I was inspired by yesterday's post from Cameron Sun (
https://news.ycombinator.com/item?id=40635397) to put a couple hours into expanding this game concept. I wanted a better computer opponent, so I implemented one with expectiminimax (i.e. game theory, not really "AI") to pick optimal moves (
https://en.wikipedia.org/wiki/Expectiminimax). I also created a more basic web-based implementation with some minor UI updates to the original.
This was mainly for proof of concept of expectiminimax as a viable computer strategy for playing this game, so it's not mobile friendly and might have bugs. Code is open-source for anyone interested: https://github.com/keshavsaharia/pt3