"Is the fetus growing in [pregnant woman you both know]'s uterus a boy or a girl." There is a tremendous amount of people who think that order of birth, shape of belly, pulse on pinky, etc can be used to predict the baby's sex [anecdotally these results, combined with peoples' guesses come to 50/50, go figure!] , which if enough of them agreed could arguably be used to marginally move away from the base of 51/49. This probability could be updated after an ultrasound, but still not move to 100% as there are well documented error rates associated with sexing a baby via ultrasound. Even more fun, these error rates are result specific, (i can't remember the numbers, they are buried in a baby book) a "boy" result is wrong less often than a "girl" result, and I seem to remember there is user error involved (some technicians are wrong less often than others). If a blood test is done on a fetus it could be used to further update the probability. The final test, inspection upon birth, which will update the probability again, will not quite get to 100% unless tests were done to rule out a rare abnormality (ie apparently a boy, but has an ovary)