The fact that Republicans haven't won a senate seat since Macaca-gate should tell Republican posters like the one who said "Democrats have only won Virginia twice" something...
Haven't there only been like 2 Senate elections since, both in Presidential election years? I remember Mark Warner and Obama winning in 2008 and Tim Kaine and Obama winning in 2012. And if memory serves the GOP ran joke, scandal-tainted candidates in both elections. I don't really see how throwing in the Senate helps you avoid the small sample size critique. I mean, between 2008 and 2012, the North Carolina GOP newly gained: control of the state legislature, a supermajority in the state legislature, the governor's mansion, a majority of Congressional seats, and the electoral college votes for President. Looking at that sample size I predict North Carolina will never ever ever ever never go Democrat again, because the 4 year window I chose to look at supports this theory.
That's a fair argument. Democrats won three Senate races in favorable circumstances.
There is the counterargument that Democrats also won two of the last three gubernatorial elections.
Although several Democratic wins (including a likely win in 2014) come down to Mark Warner being uniquely gifted.
You won't have perfect correlations between statewide office and presidential runs. Maine has a Republican Governor and a Republican Senator, but is unlikely to go to the party.
There are two arguments about where we are politically, and the Virginia results fit both.
I think it's a swing state in both senses of the term. It can go towards either party, and it'll go with the winner of the popular vote. Politics is a bit like a metronome, with parties doing well and then losing support. Currently we're at a period when the Democratic party has support, so looks at recent elections will be susceptible to sample bias, just as an analysis of 1980-1988 presidential results didn't foreshadow 1992.
If Virginia and the country were steadily becoming more Democratic, the last few elections might very well turn out to be a fairly representative sample.