i once used the monte carlo method to create a very simple but surprisingly effective AI for 4-in-a-row. it's so easy to implement it pretty much worked on the

i once used the monte carlo method to create a very simple but surprisingly effective AI for 4-in-a-row. it's so easy to implement it pretty much worked on the first try, and as a bonus the difficulty is easily tunable by changing the number of simulations (i.e. fewer simulations mean the AI makes mistakes).