Millions of people play solitaire (patience) every day. One of the (many) embarrassments of the applied probability community is: WE CAN’T FIGURE OUT SOLITAIRE!. What’s a good strategy? If one plays well, what’s the chance of winning? How does this depend on the rules? In Vegas, you can ‘buy a deck’ for $52 and get $5 for each card played forward (rules; go through once, one at a time, no splitting piles (a la Microsoft)). How close to fair is this? I will report some machine learning results (with Ben VanRoy and Robbie Yan). The main point of the talk is to introduce a simple solitaire that we can solve. One point is the connection with the deepest mathematics. Symmetric functions, tableaux combinatorics, Baik-Deift-Johansson theorem. A second point is that this variant, patience sorting, has recently emerged in semigroup theory!–work of Cain,Malheiro,Silva and others. In this introductory talk, I hope to start a conversation between these two communities.