If this were a normal year, this past weekend would have signaled the beginning of a new college football season. The month of August would have been filled with stories of position battles, scrimmage results, and the scouting of future opponents. But, at least in Big Ten country, we only have stories of COVID testing results, online classes, and proposed new schedules. The picture literally seems to change every 48 hours.
Will the Big Ten season actually start in January? Will it get pulled ahead to November? Will the ACC, SEC, and Big 12 leadership wind up looking like genius, fools, or reckless criminals? I have no idea, and neither do the smartest people on the planets. Sports and the world is still very much in limbo. To think, a lot of us actually thought back in March that maybe “May Madness” might be possible. Sometimes being an optimist sucks.
In a normal year, I would strive to provide two weekly articles where I break down the action of the week, give some “bad betting advice” and provide a detailed analysis of the results and the odds for the remainder of the year. As for 2020, we can only play the hand that we have been dealt. While the world is still locked in chaos, it is nice to imagine a world where there is some semblance of tradition, fun, and just plain normalcy.
Fortunately, we can use the magic of mathematics to a take a peak into just such a world.
Back in March, as a part of my own personal mourning process, I prepared and “reported” on an imaginary version of the NCAA Men’s Basketball Tournament. While the outcome may have been predetermined. The entire tournament was a mathematically consistent version of what could have happened, including a historically reasonable number of blow-outs, buzzer beaters, and upsets. Somewhere in a parallel universe, maybe that is what did happen.
As for college football, I have developed the mathematical tools to do essentially the same thing. As I outlined in detail in my preseason analysis, I know the accuracy and historical variance in the preseason rankings. I also know how to generate point spreads and win probabilities for each game. With these two factors taken together, I can easily generate a self-consistent simulation of the entire college football season, right down to national championship game.
In my preseason analysis, I used this methodology to perform one-million Monte Carlo simulations of the original schedule. In effect, this is akin to performing a very large number of coin flips using an array of properly “weighted” coins. Taken in total, the one-million simulations provide what I believe to the be “true” odds for each team to win its division, conference, or even the national title.
From a certain point of view, the results of each of the one-million simulations represents the results of the entire college football season in a specific “parallel universe.” While each “universe” is generated using random numbers, the collection of results does represent a “normal” amount of chaos that one would expect to see in any given year of college football. Some teams wind up being better than expected, while some are worse. Some teams get lucky a score a big upset win or two. Other teams are not so lucky. My simulation takes all of these possibilities into account.
So, my plan for this fall is to report here weekly on the results of one such parallel universe, based on the original schedule. The trick, of course, is which universe to pick. Let’s just say that I made sure to select a universe that trends on the optimistic side, but is still reasonable, based on preseason expectations. The “season” will have its ups and downs, but I made sure that it has some satisfying moments for Michigan State Spartans fans. I think that we could all use something to feel good about.
No one knows what the future may bring. If the Big Ten can somehow play some games this fall, I may need to abandon this plan and focus on some actual football. But, for now, this is the plan going forward. It should be fun.
To close things out for today, here are the results of the seven total games that were originally scheduled for “Week Zero” from the parallel universe:
As Table 1 shows, no upsets the occurred in Week Zero in the parallel universe, but there were a couple of contests that were a bit closer than expected. Cal was a three-touchdown favorite over UNLV, but only escaped with a six-point win in Vegas. Marshall was only a six-point favorite at East Carolina, but they needed a late touchdown to avoid an upset loss.
The biggest game of Week Zero was clearly the contest in Dublin, Ireland between Navy and Notre Dame. Despite the Irish being an 18.5-point favorite, the Midshipmen were competitive through out. Norte Dame led for most of the game, but Navy cut the lead to only four points on a late touchdown. The Fighting Irish were able to cover an onside kick with 90 seconds remaining to stymie the potential comeback.
Looking ahead to Week One in the parallel universe, there are a few national games of note, including:
- West Virginia at Florida State (-4.5)
- Michigan at Washington (+1)
- Ole Miss at Baylor (-6)
- TCU at Cal (+3)
- USC vs. Alabama (-7.5)
Of course, it will also mark the beginning of the Mel Tucker coaching era in East Lansing, as the Spartans (-3) play host to the Northwestern Wildcats. Will Coach Tucker start off with a win? Stay tuned.