This is a toy implementation of counterfactual regret minimization (CFR), used to solve a very simple modified game of rock paper scissors (where winning with rock wins double points, and losing with paper loses double points) and compute the nash-equilibrium strategy.
You can compile the solver with any modern C++ solver, with say
$ g++ solver.cpp -o solver -O3 -std=c++11
$ ./solver
And it should output the computed solution:
rock: 0.248074, paper: 0.252397, scissors: 0.499529