Alabama Football: Math-based game and spread prediction for Duke

The time for predictions, spreads, and deep statistical dives is upon us. Based on Duke and Alabama football stats from a season ago, here is a mathematically based prediction of the winner and the spread.

I had the idea a year ago to make a prediction model that would accurately weigh meaningful statistics to better predict the spread of football games, specifically Alabama football games. There are many models out there all claiming to be the best, but each has flaws.

I’m not going to claim this model is perfect because math, while a very powerful tool, cannot accurately predict the random aspects of sports. Injuries, mental mistakes, weather, etc. can all affect the outcome of games and are hard to quantify and place in a mathematical formula. I will say, however, that this model uses the most important statistic in football. Points on the scoreboard.

Almost everyone has heard the joke at some point. Your friend looks at you and says, “Hey! I know who’s going to win the game this Saturday.” You take the bait and unwittingly ask them, “Who?” Of course, the response is typically something like, “The team who scores the most points.”

Well, that is essentially what this formula is going to look at. How many points will each team score and how many points will each team give up. I use a variety of statistics and basic statistical principles to generate a set of two final scores. One looking at it from the perspective of each team. I then use those two sets of final scores to create one final score prediction and spread prediction.

The main problem with this formula is as of now there haven’t been any games played so there aren’t any statistics to look at yet. I’m using the statistics from last season to generate these final scores until enough games have been played this season to make accurate predictions using current statistics. So, for at least the first two or three weeks of the season, the formula will be more inaccurate.

That is to say that this formula will probably overestimate Duke football and potentially underestimate Alabama football. Overall though, I am very happy with the final score and spread that was predicted and I think my fellow Alabama fans will be happy as well.

Enough with the math mumbo jumbo. Let’s talk predictions.

I wrote a preview of Alabama vs. Duke earlier this summer, and I talked about how much better Duke has been in recent years, but they were undergoing a rebuild this year and would probably take a step back. I still stand by what I wrote back then and my formula seems to back that theory up, although I believe Duke’s score is a little high.

Alabama is on a mission this year and I don’t look for them to take it easy on Duke. They want to come out and make a statement and I look for them to do exactly that when they take the field in Atlanta.

If you look at Vegas the current spread is set at 35. I’d take Alabama to cover that.