Beat the Dealer
We are currently working on a software system designed to analyze certain casino (as in Las Vegas) games and generate a winning strategy. Most gambling games have a house advantage that cannot be beaten by any strategy, but some games can definitely be beaten. For example, Blackjack can be beaten under certain circumstances by keeping track of the cards that have been played and using this information to infer which cards are left and then using this information to modify both playing and betting decisions. We have identified several other games, which are quite complicated, but which we believe can also be beaten.
Several years ago we created a software system named WIZARD that successfully predicted American football results. Successful here means an accuracy of greater than 60% against the betting line. This system was featured weekly on a local NBC TV affiliate sports program. We are working on a new WIZARD that uses similar principles to analyze betting patterns in more complicated sports games.
Every now and then we publish technical papers that are spin-offs from our research.
- Bacterial Foraging Optimization, MSDN Magazine, April 2012 (Vol. 27, No. 3), pp. 82-85
- Probit Classification using C#, MSDN Magazine, October 2014 (Vol. 29, No. 10), pp. 64-68.
- Neural Networks using Python, Visual Studio Magazine, November 2014
- Simulated Protozoa Optimization, Proceedings of the 13th IEEE International Conference on Information Reuse and Integration, August 2012, pp. 179-184