Just a quick one, but something that really needs to be mentioned about statistics.
Every election cycle, we see all kinds of spurious "predictors" of the race. Here are a few recent examples:
The loser of the first debate wins the election.
The candidate who leads in the polls on Labor Day wins the election.
Sitting Senators rarely win presidential races.
The taller candidate usually wins.
And so on, and so on. The trouble with all of these is they are not predictive. Given the thousands of factors that are evaluated (obsessively) in every presidential race, it is a given that some patterns will emerge. However, unless the pattern has some causative relationship to the race, it's meaningless. Otherwise, the pattern is just as likely to be broken as upheld in the next race. Any number of patterns may be apparent in the rear view mirror, but have no predictive value whatsoever.
This kind of tea-leaf reading is totally irrelevant to the race. Don't let it get you down, and don't let it get your hopes up. The only thing that matters is what the American people think of each candidate on Nov. 2.
Keep on fighting!