My second one was actually useful to me: In the game of Risk, which I played often with others in the back of our HS Math Lab, there are rules for attrition of armies depending on the values of the dice thrown and whether you're an attacker or defender. If I and my opponent each had huge armies facing each other in adjacent countries, and I knew eventually there'd be a battle between us, my question was: Is it better to wait for him to attack me, or better if I'm the attacker?
There was no web to search in 1970, and I hadn't yet learned the math tools (combinatorics) to solve it on paper, so I wrote a FORTRAN program to use random #s to simulate 1000 throws of the dice - attacker has 3 dice, defender has 2, all ties won by defender. (Results: attacker lost ~85 armies for each 100 lost by defender).
But I'd already used BASIC for a couple of years before that for more than just fun & games. I wrote programs to number-crunch statistics like chi-squared for Biology labs.
That reminds me of fifth or sixth grade before we had learned algebra and we had math word problems. You had to somehow intuitively solve them by inspection and brute-force trial and error. In your head and on paper without a computer.