Editor’s Note: This is the first in a series of fight predictions by guest columnist Jason Noriega. Jason works as an analyst at a U.S. National Laboratory. He has a keen interest in predictive modeling, and lectures at conferences around the United States. Jason also participates in predictive modeling competitions. He has placed in the top 10% in the following competitions using various machine learning algorithms,
Ford Motor Company – Predicting Driver Alertness
University of Melbourne – Predicting the Outcome of Grant Applications
The predictive model developed by Jason Noriega for Jason’s Picks is the copyrighted intellectual property of Jason Noriega, Nery Castillo-McIntyre, and Artes Marciales Mixtas en Español.
Wouldn’t it be incredible to have an automatic predictive model you can use to accurately predict the successful outcome of MMA fights? To know, statistically, who you should put your money on in a match. That is my New Year’s resolution for 2012: to develop a predictive model that can be used to make money in the MMA betting world.
How am I going to do this? By using state of the art machine learning algorithms that can sift through large amounts of data and variables to find predictive patterns in MMA fights. I will weigh numerous variables such as striking accuracy, submission skills, takedowns, takedown defense, reach, age, and many others.
Perfection will not come easy. I will have to make many changes to the predictive model as time goes on and it becomes evident which variables deserve more weight and which deserve less. For the first version of this model, I used a machine learning algorithm called CHAID with 12 different fighter variables. The challenge for the time being is the lack of sufficient historical data that would allow me to have confidence on the patterns I have found. However, as the year progresses, the model will only get better.
Below are my picks for this coming weekend’s UFC Rio main card based on the probability of winning. As the year progresses and I make more predictions based on the model, I shall provide an overall accuracy “rating”.