Came across this article on arxiv.com showing how π can be generated by flipping coins so spent/ wasted a few hours adding it to my coin-flipping website. It will popup as an experiment around 14 March, but you can force on other days of years with https://flipmeacoin.com?piday
In experiment mode it rains coins (10 a second) and keeps track of the above (from paper). The longer it runs, the closer (in theory) the estimated-π ought to cover to π. Anyway.