College Football: Why the Metrics Are Down on Iowa and Michigan State
If you trust the most popular advanced statistics, this year’s College Football Playoff is setting up as a very easy game of One of These Things Is Not Like the Other.
With the Big 12 title already in hand, Oklahoma is essentially the first team to secure a spot in the Final Four. Barring major upsets this weekend, the Sooners will be joined by SEC champion Alabama (currently a 17.5-point favorite over Florida) and ACC titlist Clemson (5-point favorite against North Carolina).
All three teams have separated themselves on the field and the stat sheet. Consider Alabama, Clemson, and Oklahoma all rank nationally top 4 in numberFire’s team efficiency ranking, known as nERD; top 10 in average margin of victory; top 10 in yards per play allowed; and top 20 in defensive S&P+ and in points allowed on a per game basis.
These teams represent a stark contrast to either of the two schools likely to join them in the playoff, Iowa and Michigan State. The Hawkeyes and Spartans will do battle on Saturday in the Big 10 Championship, a de facto play-in game. There will be no complex formulas or series of what-ifs.
But an examination of the statistical profiles of both teams suggests that they are at least a tier below the college football elite. Here’s why any talk of Iowa or Michigan State as national champion needs to be taken not with a grain of salt -- but with an entire salt shaker.
Using 4 metrics-driven rating systems (nERD, Fremeau Efficiency Index, S&P+, and ESPN’s Football Power Ranking), Iowa is currently ranked only as high as 18th and as low as 29th. Yet, the College Football Playoff selection committee has seeded the Hawkeyes fourth two weeks in a row.
So why the Grand Canyon-sized chasm between what the metrics say and the committee’s assessment?
First, the committee is rightly responding to Iowa’s unblemished 12-0 mark. But SB Nation’s Bill Connelly has surmised that the Hawkeyes “have mastered the art of close wins.” Put another way, the metrics’ disrespect for Iowa comes down to the Hawkeyes being more fortunate than impressive.
To be sure, the Hawkeyes have collected their fair share of narrow victories. Their 12 consecutive wins include 6 contests where they prevailed by single digits, all of which came against teams power ranked no higher than 29th by nERD. Further, Iowa needed a last-second field goal to beat Pitt, was tied with Iowa State in the fourth quarter, and led Minnesota by only one point with five minutes remaining.
A deeper dive shows that Iowa could and possibly should have lost at least two games. Using Connelly’s win expectancy metric -- which measures how frequently a team would have won a specific game given that game's primary statistics -- Iowa was only 25 percent likely to beat Wisconsin (the Hawkeyes won 10-6) and 23 percent likely to defeat Minnesota (the Hawkeyes won by 5).
Iowa’s signature showing was a 40-10 dismantling of Northwestern on the road. But Northwestern, itself, is considered drastically overrated by a series of advanced statistics. Despite the College Football Playoff selection committee’s gaudy ranking of 14th for the Wildcats, Northwestern is rated outside the top 50 by nERD, S&P+, and FPI, and 39th by FEI.
Without a series of wins against even above-average opponents, it’s no surprise that Iowa’s strength of schedule sinks the Hawkeyes’ resumé like an egregious typo. Using Prediction Machine’s strength of schedule ranking -- which considers as factors margin of victory and wins and losses of opponents and the opponents of a team's opponents -- Iowa’s schedule rates just 62nd in the FBS. The S&P+ rankings are even less generous: the Hawkeyes’ strength of schedule is 104th, the worst of any Power 5 school except for Duke and potential playoff-crasher North Carolina.
Finally, Iowa doesn’t do any one thing on either side of the ball particularly well. According to numberFire’s efficiency ratings, the Hawkeyes’ run defense is ranked 17th in the FBS, but their rushing offense, passing offense, and passing defense ranks 31st, 29th, and 28th, respectively.
Iowa’s efficiency ratings seem to epitomize the team’s entire 2015 campaign: good, not great.
Michigan State’s problem is certainly not resumé. The Spartans can boast wins at Michigan and Ohio State and a non-conference victory over surging Oregon. The Wolverines and Buckeyes are inside nERD’s top 20.
As such, Michigan State is ranked higher than Iowa by each nERD, FEI, S&P+, and FPI. FEI ranks the Spartans 6th, a clear outlier, while the other three systems rate Michigan State an average of 14th.
But the metrics show that, like Iowa, it’s difficult to find phases of the game where Michigan State truly excels, and the Spartans simply are not efficient. Sure, the Spartans’ rush defense is rated 12th in efficiency, but their average yards per play allowed (5.6) is good for just 51st in the country. Michigan State’s offense is solid but unspectacular, with a passing attack ranked 24th by numberFire and a rushing offense a dismal 74th.
Michigan State has also been the recipient of some turnover luck, ranking seventh in the nation with an average turnover margin of plus-1.2. Bill Connelly’s Turnover Luck metric shows that the Spartans’ actual turnover margin (plus-14) ranks Michigan State third in the nation while its expected margin (plus-4.08) is just 32nd. Connelly calculates that the difference between actual and expected turnovers results in an additional 4.13 points per game for the Spartans.
Michigan State’s struggles against inferior opponents also hurt the Spartans. Despite its fluky nature, their loss to Nebraska is as perplexing as any result this side of the Red River Rivalry. Big 10 bottom feeder Purdue only lost to Michigan State by 3 points and Rutgers fell on a Michigan State touchdown with 43 seconds left.
There is little doubt coach Mark Dantonio has gotten this Michigan State team to exceed even the loftiest of expectations, especially in the face of a rash of injuries and after losing Pat Narduzzi, arguably the nation’s best defensive coordinator.
But Michigan State, if they can get past Iowa on Saturday, will likely be thrust back into the all-familiar role of little brother.
Vegas gives the slight edge to Michigan State in the Big 10 Championship. The Spartans are currently favored by 3.5 points over Iowa. But much like the metrics, bookmakers are far from bullish on both squads. According to the Westgate Las Vegas, Michigan State’s odds to win it all are 8-to-1 and Iowa’s are 20-to-1.
Alabama, Clemson, and Oklahoma appear too talented, efficient, balanced, and dominant. Credit to Michigan State and Iowa for their unanticipated runs to the Big 10 Championship Game, but the metrics suggest that the conference champion’s road to the national title will probably begin and end in Indianapolis.