I read a really interesting post on Reddit explaining Markov Chain Monte Carlo, which drew the analogy between MCMC and a long game of Frogger.

Frogger is a game where you play a frog trying to cross a road by jumping. Imagine that you keep track of each jump from place to place. If the frog dies at a position, you know the position is dangerous. Now imagine repeating a whole series of trials and working out the probability distributions of each positions by aggregating the trials.

It's a neat analogy!