Finding NCAA Risers and Fallers Using Pythag Expectation

[dt_divider style=”thick” /]Not every loss is created equal and not every team benefits from the same amount of luck from year to year. A bounce of the ball here and a bounce of the ball there tend to even out in the end, but – during the course of a season – these little things add up, whether it be injuries, suspensions or, close games, and are often the difference between a great season and a lousy one.

Luck is a huge factor within any college football season, but when it’s over, the reset button is hit. Wins and losses aren’t the only things that start back at zero. So does luck. And often times, from season to season, an unlucky team in one season becomes a lucky team in the next.

A way to measure this luck is by using Pythagorean expectation – a formula that takes points scored and points allowed then converts them to an expected winning percentage. Those teams with a positive difference between actual winning percentage and expected winning percentage are believed to have been more “lucky” throughout the year while the inverse is true for those teams deemed “unlucky” by the formula.

There is significant evidence that utilizing the Pythagorean expectation equation to identify potential risers and fallers from year to year is effective. Since 2011, there have been 66 “lucky” teams with a difference between expected winning percentage and actual winner percentage of greater than 10%. Of those 67, 47 had a worse winning percentage the next season, 17 were better, and two stayed the same. There were 82 “unlucky” teams over the same time period. During the following season, 61 teams had a better winning percentage, 16 were worse, and five stayed the same.

Obviously, luck isn’t the only factor in all of this. Coaching and personnel changes often sway team performance the most. Each year the best teams create their own luck by simply being better than their opponents, and the teams at the top at the end of the season more often than not have benefited from a great amount of luck. But there is something to be said for a team getting the short end of the stick in close games one year, learning from it, and then benefiting in the next season.

Last season, there was evidence that those teams at the top of the Football Bowl Subdivision, in terms of rankings, were those that saw the most movement in their “luck” from year to year:

Pythag Expectation

There were 42 schools that garnered votes in last season’s preseason Associated Press Poll. Of those teams, 20 had a negative difference between winning percentages, signaling a potential decline in 2016. At the end of the season, 15 were worse, four stayed the same, and one (Miami) was better. There were only eight teams that could qualify as “lucky’ the previous season, again, because in order to be up there in the polls, you better be the beneficiary of some luck. Six of those teams were better and two stayed the same.

The biggest mover last season was the Washington Huskies. They had the largest difference between expected winning percentage and actual winning percentage (22.2%) in 2015, signaling a potential huge jump in 2016, which they accomplished making it all the way to the College Football Playoff.

Who were those teams garnering votes in the 2017 AP Preseason Poll that could be “risers” and “fallers” this season? Below are the five biggest potential movers along with some other important notes:

Pythag Expectation

1) Notre Dame – Preseason RK: 28, ‘16 W%: 33.3%, ‘16 Pythag. W%: 56.2%

For a team to go 4-8 and score more points than their opponent is a truly unique accomplishment. But for everything that went wrong last season with the Irish, including the unnecessary quarterback jostling, Notre Dame played far better than their record would indicate. They lose quarterback DeShone Kizer, but many believe Brandon Wimbush could be a better fit in Brian Kelly’s offense. Notre Dame returns almost every other offensive skill player of consequence and four starters along the offensive line, including two of the best in the nation (guard Quenten Nelson, tackle Mike McGlinchey) in addition to seven defensive starters and new defensive coordinator Mike Elko (from Wake Forest). Their schedule isn’t doing them any favors with Miami and Stanford on the road during the last month of the season, but they get Georgia, USC, and North Carolina State at home and the rest of the slate seems manageable enough to get above .500.

2) Auburn – Preseason RK: 12, ‘16 W%: 61.5%, ‘16 Pythag. W%: 80.6%

Count me among the Auburn believers this season and that is independent of the numbers presented here. Although not much has been seen yet of Jarrett Stidham actually playing football, it is anticipated that he will be a significant improvement over Sean White under center. You know their running game is going to be ferocious with Kamryn Pettway and Kerryon Johnson behind an experienced offensive line. Their crop of highly rated recruits at receiver may actually get some balls thrown their way, and seven starters return to their defensive unit – including safety Tray Matthews and linebacker Deshaun Davis – that ranked seventh in the nation in points allowed per game. Plus, the Iron Bowl is in their own backyard.

3) Michigan State – Preseason RK: 44, ‘16 W%: 25.0%, ‘16 Pythag W%: 41.6%

It was hard to believe that Mark Dantonio’s team stumbled to a 3-9 record last season after being bowl-eligible in each of his previous nine seasons with the Spartans. Most of the issues were pointed at the lack of consistent quarterback play after the departure of three-year starter Connor Cook, but the defense wasn’t as tough (or as good) as it had been in previous seasons either. There is less drama around who will be manning the huddle this offseason as dual-threat Brian Lewerke – recovered from a broken leg last October – appears to be the man, and running back L.J. Scott should provide plenty of support in the run game. They start 2017 with four straight games in East Lansing, and, despite having to travel to both Ann Arbor and Columbus, they avoid Nebraska and Wisconsin from the West.

4) South Florida – Preseason RK: 19, ‘16 W%: 84.6%, ‘16 Pythag W%: 68.4%

A trendy pick by some to be the Group of Five representative in the “New Year’s Six” bowls behind explosive quarterback Quinton Flowers, the Bulls could instead be in for a bit of a step back. Head Coach Willie Taggart departed for Oregon this offseason and while the much-maligned Charlie Strong appears to be a good fit in Tampa, he will have to repair a defense that gave up 31.6 points per game and almost 500 yards per game of total offense last season. Giving up 39 points in a bowl game to a South Carolina team that averaged less than 16 PPG in SEC play isn’t something to hang your hat on.

5) Michigan – Preseason RK: 11, ‘16 W%: 76.9%, ‘16 Pythag W%: 92.3%

If you watched the Wolverines at all last season, you wouldn’t be surprised at all by the numbers above. Harbaugh’s squad looked like one of the top four teams in the nation for most of last season. They finished the year at 10-3, but their three losses (all to bowl teams on the road or at neutral sites) were by a combined five points. They scored at least 40 points seven times, held opponents under 10 points five times, and were essentially three plays away from an undefeated season. This year, they have the unenviable task of replacing 10defensive starters, most of whom will be playing on Sundays this fall. But when you recruit like Jim Harbaugh does, you have guys like defensive linemen Maurice Hurst and Rashan Gary to take up the leadership mantle and spotlight on defense. Plus, budding offensive stars QB Wilton Speight and RB Chris Evans have yet to realize their potential, an improvement certainly could be possible. Their schedule is tougher this season with an opener against Florida and trips to Happy Valley and Madison, but they get Ohio State at home to end the year.

Other Notes

  • The only team in the top seven of the AP Poll with a positive difference pointing to potential improvement is Ohio State.
  • After Notre Dame, Auburn, Michigan State, and Michigan, rounding out the top 10 potential “risers” are: LSU, Miami, UCLA, Louisville, TCU, and Washington State (all at least +9.1%)
  • South Florida was the only “falling” team mentioned above, but rest of the top 10 “fallers” are: Nebraska, Stanford, West Virginia, Oklahoma, Georgia, Florida State, USC, Tulsa, Clemson, and Penn State (all at most -6.9%).
  • Two teams I’ve pegged as being a tad overhyped going into this season – USC and Penn State – both come in as teams expected to decline this year.
  • Two teams I’m higher on than most – Stanford and Georgia –  are actually expected to decline based on the information provided. Part of me can understand Georgia because they’re still an extremely young team, but Stanford is a tough one for me to understand unless Keller Chryst simply can’t find “it” this year.

Follow Jeff on Twitter @jfey5 and find his other work here, including a comparison of modern Super Bowl teamsthe best playoff QBs in recent memory, and what went into the Texans’ decision with their 2006 draft.

Want more Inside the Pylon? Subscribe to our podcasts, follow us on Twitter, like us on Facebook or subscribe to our Youtube channel.

Leave a Reply

Your email address will not be published. Required fields are marked *