On Sat, 31 May 2003 10:17:01 -0700 (PDT), Adrian Tymes <[EMAIL PROTECTED]> wrote:
>> >if you can >> >demonstrate that P(X_fail) and P(Y_fail) are each >> at >> >most 1E-3 (say, run 1000 test flights in which >> neither >> >X nor Y fail) >> >> You need 2998 flights to demonstrate that with 95% >> confidence. > >Eh? 1000 runs where X does not fail, therefore >P(X_fail) < 1/1000. Ditto Y. These just happen to be >the same 1000 runs. Your probability numbers are not very meaningful unless you can demonstrate that they have some correlation with reality. This correlation is called confidence. If you only do 1000 runs, your confidence in your P(X_fail) being less than .001 is only 68%. That's a 1 in 3 chance that your good luck is just that, luck. Most experimenters use a 95% confidence level. It's three times as strict in terms of number of samples it requires. >I thought 1E-6 was the threshold for a single launch, >while 30E-6 was the threshold for a group of launches. >My calculations here were for a single flight. You need to demonstrate EC < 30E-6 for every flight. -R -- "We've all heard that a million monkeys banging on a million typewriters will eventually reproduce the entire works of Shakespeare. Now, thanks to the Internet, we know this is not true." -- Robert Wilensky, UC Berkeley _______________________________________________ ERPS-list mailing list [EMAIL PROTECTED] http://lists.erps.org/mailman/listinfo/erps-list
