The Law of Large Numbers Can Fail!

We show that the law of large numbers can fail if some of its assumptions do not hold. We illustrate this using the St. Petersburg paradox and the Cauchy distribution. We also show a real-world dataset where extreme values completely distort the sample mean.