They stepped in it big time. I'm ignoring them. Vote!

in place that is completely different from ANY OTHER ELECTION. The pollsters typical election polling models are pretty much blown up as so many interactive variables now exist that they have zero way of measuring the impacts in a combination of the factors.

My view from having done statistical analysis on and off over the years is no pollster has a model that is going to have a "good" chance" in prediction.

this is what we do know that is different in 2018:
- 2 years of special elections which saw nothing BUT democratic building enthusiasm AND voting
- the first time in decades the democratic base is thoroughly angered = motivation = participation = contributions = VOTES
- a president and a GOP party hellbent on destruction of america for profit of the .01% and policies they run away from
- demonization of women, of blacks, of hispanics/latinos, continuing non-stop
- OVERT in your face racism and bigotry echoed thruout the campaign by GOP candidates and supporters.


how do you measure any of this as it all resides outside the normal factors in the models themselves? All I know is what we see as indicators such as early voting and in specific areas if the data allows breakout of demographics of the early voter. We do know that higher rates of women have voted, higher rates of minorities have voted and in some instances and locations, higher rated of younger voters (under 35). Anecdotally, we can factor these votes will be more democratic than republican. Independent votes (non-party affiliated), if we take a conservative approach to say those votes are evenly split. We see a significant unpredictable advantage with democrats, based on past computer models. Its fair to say, that pollsters can easily understate the democratic turnout and percentages by 3-5% and that built-in error can easily win 50 seats or more in the House and actually allow a Senate majority as 4 or 5 senate races will be decided by 5% or less.

This is definitely a weird year. The evidence that more people are voting, especially women, minorities and (maybe) younger people, is probably a positive sign for Dems but beyond that it seems difficult to predict anything.

looking at this as the margin is being understated which is good for all of us. Nov 6th will boil down to ground game and GOTV in key states and districts


I want to watch the election returns and hear all the time "well, we didn't see this coming"....

Maybe all weekend?

They are all but saying jobs numbers today are Republicans October surprise.

Eck

investors care, but average americans care about what cost them...

But for some reason, our media outlets make an editorial decision not to share quite all that they know. I'll give you a f'rinstance:

Oregon is having a gubernatorial election, Democratic incumbent Kate Brown pitted against Republican and former legislator Knute Buehler. Oregon isn't heavily polled, but there have been occasional snapshots in the last few months. Brown is leading Buehler in each poll by 3-5%, but her lead is consistently characterized as "within the margin of error." No dispute about that, but by not talking about the polling history, it seems that the race is far closer or more volatile than perhaps it is. Brown will not cruise to a landslide victory, but she will probably win.

Why do I think that? Because of another little datapoint, early returns of ballots. Oregon has vote-by-mail, and ballots have been sent in and dropped off at county registrar offices for the last couple of weeks. Overall turnout as of a couple of days ago was 21.6% of all ballots had been returned. Of all ballots sent to Democratic voters, 27.4% of ballots had been returned, as against 27% of Republican ballots. Oooh close, right? Except in Oregon, Democrats voters outnumber Republicans, so even though both parties have a similar percentage of ballots turned in, the actual numbers would be in favor of Democrats.

Here's why: Let's assume there are 125,000 registered Democratic voters and 100,000 registered Republicans. If both parties have returned 27% of their ballots, that means that about 6,000 to 7,000 more ballots have been returned by Democrats. Obviously, not every Democrat will vote for the Democratic candidate, but neither will every Republican voter cast a ballot for the Republican. By reporting just the percentages but not the numbers behind those percentages (which the media are fully aware of), viewers and readers are left with a distorted picture of how things are progressing.

Since nobody's supposed to opine about what will happen, not much else to talk about...