On Nagel’s Truth-Frequency Theory of Probability