Iowa's special-teams edge is 0.8 points, and three games have shrunk it
SP+ puts Michigan's special teams 108th and Iowa's 7th. Michigan has kicked better, Iowa has returned better, and the rating separating them is still mostly a preseason projection.
Iowa's special teams are worth eight tenths of a point a game more than Michigan's.
That is the SP+ gap after three weeks. SP+ is Bill Connelly's opponent-adjusted efficiency rating at ESPN, and it measures every unit in points per game against an average FBS team on a neutral field. Iowa's special teams rate 0.5. Michigan's rate -0.3. Subtract and you get 0.8, which is less than one extra point.
The ranks attached to those two numbers are 7th and 108th. That is 101 places. It is the figure that travels, and it is the less useful of the two.
Why 101 places buys so little#
Put the three units side by side and the ranks and the points stop agreeing with each other.
Iowa's special-teams edge is the same size as its defensive edge
Iowa's SP+ advantage over Michigan by unit, in points per game on a neutral field, after Week 3 (table dated September 20, 2026). Labels give each team's national rank of 138. Positive favours Iowa on every row; on defence a lower SP+ rating is better. The three gaps sum to 2.1 against an overall gap of 2.2; the difference is rounding in the published ratings.
Read the bottom two bars together. The defences are two places apart and the special teams are 101 places apart, and the model prices both gaps at 0.8 points. The offences are four places apart and worth half a point.
The reason is the width of the scale. On offence, SP+ ratings run from 11.9 to 42.3 across the 138 FBS teams, a spread of 30.4 points. On defence the spread is 30.0. On special teams, every team in FBS sits between -0.6 and 0.6. Ole Miss is first at 0.6 and Oregon State is last at -0.6, and the best and worst units in the country are 1.2 points apart.
Pack 138 teams into 1.2 points and a tenth of a point is worth about ten places. Eleven teams print at Michigan's -0.3, and they occupy every rank from 100th to 110th. Thirteen teams print at Iowa's 0.5, from 4th to 16th.
So 0.8 is large against the special-teams scale, two-thirds of the distance from first to last. Against the game it is small. SP+ has Iowa 2.2 points better overall, and 1.3 of those points sit on offence and defence.
Michigan's counting numbers do not look 108th#
Here is what Michigan's specialists have done in three games, next to Iowa's.
| Through three games | Michigan | Iowa |
|---|---|---|
| Field goals | 2 of 2 (40, 49 yards) | 5 of 6 (miss from 50+) |
| Extra points | 10 of 10 | 12 of 14 |
| Punts, gross average | 14, 43.5 yards | 13, 44.1 yards |
| Punts downed inside the 20 | 8 | 9 |
| Opponents' punt returns | 4 for 17 yards | none |
| Own punt returns | 4 for 19, 4.8 a return | 11 for 187, 17.0 a return, 1 TD |
| Kickoffs returned by opponents | 2 for 39 | 5 for 77 |
Every number in that table is a count, and every count is small. The Michigan column comes from the club's cumulative statistics. That page labels Cam Brown's 43.5 as yards per game; it is 609 yards on 14 punts, so it is yards per punt. The Iowa column comes from the three game box scores.
The Michigan column is clean. The kicker, Trey Butkowski, has not missed. Brown's net average is 42.3 yards, and he has had no touchbacks. Opponents have returned four of his 14 punts and gained 17 yards doing it. Eleven of Jacob Baggett's 14 kickoffs have gone for touchbacks.
The Iowa column is uneven. Caden Buhr has missed two of 14 extra points and one of six field goals. On placekicking alone, Michigan has been the better team. Iowa's case is the return game: 187 yards on 11 punt returns, 77 of them on Evan James's touchdown against Northern Illinois, and no opponent has returned an Iowa punt at all.
The samples: n = 14 Michigan punts, n = 4 Michigan punt returns, n = 3 Michigan kickoff returns. Three kickoff returns is three plays. These figures complicate the 108th. They do not refute it, and "Michigan's special teams are fine" off three returns makes the same error from the other direction.
Where the 108th comes from#
There are four candidates.
Field goals. Not the cause. Michigan is 2 of 2 on field goals and 10 of 10 on extra points.
Opponent adjustment. SP+ adjusts every rating for the quality of the opponent. Two of Michigan's three opponents, Western Michigan and UTEP, rate 80th and 130th overall. A clean day against a weak coverage unit earns less credit than the raw count suggests. The public table does not break out that adjustment, so its size cannot be measured here.
Efficiency against counting. SP+ is an efficiency rating and the club's page is counting averages, and the two can disagree about the same kicks. A 43.5-yard punt fair caught at the 35 and a 43.5-yard punt downed at the 8 have the same average and very different value. An average cannot tell them apart. An efficiency measure can.
The preseason projection. ESPN prints this under the table: early in the season the ratings are based primarily on preseason projections, special teams included, and the preseason numbers are phased out week by week. The data say that the fourth candidate carries most of the weight.
In the final preseason table in August, Michigan's special teams rated -0.5 and ranked 113th. Iowa's rated 0.6 and ranked 10th. Across all 138 teams, the Week 3 special-teams ratings line up with the preseason ones almost perfectly (a correlation of 0.98). Sixty teams print exactly the rating they printed in August. Of the 138, 121 have moved by 0.1 or less. None has moved more than 0.3.
Michigan has moved up 0.2, one of ten teams to rise that far. Iowa has moved down 0.1. The gap has gone from 1.1 points to 0.8 in three weeks, and the evidence from those three weeks points toward Michigan.
The preseason -0.5 was set before Butkowski or Brown had kicked for Michigan. The 2025 kicker, Dominic Zvada, made 17 of 25 field goals and is no longer on the roster. The August number could not have seen this unit play.
So the 108th is mostly an August number, and the three games since have been pulling it up.
What would change the read#
If Michigan's rating stays near -0.3 as the preseason weight drains out, then the model sees something in the play-by-play that three games of clean counting numbers do not show, and the opponent adjustment and the efficiency scoring are doing real work against Michigan. If the rating keeps climbing, the 108th was a projection being overwritten by evidence. Either is possible with n = 14 punts.
On Saturday, watch Brown's punts against Iowa's return unit. Iowa averages 17.0 yards a punt return, and Michigan's coverage has allowed 17 yards on four returns all season. Those two numbers rest on 15 returns between them, and one long return would move either.
The call#
In the first SP+ table ESPN publishes after Michigan's regular season ends on November 28, Michigan's special-teams rating is 0.0 or better.
The case for it: the preseason projection appears to be what holds the rating down, it is being phased out, and three weeks of in-season data have moved Michigan up. A rating of 0.0 is an average FBS unit, and Michigan's kicking and punting numbers are clean.
The case against it: the rating has moved 0.2 in three weeks and needs another 0.3. No team in FBS has moved more than 0.3 in that time. Michigan's own punt returns average 4.8 yards on four returns, against Iowa's 17.0. The rest of the schedule includes Penn State, Indiana, Oregon and Ohio State. And two field-goal attempts is not a kicking record; one miss in a close game moves the unit.
Revisit on December 1.
SP+ ratings from ESPN's all-138-teams table dated September 20, 2026, and its final preseason table of August 13, 2026; the correlation and movement counts use the ratings as printed to one decimal, all 138 teams. Unit gaps are computed from the printed ratings, so they can differ from the model's unrounded values by a tenth. Michigan counting statistics from the club's cumulative statistics through September 21. Iowa counting statistics summed from the three game box scores. Iowa's third game was against Northern Iowa, an FCS team. Known weakness: SP+ publishes one special-teams number per team, so the split between field goals, punts, kickoffs and returns inside the rating cannot be seen, and no public figure isolates how much preseason weight remains in Week 3.
Sources: ESPN SP+ ratings tables (Bill Connelly), September 20 and August 13, 2026; Michigan Athletics 2026 cumulative football statistics; ESPN box scores of Iowa's games against Northern Illinois, Iowa State and Northern Iowa; Northern Illinois Athletics official box score, September 5, 2026.
