You're sure one lottery ticket will win, yet sure that every single ticket will lose
Imagine a fair lottery with 1,000 tickets and exactly one winner. For any single ticket, it's rational to believe it will lose, since it has only a 0.1% chance of winning. But apply that reasoning to every ticket, and you're forced to believe all 1,000 will lose, even though you know with certainty that one of them must win. How can both conclusions be rational at the same time?
Reveal the answer
Philosopher Henry Kyburg posed this in 1961 to show that "rational to believe" doesn't behave like simple arithmetic: being individually justified in believing each of many claims doesn't guarantee you're justified in believing their conjunction. It challenged a once-common assumption that rational belief should be closed under logical combination, and remains a key test case for theories of belief and probability.
— Henry E. Kyburg Jr., Probability and the Logic of Rational Belief — 1961