Path: netnews.upenn.edu!newsserver.jvnc.net!igor.rutgers.edu!aramis.rutgers.edu!kjc From: kjc@aramis [OLD RUTGERS ADDRESS] (Kelly J. Cooper) Newsgroups: talk.bizarre Subject: Re: Guide to Horndogs II Message-ID: Date: 3 May 94 23:01:09 GMT References: <2pjgcr$6dr@dorito.cs.umd.edu> <2q31mh$5tv@senator-bedfellow.MIT.EDU> Organization: Psychology @ Rutgers University Lines: 31 In article , eds AT vivaldi DOTTY media DOTTY mit DOTTY edu (Eric Scheirer) writes: > My statement was made based only upon my knowledge of what Thomasc's legs > look like, and an intuition, based on years of observation, of the prior > distribution of the skinniness of legs among the species. In particular, > I was observing that there are very few adult humans whose legs are > more skinny than Mr Colthurst's; and thus, the probability of *your* > legs being skinnier than his is quite small; and further, that the > Bayesian expected cost with respect to said probability of eating the > aforementioned Athena termical was negligible, even in the absence of a > competing, negative-cost hypothesis. But can you really consider the problem in Bayesian terms without a much clearer base-rate? I think I'd do a considerable amount of fact gathering to get more accurate numbers before committing myself to something like consuming a computer. Unless of course Thomasc's legs are so astonishingly thin that they've been classified as some sort of record (in which case the base rate has been pre-established for you). And really, does the expected utility of the response to the humor involved outweigh the possible costs, no matter how slight the risk? Of course, although the failure to follow-through on the threat is measurable only in subjective terms, it is not a fully actionable likelihood. Thus the risk _is_ in fact negligible, but only if you (and your system) are equipped with the ability to parse a joke. Hmph. I think I'll go back to my moldy M&Ms. -- Kelly J. Cooper ...Melting my mouth.