You're being ridiculously hackish to equate MI and MS. Like, okay, some polls in 2008 overestimated Republicans in Michigan, so that just blows out all the other evidence we have? That nearly every election in the past 20+ years supports the idea that MS is substantially more Republican than MI is Democratic? That polls in MI indicate a close race but polls in MS don't? That even in the same year where MI's polls were supposedly crappy, MS polls also underestimated Republican strength? That all of the prognosticators strongly disagree with you?
Democrats winning strongly in MI Senate seats in recent years is totally consistent with it being more elastic. The past few Senate races have been in Democratic years with strong Democrats and/or weak Republicans, so it's not surprising that Democrats would win big then. In Republican years like 2010, Republicans won big (there wasn't a Senate seat up then, but there's no reason to believe it couldn't have been competitive when Republicans swept the rest of the Midwest that year).