Some analysis on early voting and polling that may be missed
Polling, although I am confident of it's accuracy, is one thing and early voting, another.
Let me explain my opinion. Early voting data just reveals the political party the voter is registered in. Still, anywhere from a quarter to a third of voters are un-enrolled. Every poll has this voting block breaking solidly Democratic.
In addition, even though Republicans have an 85% to 90% approval of Trump, and say they are voting for the Republican for Senate or the House, polls show that there are fewer and fewer Republicans self describing as Republicans. I surmise that when polled, these individuals are ashamed to say they are enrolled as Republicans and tell the surveyor that he or she is un-enrolled, when in fact, they are still on the books as a Republican. After all, who goes down to city hall to just change their registration, unless they vote in a primary? I surmise that a lot of enrolled Republicans, voting early, in fact consider themselves to be un-enrolled and are voting Democratic. I think the pundits are assuming that a big Republican turnout reflects the 90% in the polls. That will result in the biggest surprise on election night. "Where did all these Democratic votes come from?" They came from registered Republicans that consider themselves Independents and voted Democratic.