Every once and awhile, voters deliver a result that none of the experts would have predicted.
In this case, experts were so wrong about the Cantor race for a number of reasons. Part of this was simply that there wasn't much polling done in the race, and the polling that did dribble out wasn't reliable - such as the now infamous report that a Cantor internal poll showed him with a 34-point lead. The Daily Caller released a "shock poll" last week showing a closer race, but it still had Cantor up by double-digits, with Brat at just 39 percent including leaners (he won with 56 percent of the vote).
Another reason is that it's hard to poll primaries, because pollsters can only make educated guesses about what the electorate will look like. In the 2012 Republican primary in the district, roughly 47,000 people voted, but Tuesday night, there were more than 65,000 votes cast - so it was a different pool of voters than what existed just two years ago.
In addition, the fact that Senate Minority Leader Mitch McConnell's easy victory in the Kentucky Republican primary came first cemented the assumption that Cantor would coast. Had Cantor's primary come before McConnell's, it would have attracted more attention and perhaps more outlets would have picked up on Brat's momentum. Alternatively, had McConnell lost or merely won narrowly, the possibility of an upset in Virginia would have been given a greater likelihood.
All this having been said, it was amazing to follow the news and see how the same experts who had no idea this upset was brewing - and who hadn't even followed the race closely - could turn around within seconds to confidently explain precisely why Cantor lost, and even how, to the extent that it doomed comprehensive immigration reform legislation, that it was bad news for Republicans heading into the 2016 election.
The results themselves should be a reminder that there are still big knowledge gaps when it comes to understanding voter behavior.