“Sure they dominate in their conference, but they could never do that well in the SEC.”

If I had a dime for every time I’d heard it, I’d be one rich man. While there isn’t much argument that the SEC has produced the strongest teams overall in the BCS era, how much does playing an SEC schedule hurt a team’s odds of, say, going undefeated?

I originally embarked on this bit of research to answer a simple question: “How would FSU do if they played Alabama’s schedule?” The answer I found was actually quite surprising. It turns out Florida State’s odds of going undefeated would only fall by 1.1% when playing Alabama’s schedule rather than their own.

How could this be? I compared the average and median power ratings of both teams’ opposition, which may shed some light on the reason:

FSU Opponents: 71.28 Avg, 71.00 Median

Alabama Opponents: 69.39 Avg, 72.10 Median

Surprisingly, the two schools play a very similar strength of schedule. FSU’s scheduling of Notre Dame and Oklahoma State boosts them above the ACC average, and games at Louisville and Miami keep the strength of schedule boat afloat.

Meanwhile, while Alabama faces LSU and Auburn, as well as two very underrated teams in Ole Miss and Mississippi State, their strength of schedule is weighed down by a poor FCS team in Western Carolina, and a Southern Miss team that finished the season rated lower than most FCS teams.

A significant disclaimer must be attached to these statistics though. Florida State playing Alabama’s schedule implies that they wouldn’t have to play against Alabama. If they were simply added to the SEC West and faced Alabama in addition to all of those opponents, their odds of going undefeated would be cut in half.

We often talk about the most improved teams by our own eyeball test. We’re astounded at how quickly Auburn turned around, and are certainly impressed with Missouri’s performance last year as well. But there is improvement going on around the country as well in some teams that are often overlooked. The below numbers show the +/- change in each team’s Sagarin power rating from 2012 to 2013. The difference is the amount in which each team improved over the course of the past year.

Most Improved:

Auburn +24.95

Washington +15.75

Washington State +15.37

UNT +15.29

Colorado +14.96

Houston +14.10

South Alabama +13.29

Least Improved:

LA Tech -26.50

FIU -22.76

Tulsa -21.36

Miami OH -16.47

SMU -15.14

Florida -15.10

While we can never predict injuries, arrests, and other black swans, we as statisticians do our best. With the information given at this time, I’ve used my simulation results to determine the odds of any team going undefeated. Just because your team isn’t on there doesn’t mean that they can’t go undefeated…they’re just not as likely to as Florida State right now.

Florida State (46.8%) – Our favorite to go undefeated comes as no surprise. The Noles’ video-game-like numbers combined with their favorable schedule make them an absolute frontrunner in the odds to go undefeated…again. While I say their schedule is below average, it is by no means easy. We discovered late last year that the Clemson team they whipped 51-14 turned out to be pretty good and that Duke team they beat 45-7 matched up quite well with Texas A&M. So while their high odds of going undefeated are partially due to a favorable schedule, it’s even more due to the fact that they’re just an amazing team.

Alabama (28.2%) – Yes, the SEC West is a tough road. Alabama travels to Ole Miss on October 4th and travels to LSU on November 8th, and this year’s Iron Bowl will be a heated rematch in Tuscaloosa. Those will be by far three tough challenges. But beyond that, home games hosting Texas A&M and Mississippi State are very manageable, as is the Tide’s road slate.

Marshall (21.7%) – Wait…who?

That’s right – the Thundering Herd come in at #3 on the list of most likely to go undefeated. They’re a similar version of Florida State’s case. FSU is a great team playing against very average competition. Marshall is a very good team, by no means playoff quality, but will likely be worthy of a spot in the Top 25 by the end of the year. They play one of the easiest schedules in the FBS. Last year it was debatable whether the C-USA East or MAC East was the worst division in FBS. With ECU’s departure to the AAC, Marshall’s C-USA East should be a cakewalk.

Bill Connelly dubbed Marshall the “overlords of Conference USA” and I can’t think of a better word. Led by senior quarterback Rakeem Cato and a high powered offense. they are -500 to -700 to win the conference, miles ahead of the rest. Marshall’s biggest stumbling block last year was road games, but this year Marshall’s “toughest” opponents come to Huntington. Statistically, Marshall is likely to trip up somewhere along the way. But my gut tells me otherwise, as I have a sneaking suspicion that the Herd will run the table, yet still not be invited to the College Football Playoff. And as a result they may be one of the key pieces to expanding this playoff system to eight teams.

Every May or June, a handful of bookmakers release the first odds. I’m not quite sure what their method is, but I’m happy to explain my own. In doing so it’s easy to identify lines that are just plain wrong. In my last article, I detailed the methods of generating win totals. Now I’m releasing the results from my simulations.

My favorite line released this year was Notre Dame 9.5 wins, -150 under,  in a schedule with games at Florida State, Arizona State, and USC, in addition to facing Stanford, Michigan, and a very underrated North Carolina team). The line has since moved dramatically.

Below are all the teams ranked by average win totals.

** Utah State, Washington, and Hawaii play 13 games.

How does one go about setting the win total for every college football team? A simple Monte Carlo simulation is able to do the trick! In this article I will detail exactly how you can generate your own win totals for the college football season.

Step 1: Generate power ratings of each team using a nonlinear regression (or use someone else’s). 

You can generate your own power ratings using the Solver tool in Excel for all FBS and FCS teams. In an upcoming post I will detail the exact process for doing this.

Step 2: Build a list in Excel for every college football game that is being played.

This data can be found on any site such as ESPN and formatted into columns. You must then use the VLOOKUP function in excel to look up the power ratings for each team. (You’ll need to make adjustments to those power ratings based on returning starters, but that’s a discussion for another day). Add 3.5 points to each home team’s power rating to account for home field advantage, then find the difference between the two. This difference is the point spread that you would expect for the game.

Step 3: Convert Spreads into Win Probabilities

I did a regression of wins and losses of all games based on the closing point spreads, and in doing so created the following formula to estimate win probability based on a point spread:

Win Probability = 55.63% – 1.71% * (Point Spread)

For example, if a team is a 14 point favorite, or ” -14 ” as we will say, then their odds of winning are 55.63% – 1.71% * -14 = 79.57%.

Step 4: Randomly Generate Game Outcomes

In a column next to each game, generate a random number using the =RAND() function in excel. Next to it, use an =IF function to determine the winner. In the example above, let’s say Alabama is a 14 point favorite over Oklahoma. We would write a formula as such:

=IF(rand()<79.57%,”Alabama,”Oklahoma”)

This means that if the randomly generated number is anywhere from 0% to 79.57%, Alabama will be deemed the winner, but if it is any higher then Oklahoma is the winner. By inserting these formulas for every game in an entire season, we can simulate every game. By pressing F9, a new batch of random numbers is generated and another full season simulated.

Step 5: Create a Win Totals List

In a separate column, create a list of every college football team playing. In one column, do =COUNTIF(*columns listing teams playing*,”Alabama”,”W”) and in another do =COUNTIF(*columns listing teams playing*,”Alabama”,”L”).

Step 6: Run Monte Carlo Simulation

Don’t freak out if you’ve never done this before. All you have to do is record a macro that copies and pastes each team’s win record into a column A on a new sheet, and then moves over to column B. Title tha Macro as “Macro 1”. Then go into Developer > Record Macro and create a new macro called “Loop.” The screen will open up with a development tab for you to write VBA. Type in:

i = 0 

Do While i < 10000

Call Macro 1

i = i + 1

Loop

This simple code will repeat the simulation 10,000 times (or however many you choose). Once the simulation is completed, you can find the average wins (or the mode wins) and use it as your win total.

Welcome to WordPress. This is your first post. Edit or delete it, then start blogging!