Picking the Winner of Super Bowl XLVIII with Monte Carlo Analysis

With the 2014 Super Bowl in New York City, there has been talk of Super Bowls in other cold locations like Philadelphia or perhaps limiting the Super Bowl to warmer or enclosed locations such as New Orleans, Florida or California. In this article, we bring the Super Bowl to Monte Carlo — at least in the numbers world of applying Monte Carlo analysis.

Monte Carlo analysis applies a random process to solve complicated problems. For example, Monte Carlo analysis can be used to find the area of an irregularly-shaped object. By placing the object inside a rectangle and then throwing darts at the rectangle, the area of the odd shape can be computed by taking the area of the rectangle and multiplying it by the percentage of darts falling within the rectangle.

In a similar manner, football games can be modeled in a probabilistic manner based on certain random variables and statistics. Previously, I have applied Monte Carlo analyses to a variety of sports including last year’s Super Bowl, the past two BCS college football championships, basketball and tennis. This type of probabilistic approach is interesting because it can yield a better understanding of possible outcomes.

We simulated over 10,000 Super Bowls and the final score that comes up most frequently is Seattle Seahawks 24 – Denver Broncos 21. The average simulation result is Seattle winning by 1.5 points with 47 total points scored. For those interested, the latest line at Las Vegas sportsbooks has Denver favored by -2.5 and a total of 48.5. Here are some other results:

• Probability of Seattle winning: 54.4 percent
• Probability of Seattle winning by 3 or more: 48.6 percent
• Probability of Seattle winning by 7 or more: 36.9 percent
• Probability of Seattle winning by 14 or more: 19.9 percent
• Probability of Denver winning: 45.6 percent
• Probability of Denver winning by 3 or more: 40.1 percent
• Probability of Denver winning by 7 or more: 28.5 percent
• Probability of Denver winning by 14 or more: 14.5 percent

Monte Carlo simulations are interesting because they give us a more complete picture. For example, although the Broncos and Seahawks are a great matchup, the chances of either team winning by a margin of victory of 14+ points is more than one in three.

Carlton Chin is a portfolio manager, quant researcher, and sports analysis contributor at Rant Sports. Please follow him on Twitter @QuantFacts, “Like” him on Facebook, or add him to your Google network.