This implementation of the simple Rock, Paper, Scissors game includes a computer player that attempts to learn your tactics to allow it to predict what you will play next and beat you. The longer you play, the more it will learn about your choices. After about 50 games, it should be beating you.
The computer player looks for all past sequences that match the most recent sequence of moves. It looks in your play history, its play history, and your combined history for the longest matching sequences and bases its prediction on the sequence that is most statistically significantly different from truly random play.