The month of August is an exciting time for college football fans. Training camp is now underway and the kick-off of the first weekend of games is just a few short days away.
In preparation for the upcoming action, I have completed a set of mathematically-driven simulations of the 2022 season, based on a consensus of the preseason rankings (of all 131 FBS teams) as the primary input. In the previous installment of this series, I reviewed the overall methodology and took a detailed look at Michigan State’s schedule and season prospects. Today, the series continues with a deeper dive into the entire Big Ten Conference.
Figure 1: Strengths of schedule for each Big Ten team's full schedule (left) and conference schedule (right). The strengths of schedule are calculated as the total number of expected wins an average power five team would earn with the same schedule
There are several different ways to estimate strength of schedule. My method makes the expected win calculation for an average Power Five team (ranked roughly No. 25 in the country) assuming that this reference team were to play the schedule of each of the 14 Big Ten schools.
Overall, Figure 1 suggests that the Michigan Wolverines are fortunate enough this year to earn the easiest overall schedule among Big Ten teams, and the third-easiest conference schedule. The Wolverines’ non-conference schedule consists of three home games against Colorado State (ranked No. 95 in the preseason), Hawaii (No. 121) and UCONN (No. 129).
Michigan also has the benefit of facing Nebraska and Illinois from the Big Ten West, the projected No. 5 and No. 6 best teams in that division, respectively. That said, the Wolverines do travel to Iowa City to open the month of October, and the Iowa Hawkeyes will likely present a bigger challenge.
In general, Michigan’s overall schedule is 1.3 games easier than the team with the most difficult overall schedule in the conference, Indiana. Overall, Michigan’s schedule is 0.85 games easier than Michigan State’s schedule, and 1.1 games easier than Ohio State’s schedule.
In just conference play, the Wolverines have a 0.36 game advantage over the Buckeyes, and a 0.54 game advantage over the Spartans. Also note that Michigan State’s overall and conference strength of schedule is on the difficult side of average compared to the rest of the Big Ten.
Based solely on this strength of schedule data, Purdue is an interesting team to watch in the Big Ten West. The Boilermakers own the easiest conference schedule of all 14 teams, in large part due to drawing Indiana, Maryland and Penn State at home as the team’s East Division crossover opponents.
Table 1: Big Ten win probability matrix derived from the results of the 240,000 cycle Monte Carlo simulation, including the known uncertainty in the preseason rankings.
Table 2: Summary of the preseason projections for the Big Ten conference, based on the consensus preseason rankings and a 240,000 cycle Monte Carlo simulation of the full college football season
The methodology used to generate all of this data was described in the first installment of this series. Very briefly, the expected win totals and season odds were all derived from a Monte Carlo simulation based on the consensus preseason rankings and their historical uncertainty. I believe that due to the inclusion of this uncertainty, the numbers above are the most accurate on the internet.
The two sets of projected records are based on a much simpler pair of simulations where 1.) the preseason rankings are assumed to be correct and the projected favorite in each game wins (the “most likely projection”), or 2.) a historically reasonable number of home upsets occur (the “disruptive projection”).
There is clearly a lot to unpack from Tables 1 and 2. In the Big Ten East, Ohio State has the best overall odds across the board. This makes sense due to the fact that the Buckeyes also hold the best overall preseason ranking at No. 2 in the country. But No. 5 Michigan, with the easier schedule, is just about five percentage points behind Ohio State in both the division and conference race. The gap between the two teams is even smaller in the odds to make the playoffs (about 26 percent) and to win the national title (seven to nine percent).
No. 16 Michigan State and No. 20 Penn State are neck-and-neck for the third-best odds in the Big Ten East. Similar to the Buckeyes and Wolverines, the Spartans have the better team on paper, while the Nittany Lions hold the easier conference schedule. Both teams have about a 13 percent chance to win the division and between a seven and eight percent chance to win the conference.
As for the remainder of the Big Ten East, Maryland, Indiana and Rutgers all have less than a two percent chance to win the division. Only Maryland looks likely to be in contention for a bowl game, but the data in Table 1 suggests that the Terrapins have just a 57 percent chance to finish the regular season with six or more wins.
In the Big Ten West, the odds (42 percent) pretty heavily favor No. 12 Wisconsin to win the division. After the Badgers, the next four teams are all very closely grouped. No. 29 Iowa has the next best odd at 15.8 percent, with No. 35 Purdue, No. 22 Minnesota and No. 42 Nebraska all with odds between 11 and 15 percent. Illinois and Northwestern are projected to finish in the basement of the West Division.
However, the disruptive simulation highlights a few different scenarios that might play out and a few key games that may be pivotal in the overall Big Ten race.
Specifically, the disruptive simulation suggests that there is a path for Michigan to make a return trip to Indianapolis even if it loses to Ohio State in Columbus, as expected. The Wolverines, due to their light schedule, project to be heavily favored in the other 11 games. So, an 8-1 conference record is projected in both simulations.
But the path for the Buckeyes is slightly trickier. Ohio State has to travel to both Michigan State and Penn State in 2022, and the disruptive simulation flags both games as potential upsets. In this scenario, Ohio State (7-2) would finish in a three-way tie for second place with Michigan State (7-2) and Penn State (7-2), and Michigan (8-1) would emerge as the champion in the East.
In the Big Ten West, both the most likely and the disruptive simulations pick Wisconsin to win the division. However, both simulations suggest that the Badgers will lose two or three conference games. Furthermore, Table 2 flags Purdue as a potentially interesting dark horse in the West Division.
Purdue has tough road games at Wisconsin and at Minnesota, but the Boilermakers have two potentially winnable home upset opportunities against Penn State in the season opener, and against Iowa on Nov. 5. If Purdue can hold serve at home, it has a shot to finish at 7-2 in conference play, even if the Boilermakers cannot beat the Gophers or Badgers.
Meanwhile, Wisconsin has a series of challenging road games at Ohio State, at Michigan State and at Iowa. In the disruptive simulation, the Badgers lose all three games.
In this scenario, Purdue would need to avoid an upset loss at Maryland in order to claim the Big Ten West title. Unfortunately, the disruptive simulation does favor the Terrapins in this contest. That said, the path for the Boilermakers to make their first trip to Indianapolis is wider than one might expect.
It is my annual tradition to use the results of the disruptive simulation to generate my official preseason picks. As a result, I am compelled to predict that Michigan will win the Big Ten East and narrowly defeat Wisconsin in the Big Ten Championship game to claim its second consecutive Big Ten title.
Yeah, I know. I don’t like it either, but that is what the #math tells me.
If this scenario were to come to pass, Michigan would almost certainly make a return to the College Football Playoff. But, the selection of the team to represent the Big Ten in the Rose Bowl would be interesting.
No team from the West would have better than a 9-3 record, so the Selection Committee would need to sort out which team out of the East would get the nod. In this scenario, Ohio State and Michigan State would both be 10-2, and Penn State would be 9-3, but with wins over both the Buckeyes and the Spartans. In this scenario, Penn State would have a loss to Michigan in addition to a pair of early-season losses at Purdue and at Auburn. So, style points would likely play the role of tiebreaker.
Finally, while my disruptive simulation does predict a Big Ten title for the Wolverines, there is one variable that my simulation did not take into account. In the disruptive scenario, Michigan is able to beat a Michigan State team that eventually finishes with a record of 10-2.
As I have shown before, since the Spartans joined the Big Ten in 1950, the Wolverines essentially never beat Michigan State when MSU wins a large percentage of its other games (over 70 percent). In fact, Michigan has only defeated a “good” Michigan State team once in the past 70 years, and that was in 1955.
If we take this fact into account, I could certainly envision a scenario where Michigan State is able to beat Ohio State at home and upset Michigan on the road. In this case, the 11-1 Spartans would advance to Indianapolis where I would project them to be a narrow underdog in a rematch with Wisconsin. This scenario would look frightening like the 2011 season.
While I am not predicting this specific outcome, it is one worth noting.
With the Big Ten analysis now in the books, it is time to look at how the rest of the Power Five conferences might shake out. That is the subject of the next installment of this series. Stay tuned, and Go Green.
In preparation for the upcoming action, I have completed a set of mathematically-driven simulations of the 2022 season, based on a consensus of the preseason rankings (of all 131 FBS teams) as the primary input. In the previous installment of this series, I reviewed the overall methodology and took a detailed look at Michigan State’s schedule and season prospects. Today, the series continues with a deeper dive into the entire Big Ten Conference.
Strength of Schedule
To begin, Figure 1 below shows a summary of the relative strengths and schedule for all 14 Big Ten teams. The left pane shows the overall strengths of schedule, based on each team’s full 12-game slate. The right pane shows the strengths of schedule for each team’s nine-game conference schedule.
Figure 1: Strengths of schedule for each Big Ten team's full schedule (left) and conference schedule (right). The strengths of schedule are calculated as the total number of expected wins an average power five team would earn with the same schedule
There are several different ways to estimate strength of schedule. My method makes the expected win calculation for an average Power Five team (ranked roughly No. 25 in the country) assuming that this reference team were to play the schedule of each of the 14 Big Ten schools.
Overall, Figure 1 suggests that the Michigan Wolverines are fortunate enough this year to earn the easiest overall schedule among Big Ten teams, and the third-easiest conference schedule. The Wolverines’ non-conference schedule consists of three home games against Colorado State (ranked No. 95 in the preseason), Hawaii (No. 121) and UCONN (No. 129).
Michigan also has the benefit of facing Nebraska and Illinois from the Big Ten West, the projected No. 5 and No. 6 best teams in that division, respectively. That said, the Wolverines do travel to Iowa City to open the month of October, and the Iowa Hawkeyes will likely present a bigger challenge.
In general, Michigan’s overall schedule is 1.3 games easier than the team with the most difficult overall schedule in the conference, Indiana. Overall, Michigan’s schedule is 0.85 games easier than Michigan State’s schedule, and 1.1 games easier than Ohio State’s schedule.
In just conference play, the Wolverines have a 0.36 game advantage over the Buckeyes, and a 0.54 game advantage over the Spartans. Also note that Michigan State’s overall and conference strength of schedule is on the difficult side of average compared to the rest of the Big Ten.
Based solely on this strength of schedule data, Purdue is an interesting team to watch in the Big Ten West. The Boilermakers own the easiest conference schedule of all 14 teams, in large part due to drawing Indiana, Maryland and Penn State at home as the team’s East Division crossover opponents.
Big Ten Win Probabilities and Season Odds
With the strength of schedule data in mind, it is now time to take a look at the detailed projections for the Big Ten. Table 1 below shows the win probability matrix for the entire Big Ten Conference. Table 2 summarizes the strength of schedule information above as well as the consensus preseason ranking of each team, expected win total, projected records and overall odds to win the division, conference, make the College Football Playoff and win the national championship.Table 1: Big Ten win probability matrix derived from the results of the 240,000 cycle Monte Carlo simulation, including the known uncertainty in the preseason rankings.
Table 2: Summary of the preseason projections for the Big Ten conference, based on the consensus preseason rankings and a 240,000 cycle Monte Carlo simulation of the full college football season
The methodology used to generate all of this data was described in the first installment of this series. Very briefly, the expected win totals and season odds were all derived from a Monte Carlo simulation based on the consensus preseason rankings and their historical uncertainty. I believe that due to the inclusion of this uncertainty, the numbers above are the most accurate on the internet.
The two sets of projected records are based on a much simpler pair of simulations where 1.) the preseason rankings are assumed to be correct and the projected favorite in each game wins (the “most likely projection”), or 2.) a historically reasonable number of home upsets occur (the “disruptive projection”).
There is clearly a lot to unpack from Tables 1 and 2. In the Big Ten East, Ohio State has the best overall odds across the board. This makes sense due to the fact that the Buckeyes also hold the best overall preseason ranking at No. 2 in the country. But No. 5 Michigan, with the easier schedule, is just about five percentage points behind Ohio State in both the division and conference race. The gap between the two teams is even smaller in the odds to make the playoffs (about 26 percent) and to win the national title (seven to nine percent).
No. 16 Michigan State and No. 20 Penn State are neck-and-neck for the third-best odds in the Big Ten East. Similar to the Buckeyes and Wolverines, the Spartans have the better team on paper, while the Nittany Lions hold the easier conference schedule. Both teams have about a 13 percent chance to win the division and between a seven and eight percent chance to win the conference.
As for the remainder of the Big Ten East, Maryland, Indiana and Rutgers all have less than a two percent chance to win the division. Only Maryland looks likely to be in contention for a bowl game, but the data in Table 1 suggests that the Terrapins have just a 57 percent chance to finish the regular season with six or more wins.
In the Big Ten West, the odds (42 percent) pretty heavily favor No. 12 Wisconsin to win the division. After the Badgers, the next four teams are all very closely grouped. No. 29 Iowa has the next best odd at 15.8 percent, with No. 35 Purdue, No. 22 Minnesota and No. 42 Nebraska all with odds between 11 and 15 percent. Illinois and Northwestern are projected to finish in the basement of the West Division.
Projected Record Analysis
While the odds likely paint the most accurate picture of what is likely to happen, the projected records give some interesting insight. Both the odds and the “most likely” record projections suggest that Ohio State and Wisconsin will meet in Indianapolis for the Big Ten Championship, where the Buckeyes would be favored by roughly six points.However, the disruptive simulation highlights a few different scenarios that might play out and a few key games that may be pivotal in the overall Big Ten race.
Specifically, the disruptive simulation suggests that there is a path for Michigan to make a return trip to Indianapolis even if it loses to Ohio State in Columbus, as expected. The Wolverines, due to their light schedule, project to be heavily favored in the other 11 games. So, an 8-1 conference record is projected in both simulations.
But the path for the Buckeyes is slightly trickier. Ohio State has to travel to both Michigan State and Penn State in 2022, and the disruptive simulation flags both games as potential upsets. In this scenario, Ohio State (7-2) would finish in a three-way tie for second place with Michigan State (7-2) and Penn State (7-2), and Michigan (8-1) would emerge as the champion in the East.
In the Big Ten West, both the most likely and the disruptive simulations pick Wisconsin to win the division. However, both simulations suggest that the Badgers will lose two or three conference games. Furthermore, Table 2 flags Purdue as a potentially interesting dark horse in the West Division.
Purdue has tough road games at Wisconsin and at Minnesota, but the Boilermakers have two potentially winnable home upset opportunities against Penn State in the season opener, and against Iowa on Nov. 5. If Purdue can hold serve at home, it has a shot to finish at 7-2 in conference play, even if the Boilermakers cannot beat the Gophers or Badgers.
Meanwhile, Wisconsin has a series of challenging road games at Ohio State, at Michigan State and at Iowa. In the disruptive simulation, the Badgers lose all three games.
In this scenario, Purdue would need to avoid an upset loss at Maryland in order to claim the Big Ten West title. Unfortunately, the disruptive simulation does favor the Terrapins in this contest. That said, the path for the Boilermakers to make their first trip to Indianapolis is wider than one might expect.
It is my annual tradition to use the results of the disruptive simulation to generate my official preseason picks. As a result, I am compelled to predict that Michigan will win the Big Ten East and narrowly defeat Wisconsin in the Big Ten Championship game to claim its second consecutive Big Ten title.
Yeah, I know. I don’t like it either, but that is what the #math tells me.
If this scenario were to come to pass, Michigan would almost certainly make a return to the College Football Playoff. But, the selection of the team to represent the Big Ten in the Rose Bowl would be interesting.
No team from the West would have better than a 9-3 record, so the Selection Committee would need to sort out which team out of the East would get the nod. In this scenario, Ohio State and Michigan State would both be 10-2, and Penn State would be 9-3, but with wins over both the Buckeyes and the Spartans. In this scenario, Penn State would have a loss to Michigan in addition to a pair of early-season losses at Purdue and at Auburn. So, style points would likely play the role of tiebreaker.
Finally, while my disruptive simulation does predict a Big Ten title for the Wolverines, there is one variable that my simulation did not take into account. In the disruptive scenario, Michigan is able to beat a Michigan State team that eventually finishes with a record of 10-2.
As I have shown before, since the Spartans joined the Big Ten in 1950, the Wolverines essentially never beat Michigan State when MSU wins a large percentage of its other games (over 70 percent). In fact, Michigan has only defeated a “good” Michigan State team once in the past 70 years, and that was in 1955.
If we take this fact into account, I could certainly envision a scenario where Michigan State is able to beat Ohio State at home and upset Michigan on the road. In this case, the 11-1 Spartans would advance to Indianapolis where I would project them to be a narrow underdog in a rematch with Wisconsin. This scenario would look frightening like the 2011 season.
While I am not predicting this specific outcome, it is one worth noting.
With the Big Ten analysis now in the books, it is time to look at how the rest of the Power Five conferences might shake out. That is the subject of the next installment of this series. Stay tuned, and Go Green.
