Wait, there's an alien overlord for which we can vote? Please?
So it's been ten days (or in 2020 terms, six excruciating months) since the last time I ran the numbers. Let's see what the polls are showing us after having some time to digest Trump's COVID diagnosis, the way that he handled it, and the VP debate.
When last we looked, based on the +/- 7% threshold, there were 14 states in play.
Alaska (R+3, 3 EV, Trump) - unchanged
Arizona (D+5.7, 11 EV, Trump) -- now D+3, a change of -2.7
Arkansas (R+2, 6 EV, Trump) - unchanged
Florida (D+1.2, 29 EV, Trump) - now D+1.1, a change of -0.1
Georgia (R+1.3, 16 EV, Trump - now D+0.3, a change of -1.6 and a flip
Iowa (R+1.7, 6 EV, Trump) - now D+0.5, a change of -2.2 and a flip
Michigan (D+4.2, 16 EV, Trump) - now D+5.2, a change of 1.0
Nevada (D+4, 6 EV, Clinton) - now D+5.3, a change of 1.3
North Carolina (D+0.8, 15 EV, Trump) - now D+0.5, a change of -0.3
Ohio (D+2.4, 18 EV, Trump) - now D+3.3, a change of 0.9
Pennsylvania (D+4.3, 20 EV, Trump) - now D5.7, a change of 1.4
South Carolina (R+5, 9 EV, Trump) - now R+5.7, a change of 1.4**
Texas (R+3.5, 38 EV, Trump) - now R+3.2, a change of -0.3
Wisconsin (D+6.4, 10 EV, Trump) - now D+5.5, a change of -0.9
Here's the new batch of swing states, based on the current (10/12/20) RCP averages (footnotes below):
Alaska (R+3, 3 EV, Trump) - unchanged
Arizona (D+3, 11 EV, Trump) -- now D+2.7, a change of -0.3[1]
Arkansas (R+2, 6 EV, Trump) - unchanged
Florida (D+1.1, 29 EV, Trump) - now D+3.7, a change of 2.6[2]
Georgia (D+0.3, 16 EV, Trump - now R+0.4, a change of -0.7
and a flipIowa (D+0.5, 6 EV, Trump) - now D+1.2, a change of 0.7
Missouri (R+6, 10 EV, Trump) -- NEW CHALLENGER! [3]
Nevada (D+5.3, 6 EV, Clinton) - now D+6, a change of 0.7
North Carolina (D+0.5, 15 EV, Trump) - now D+1.4, a change of 0.9
Ohio (D+3.3, 18 EV, Trump) - now D+0.6, a change of -2.7 [4]
South Carolina (R+5, 9 EV, Trump) - unchanged
Texas (R+3.2, 38 EV, Trump) - now R+4.4, a change of 1.2 [5]
Wisconsin (D+5.5, 10 EV, Trump) - now D+6.1, a change of 0.6
Footnotes:
[1] -- Average seems to be skewed by one contrarian poll that has Trump up 4 in Arizona from The Trafalgar Group, a Republican polling firm. Trafalgar Group's CEO, Robert Cahaly, has been giving a number of interviews lately talking about what he calls "social desirability bias" and the "shy Trump voter". Basically, saying that because so many people are so vehemently opposed to Trump, a number of Trump supporters are too intimidated to admit it to a complete stranger over the phone. Which may hold some water -- not all Trump voters are the full-throated MAGAt crowd -- but this seems to me to be reminiscent of the "unskewing the polls" narratives that were so popular in the 2012 election, right before Mitt Romney got his ass handed to him. It's also a take on the "silent majority" narrative that every campaign that's behind in the polls invokes. So take this slight dip in AZ with a grain of salt.
[2] -- The widening gap in Florida seems to be less about Trump voters flipping to Biden as it is Trump voters just disappearing. Biden's support is staying level at about 48%, while Trump's average support has dropped from 47 down to around 44.5. These could be respondents who are just going to sit out the election, or vote third party, or are just back into the undecided camp. Would not be surprised to see third parties outperform in Florida (by which I mean 4-5% instead of 1-2%, and I would assume Jorgensen would be the primary beneficiary).
[3] -- Missouri has chosen to enter the swing state Octagon by dropping to an R+6. It was previously averaging Trump by 8, so this is wouldn't be a huge shift except that it's 2 points in ten days. Whether this is statistical noise or a genuine weakening of support will take another week and a few more polls to discover.
[4] -- As with [1] above, this precipitous dip in Ohio is mostly due to an outlier poll from The Trafalgar Group. Take with a grain of salt, although I suspect there has been a very tiny actual dip in support. Ohio will in all likelihood remain one of the closest races in the country.
[5] -- This one bothers me, because using RealClearPolitics' poll aggregates is the backbone of my methodology here, because it was (I assumed) a complete listing of significant polling without any formulas or massaging of the data. However, when comparing it to 538's info, I see that there are several recent (October 7-9) polls for Texas which RCP is not listing, and all of these are in Biden's favor, showing either a slight lead for Biden or a dead-on tie. For now, I'm going to assume that RCP is just lagging behind in getting these added to their site, but if they continue to omit them I'm going to have to assume they're doing some selective cherrypicking because their viewership and editorial content has shifted significantly to the right. In which case, I may have to find a whole new poll aggregator and rebaseline this whole thing.
[6] -- Astute observers may have noticed that Michigan and Pennsylvania are no longer listed. This is because they have met (MI D+7) or surpassed (PA D+7.1) the 7-point threshold and are now being counted in the "likely" Biden bucket. I considered retaining Michigan as it's right on the line, but if I did I would also need to include Kansas, which is now at R+7. For now, any state which is at exactly 7 points will be considered out of play.
So, after removing the likely/solid/no-chance-in-Hell states off the board, the tally is 263-98, with 38 states accounted for. At this point, Trump needs to win almost every state still in play, while Biden needs to win at most 3 of the 13 (Alaska, and two of Arkansas/Iowa/Nevada) and can win outright with a majority of the single states. If Wisconsin, sitting at D+6.1, were to move up past the 7% threshold, then it would be highly unlikely that Trump could win without the poll data being horribly inaccurate across the board and/or shenanigans. (As it is, it's already getting statistically highly unlikely Trump will win without some major shifts in his direction. Or shenanigans.)
My gut says of these 13, Trump will win Alaska, Arkansas, Georgia, Missouri, Ohio, South Carolina, and Texas. For a total of 100 electoral votes and a final tally of 340-198. A solid win, but not a total asskicking. However -- Ohio, Georgia and Texas are honestly all too close to call (as is North Carolina, to be fair). Even a subtle shift of half a point could be enough to put them in Biden's camp, there's still three weeks left, and the general trend has been a widening of Biden's lead rather than the narrowing one normally sees in the last stages of an election campaign. Those three flipping would make the final total 412-126, an asskicking almost on the level of Bush-Dukakis in 1988 (426-111).
Next time, I'll break down the Senate chances and key races.