Use of Probability - 2 small points

[John Aspinall <jga@xxxxxxx>]

[1] Conditionality is not causality.  (The Jaynes book has a good
discussion about this issue.)  While a circular set of *causal*
dependencies should set off alarms, a circular set of conditional
dependencies is no big deal.  After all, Bayes Rule relates P(A|B) and
P(B|A).

[2] Eric Baum, currently at NEC I believe, has done some work on Bayesian
game players.  See http://www.neci.nj.nec.com/homepages/eric/eric.html .
The fact that your opponent "conspires" to chose the worst move against you
must be taken into account, of course.  It can be, without invalidating the
entire approach.