Why I won't be in suspense this presidential election night (in tweets)

Statistician Drew Linzer was pretty accurate predicting an Obama win in 2012 and Republican wins in 2014 Senate races. Here's his take on 2016:

I've analyzed 1,221 polls of the presidential election in all 50 states. They point to a Clinton victory. https://t.co/Xmp9oeVHo6

— Drew Linzer (@DrewLinzer) November 4, 2016

People who study this stuff know what data points to what outcomes. A Trump win would belie a wide range of theory and evidence. https://t.co/P3Mxp9qN0h

— Drew Linzer (@DrewLinzer) November 4, 2016

You need to understand Electoral College math:

Trump's challenge: If Clinton wins VA (leading +8), PA (+6), and NC (+3) then Trump would have to sweep FL, OH, IA, NV, AZ, CO, WI, and NH.

— Drew Linzer (@DrewLinzer) November 1, 2016

But in Nevada:

Final (almost) NV early #s are in:
Trump is dead.
GOP in big trouble in #nvsen, two House seats and #nvleg control.https://t.co/1pS7nvPBhm

— Jon Ralston (@RalstonReports) November 5, 2016

One more explainer of why the Electoral College map is so favorable for Clinton (and why national public-opinion polls late in a Presidential race are fairly useless):

Here's a roundup of some of the major data-informed forecasts:

Plus my friend Matt Kerbel's prediction (I created the map but not the forecast for his Wolves and Sheep blog):

And Linzer's model implemented in R code:

Nate Silver, so accurate in 2008 and 2012, was off the mark early in this year's GOP primaries, discounting Trump's chances of winning the GOP nomination. He's now the most bullish about Trump's general election chances among major independent forecasters (although he still says odds favor Clinton, just not by as much as other models). One analyst believes Silver's model may fall short as conditions change:

While another accuses him of mucking around a bit too much with raw poll data:

One of Silver's safe-for-work responses (not all of them were):

But not all serious political data analysts agree with that:

Moving on from model disagreements, some info on early-voting demographics:

Data show that this race has actually been quite uneventful in terms of likely winner, despite some media efforts to make it look otherwise. As Sam Wang told CNN: "There's drama, and then there's data."

The remarkably stable Presidential campaign, animated.
h/t David Elkhttps://t.co/SPS5BztC7g pic.twitter.com/3PguB5wYBe

— Sam Wang (@SamWangPhD) October 12, 2016

Daniel Drezner, professor at the Fletcher School, looks at what he thinks the Electoral College map would look like if Nevada early voting is signaling the way overall US Election Day voting will go:

Analysts I respect argue that Nevada is somewhat of a special case and you can't extrapolate national trends based on what's happening there. If positive early voting isn't at all predictive of Election Day, this is Drezner's scenario:

Good advice below: Don't assume that early returns on Election Night signal some sort of upset, since small, GOP-leaning states will likely be called first. Here's a timeline of what may happen when:

I'll let Dr. Wang have the last word here: