SUBSCRIBE NOW
As low as $49 for one year. Save 59%.
SUBSCRIBE NOW
As low as $49 for one year. Save 59%.

Feng: What are chances MSU or UM can knock off OSU?

Ed Feng
Special to The Detroit News
Jim Harbaugh has an impressive record of seemingly getting the most out of his players.

Which team will win the Big Ten East in 2015? Can Michigan and Michigan State challenge Ohio State, the nation's top team by just about every subjective and objective metric?

Most media predictions offer a single team (Ohio State) when making predictions. However, college football features too much randomness to make any prediction with certainty.

Instead, my calculations at The Power Rank assign a win probability to each team in the division.

Here are probabilities for Big Ten East teams winning the division.

These calculations start with preseason rankings that combine my preseason model at The Power Rank with subjective input from the preseason AP and coaches' polls. Over the past two seasons, these ensemble rankings have predicted the winner of 72.1 percent of games (1,080-417) before the start of the season.

These rankings imply a win probability for each of the 56 conference games between Big Ten teams. A computer simulation uses random numbers to turn these win probabilities into a winner for each game.

For example, Michigan has an 83 percent chance to beat Rutgers this season. The simulation flips a coin that comes up heads 83 percent of the time, recording a win for Michigan on each head and a loss for each tail.

After simulating each of 56 conference games, a division winner is determined based on record and tie-breakers. This simulation is repeated 50,000 times to determine the win probabilities in the visual.

Let's look at the primary contenders for the Big Ten East this season.

OHIO STATE

Preseason rank: No. 1. Division win probability: 74.6 percent.

Ohio State is the consensus No. 1 college football team in the nation. The Buckeyes won the first playoff last season and return so many starters that coach Urban Meyer still hasn't named his No. 1 QB.

With so many accolades, it's hard to find fault with the Buckeyes. But it is possible.

Let's look at Ohio State's performance over a three-year period before last year's Big Ten championship game. For each year, I show OSU's rank based on The Power Rank's team rankings, which take margin of victory and adjust for strength of schedule.

2012: 12-0, 14th.

2013: 12-2, 15th.

2014 before Big Ten championship game: 11-1, 13th.

In 2012, Ohio State went undefeated but had a 6-0 record in games decided by a touchdown or less. Randomness plays a big role in a team's record in close games.

In 2013, the Buckeyes looked like a contender for the BCS title game before losing to Michigan State in the Big Ten title game and Clemson in the Orange Bowl.

In 2014, Ohio State lost to Virginia Tech at home and needed overtime to beat Penn State. The defense had allowed 5.1 yards per carry (eighth out of 14 Big Ten teams), which seemed like a bad omen against Wisconsin in the Big Ten title game.

Then Ohio State played an unbelievable three-game stretch of football to beat Wisconsin, Alabama and Oregon and win the first college football playoff. Now, with their wealth of returning starters, it seems like the Buckeyes will never lose again.

Don't let a small sample size of three games fool you. Just like every football team, Ohio State has its good and bad moments over the years.

Mark Dantonio and Michigan State appear to be the most likely challenger to Ohio State.

MICHIGAN STATE

Preseason rank: No. 7. Division win probability: 15.7 percent.

In my Michigan State preview I noted how my preseason model had them underrated at 16th in the nation. In contrast, the AP and coaches' poll had the Spartans fifth and sixth, respectively.

With the inclusion of these human polls, Michigan State moves up to seventh, a more reasonable estimate. However, Michigan State still has a significantly smaller likelihood to win the Big Ten East than Ohio State.

MICHIGAN

Preseason rank: No. 34. Division win probability: 6.3 percent.

As I wrote in my Michigan preview, the preseason model doesn't consider coaching changes like Michigan's hiring of Jim Harbaugh. The pollsters also don't seem impressed with the change, as Michigan got one and two points in the AP and coaches' polls, respectively.

However, it's reasonable to believe that Harbaugh will turn the program in the right direction. It's not just that he turned a one-win Stanford team into a national powerhouse in four years. It's how he gets the most out of his players no matter where he coaches.

This starts with how Harbaugh understands people. A story about his wife in a recent Sports Illustrated profile explains this best.

From the story:

People often ask Sarah what it's like when Jim yells at her, and she tells them, "He's never yelled at me." She yells sometimes, but he doesn't yell back. He just tells her the real reason she is mad, and he is usually right, and yes, that can be maddening. Sometimes it would be easier if he just yelled.

Harbaugh seems to understand people like a Mike Krzyzewski or Phil Jackson. This should excite Michigan fans more than any football decisions the coach makes.

In addition, Michigan gets Michigan State and Ohio State at home this season, which contributes to their 6 percent chance to win the division.

PENN STATE

Preseason rank: No. 37. Division win probability: 3.2 percent.

To understand Penn State's chances this season, let's first go back to last season.

At The Power Rank, I rank offense and defense by taking yards per play, a powerful efficiency statistic, and adjusting for strength of schedule. Penn State ranked an amazing second on defense but a horrific 98th on offense.

The offensive line struggled last season, allowing sacks on 8.1 percent of pass attempts and not opening many holes for the running game. They return three starters this season and should see improvement similar to what Michigan saw last year on its line.

If the Nittany Lions can improve their line play, QB Christian Hackenberg, a highly-regard pro prospect, should lead this offense back to respectability. In addition, he has many top recruits at receiver.

If the offense can make it back to FBS average, Penn State might make this preseason rank of 37th seem silly because of its defense. Under coordinator Bob Shoop, this unit returns a wealth of starters that keeps them in the top 10.

Ed Feng has a Ph.D. in chemical engineering from Stanford and runs the sports analytics site The Power Rank. Email Ed Feng here.