gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (12/21/85)
> One of our Directors has asked me to inquire about a reputed Bell labs > study from 7 or so years ago, which he heard about at a conference. This > study was on "failure probabilities"; one of the statements or > conclusions he recalls was that if you have a string of five sequential > decisions, one after the other, each based upon the preceeding, the > reliability of the result is at the 59% level. I don't really have much > other than this to go on, so, if this comment rings a bell with you, and > you know the study (or studies) that this sort of conclusion came out > of, I would greatly appreciate it if you could mail me a reference. If > you know of work being done in this area by other organizations or > particular researchers, any comments or rumors or hearsay or pointers to > published work or theses would be welcomed. Gee, this is hardly revolutionary. If you assume a single decision can be made with 90% confidence, and successive decisions are statistically independent (important!), then the confidence for the composition is (90%)^5 which is about 59%. This is very elementary probability theory; almost any mathematically trained person could reproduce this -- it isn't any research specialty. Perhaps you are really interested in "fault tree" analysis or rare-event estimation, such as is used in predicting failure rates for nuclear reactors and other such systems where empirical data are few or non-existent.