of his books, but he seems to underestimate them. (His estimates are influenced by the traditional estimate of the first move as worth 10 points by miai counting. The size of komi strongly suggests that that is an underestimate.)


[1] This is not the same sense of urgent as in Play Urgent Moves Before Big Moves. Both kinds of moves are urgent in this sense. -- Bill Spight



Path: CGTPath   · Prev: Minimax   · Next: Environment
This is a copy of the living page "Temperature" at Sensei's Library.
(OC) 2005 the Authors, published under the OpenContent License V1.0.
[Welcome to Sensei's Library!]
StartingPoints
ReferenceSection
About