I recognised today’s Riddler problem and recognised that it’s reasonably easy to attack by Monte Carlo testing rather than my (weak) probability/combinatorics.
In part 1 we saw how you can estimate difficult probabilities with a large number of simulations. Every time I updated that post my poor home desktop took 10 mins to run the simulation. And it still wasn’t good enough. Also, it doesn’t scale. 10 minutes times 100 ages times 2 sexes = NO.