Sports and Numbers

Why Simulate the Conference Race?

Predicting the outcome of a conference race is often very difficult because it's hard to keep track of who has played who where, and in this day and age, many series no longer play home-and-home.   Simply looking at the current wins and losses is not always a good way of seeing who's really in the lead, since the leading team may have played its easy games upfront.  This system is a good way of tracking just how likely any team is to win the title, taking into account any games they've already played. One big road victory, for example, can turn the tide. Here, we can capture that. We can also do what-if scenarios to see just how big a game truly is. For example, we can do an analysis of the KU vs. OU game to see just how much it actually swings the winner into favor.

Methodology

The predictions are based on thousands of computer-simulated conference seasons. For games that have not yet been played, the Sagarin ratings (predictor) from USAToday.com are used to predict the likelihood that a particular team will win a game. I use a formula that translates the difference in power rankings into a probability of victory. If two teams have identical ratings (after factoring in the home advantage), each team has a 50% chance of winning that game. The maximum chance a team can have is 99%, as no game is absolutely certain!

Disclaimer: These rankings are not trademarked or copyrighted.  They are merely derived mathematically from the use of existing public knowledge (records, schedules, power ratings, etc.).