I fail to see how in any scenario that has the Democrats retaining the White House in 2016, that they don't also take back Florida, Illinois, Pennsylvania, and Wisconsin once their voters who don't bother coming out for midterms show up. Alaska, Arizona (only if McCain doesn't run), North Carolina, and Ohio would also be possible Democratic gains in that case.
A Republican filibuster-proof majority in the Senate in 2018 requires at a minimum that the Republicans win the White House in 2016. Even then, it is unlikely.
The links between presidential elections and Senate races aren't always as clear.
It's possible that Democrats keep the White House while losing Florida. So that might help Rubio.
It's also entirely normal for candidates for Senate to run behind or ahead of the party's nominee for President. Incumbency is generally worth a few points. In 2012, Democratic incumbents won senate elections in several states Romney carried: Missouri, Montana, and West Virginia. Meanwhile, a Republican incumbent won in Nevada. Scott Brown lost by seven points in a state Romney lost by 23, so he still ran well ahead of the presidential ticket.
Also, unless the Republican president thwarts his party's extremist agenda he is never going to have 50/50 approvals. The Democrats will almost unanimously hate him, just as much Republicans hate Obama, and they will be motivated this time to go to the midterm polls.
Polarization works in two ways you know.
Look at how the guy took on the mission of pushing Bush's partial privatization of Social Security (as a Sentor representing Pennsylvania).
And look what happened to him two years later.
Republicans are better suited to opposing Presidents. A party that is philosophically in favor of reducing spending is tougher to negotiate with.