Michigan Opens Up +10 at Wisconsin

Clutch may exist, but it's not predictive.  Someone who has been "clutch" in the past is neither more nor less likely to be "clutch" in the future than someone who has not been "clutch."  Clutch is only meaningful after the fact -- and is still often subject to confirmation bias -- which means that it's useful neither for determining point spreads nor for wagering against them.

It's similar to luck.  Luck certainly exists, but I hope that nobody would attempt to build a football team by selecting the luckiest players they could find instead of the most skilled.  Someone who has been lucky in the past may be unlucky in the future.

On the other hand, Las Vegas has quite a few very tall monuments dedicated to people who believe that they are lucky. :-)

I can believe than an individual player is clutch - able to perform well under pressure - but to say it about an entire team is pushing it.

You're right -- there's not enough sample size to know whether or not Michigan under Harbaugh is clutch.  But there is enough sample size to determine whether or not clutchness exists in general.  In other words -- if you look at the entire corpus, are there any "clutch" teams / players?  That is, teams or players who consistently perform better under whatever conditions you define as "clutch."  And the overwhelming evidence is that it does not; it's a product of confirmation bias more than anything else.  We tend to remember the successes or failures that fit the particular narrative.

Granted, much of the evidence comes from baseball, which has a much higher number of individual trials than football does.  However, the principles are the same.

People aren't robots, and it is certainly possible that a player's emotional state affects his play.  But I've yet to see a single study show that there's some way for a player to "psyche himself up" to play better -- and if he could, why wouldn't he do that all the time?

As for playcalling -- I don't even know how to respond to that.  Are you really suggesting that Harbaugh's staff somehow calls worse plays against better opponents?  Or, sequences them more poorly?  How would you differentiate that effect from better teams simply being more equipped to stop Michigan than worse opponents?

Was Penn State a big game last year?  Michigan won by 39, but everybody expected Penn State to be terrible.  How about Colorado?  Florida seemed like a big game at the time, but now Florida looks awful.  Was that a big game or not?

Finally -- not your comment, but the one above it -- Iowa has a good record at home against Top 5 teams recently because streaks happen.  Home field advantage is a real thing -- mostly due to the effect on officiating, but real nonetheless -- and Iowa normally has decent teams.  Suppose they were only 25% likely to win each of those 5 games.  The number of expected wins would be 1.25, and the probability of winning 4 or more would be about 1.5%.  Suppose there are 40 teams in college football that have the same 25% likelihood to win a home game against top 5 opponents.  It is more likely than not that one of those teams would go 4-1 or 5-0 over their last five.

Streaks exist, but they're not predictive.  Given those assumptions, the likelihood that Iowa wins their next home game against a Top 5 opponent is.. 25%.  Streak stats also tend to be cherry-picked.  Why Top 5?  If you go with Top 10, they're 4-3 over the same time period.  If you stick with Top 5 but go back to the beginning of Ferentz's career in 1999, they're 6-5.  The numbers are chosen to fit a narrative.