COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub" source
C00004 ENDMK
C⊗;
.require "let.pub" source
∂AIL Prof. Judea Pearl↓Cognitive Systems Laboratory
↓School of Engineering and Applied Science
↓U.C.L.A.↓Los Angeles, CA  90024∞

Dear Judea:

	Thanks for your paper about alpha-beta.  As the "father" of
the algorithm, I must protest that you have solved the wrong problem.
The point of the algorithm is that it is to be used with a move-ordering
heuristic that tries to examine the best move first.  Results about
randomly generated trees seem to me unmotivated.

	Interesting results would involve making assumptions about how
often the best move was examined first, second, etc. and then determining
the expected number of positions examined.

	Is there something I don't understand about the question
you have investigated and the other questions mentioned in the paper
that makes my qualms irrelevant?

.reg