It's not one data point. There are many many documented asteroid impacts. If you assume that hitting ANY spot on Earth is randomly distributed (or at least ra

It's not one data point. There are many many documented asteroid impacts. If you assume that hitting ANY spot on Earth is randomly distributed (or at least randomly outside the poles), and you know the rate of meteor impacts (even within an order of magnitude), you can easily calculate the probability of destroying a city to be extremely low. Particularly given the number of cities in 1650 BCE.