[sargon]

If I remember correctly, though, these same organizations that were accurate in 2012 were very much off in 2014.

These outfits tend to be in denial and miss these wave elections. Same thing happened in 1994 (and 1980).

In elections such as 2008 and 2012, the "accuracy" of the mainstream pollsters can be attributed in large part to them getting the party splits and voter enthusiasm right.

In those other cycles that I referenced, in each case they missed the "wave", and I think that's exactly what's happening this cycle.

As for "disbelieving unfavorable polls", there aren't even that many of those left to disbelieve. Even with the obvious pro-Hillary bias, the polls have been trending in Trump's direction. And we also know historically that undecideds tend to break against the incumbent in such instances, and while Hillary isn't technically an incumbent, she's certainly a proxy for one, and those considerations still apply, IMHO.

Tournout decides elections, and turnout should favor Trump very heavily this year, unless something highly unusual happens between now and November 8. The evidence for that is ample and compelling.

The Revolution is ON!

Vote Trump!

[nathanbedford]

I too am looking for indications of a wave election and they are there to be found: The huge turnout for rallies, while not at all significant in determining who is ahead, does tell us something about intensity and it might tell us something about the latent power of the campaign; the encouraging signs of registration in many states increasing Republican count in decreasing Democrat count; the increased early voting ballots going to Republicans.

These things tell me not that we are going to have a wave election but that the tinder and gasoline are there waiting for a spark. That is probably a very good thing because wave elections seem to be limited to the midterm elections rather than presidential election, indeed we have not seen a wave election on the presidential level since Reagan. More, we haven't won a popular vote in five of the last six presidential elections and that, combined with our structural disadvantage in the electoral college, means that wave elections at this level are hard to come by. Yet, for other structural reasons, it may be that we need a wave election to win.

As you say, turnout is everything and it might be that the constituent elements of the Democratic Party simply are not aroused to turn out for midterms. That means the African-American vote. So I think it is true that Republicans are highly motivated and Democrats appear to be weakly motivated which explains why the race card is being played at the highest level by the candidate herself.

How this applies to polls is an art not a science and it will be difficult for pollsters to figure party participation at the polls when that is complicated by unknown levels of intensity. I'm inclined to agree with you, it is dangerous to over count the Democrats this time around but I will bet that the data mining system put it in place with revolutionary results by the Democrats in 2012 probably has a very good understanding of intensity. They will simply know precinct by precinct whether individual voters are motivated or not. The question is, do they know how to motivate them?

There are black swan events to be considered such as an economic crisis which occurred in 2008, a medical attack which exposes Hillary's true condition, a terrorist attack, or a revelation actually coming from wikileaks which amounts to a smoking gun. Any one of these, except an economic debacle and a terrorist strike, would be welcome as well as beneficial but that is hardly a way to plan a campaign.

I am eager to see what the legitimate polls tell us about debate performance which should begin to come in today and certainly by Friday.