Evidence rarely speaks for itself.  People tend to interpret the evidence they see as consistent with what they expect.  National sportswriters expect Michigan to be bad this year.  With that in mind, the Utah game will go a long way in setting the tone of the coverage we get this season.

Suppose first that we lose to Utah.  When we beat Notre Dame, that just proves that they're still really bad, not that we're any good.  When we beat Wisconsin, the Wisconsin doubters step up to say the Badgers were one-dimensional and over-rated; everyone else says we have a strong D-line and we've always done pretty well against Wisconsin's running game:  the true tests lie ahead.  If Illinois has already lost to Penn State, then nobody much cares when we beat them:  they've clearly taken a step back from last year.  If Illinois has beaten Penn State then people will start asking if maybe we're better than people think.  If Illinois has beaten both Penn State and Missouri, then people will take notice of a Michigan win.  Finally, if we beat a Penn State team that has already lost to Illinois and Wisconsin, then the pundits will say that we have their number psychologically and they're probably not a good team anyway since it's half-way through the season and they've beaten nobody but cream-puffs.  A win over Penn State gets us favorable press only if they've beaten at least two out of Wisconsin, Purdue, and Illinois.