It's worthwhile to pause my incessant checking of the NYTimes polling site and the fivethirtyeight assessment of probable and possible outcomes and consider the unthinkable:
Maybe, just maybe, the polls are off and Trump's packed rallies represent something more than the enthusiasm of a set of niche voters. It seems that pollsters and analysts may have reassessed their performance in 2016 and have changed their methods and approach in 2018. I hope so. But it's also possible the pull of a conventional wisdom is still working.
We should know in five days time, although the worse thing I've seen today is the idea it will take weeks to find out who controls the House. (The reason: a lot of close races and the long time it takes to count mail ballots, particularly in CA.)