Mathametics probability question....pala...!!!!?

assuming that the earlier lot also has 20 defectives in 100'

P[none defective in sample of 50] = 80c50/100c50

= 8.793e-8

-----------------

i believe my argument will no longer be nicely received. I say the chance is 50% let a ??, let b ?? and randomly elect the values for a and b. As already said, for a ? 0, P( a < b²) = a million, it is trivial. in reality somewhat a lot less trivial is the concept P(a < 0 ) = a million/2 and subsequently P( a < b² | a ? 0) = a million and P( a < b² ) ? a million/2 Now evaluate what occurs at the same time as a > 0 For a > 0, at the same time as that is difficulty-free to reveal there's a non 0 chance for a finite b, the reduce, the chance is 0. a < b² is an same as declaring 0 < a < b², bear in mind we are in reality searching at a > 0. If this a finite period on a limiteless line. The chance that a is an ingredient of this period is 0. P( a < b² | a > 0) = 0 As such we've a finished chance P( a < b² ) = P( a < b² | a ? 0) * P(a ? 0) + P( a < b² | a > 0) * P(a > 0) = a million * a million/2 + 0 * a million/2 = a million/2 bear in mind, it is because of the countless gadgets. whatever style of period you draw on paper or on a computing gadget you'll detect a finite chance that seems to mind-set a million. yet it is because of the finite random decision turbines on the computing gadget and if we had this question requested with finite values there's a a answer more beneficial than 50%. i do not mean to be condescending, yet please clarify why utilising the Gaussian to approximate a uniform distribution is a reliable idea? are not countless numbers relaxing. Cantor at the same time as mad operating with them! :)