It's Monday (and Cakeday), so let's pause the neverending leftist navel-gaze of "Are we the baddies?" (I can assure you, the Right does not bother to ask this question. Ever.) and crunch some numbers.
Let me start by saying I have just flat-out given up on RCP, for at least four reasons:
1. They have no consistency to their thresholds for tossup/lean/likely/solid (for example, Virginia at D+13 is "leans Democratic" while Arkansas and Alaska at R+6 are "likely Republican". Indiana is "Lean Republican" at R+7, while Kansas is "Likely Republican" at the same R+7)
2. Cherrypicking which polls to include and which to leave out, and showing a preference for Trump-friendly pollsters like Rasmussen and Trafalgar Group.
3. Multiple states for which the most recent poll is this past spring, or for which they don't even have any polling data.
4. Their burning dumpster fire of an op-ed page. It's tough to believe a site is unbiased when they have headlines like "No, Really -- Don't Pay Attention to Early Voting Data", "Lewandowski: When Donald Trump Wins Florida Big It's Going To Set The Path For Rest Of The Night" (RCP's own poll average for Florida has Biden up 1.5%), and "A Trump Victory Will Be a Giant Middle Finger to the Cultural Left".
So for this week and next week (the final number crunch before Election Night), I'm using 538's polling averages, which I've been tracking concurrent with RCP since 10/12. The overall movement this week has been half a point to a full point in favor of Trump in many places. This isn't unsurprising, given that Biden has been riding the largest polling advantage for a challenger since scientific polling of US election began in 1936. Most of the largest decrease has been in the deepest blue states, which also makes sense and is of the least concern. Dropping 2.2 points in Illiinois or 2.8 points in Delaware isn't going to keep anyone up at night on the Biden team.
It's tempting to say this may be reversion to the mean, but there's not a corresponding dip for Trump in most of his safe states. He's generally gained about the same margin (0.5-2.0) in the deepest red states, with one interesting exception: The Mountain West.
He's lost 0.5 in Idaho, 0.7 in Utah, 1.4 in both of the Dakotas, 1.5 in Montana, and Biden has extended his lead by the merest fraction of 0.1% in Colorado. The outlier is Wyoming where Trump has gained a point (for perspective, 1% of Wyoming is roughly 58,000 people, or a modest suburb in most other states).
So on the whole, it's been a decent week for Trump. Maybe people were impressed that he just seemed like a crotchety old man at the second debate and less like a roid rager with ADHD. Maybe Laptopgate is actually peeling off a few folks here and there. Maybe people are cringing at the thought of four years of watching Jim Carrey play the President in SNL skits (I know I am.)
But, it's almost certainly too little too late. 61.27 million people (slightly more than a quarter of the estimated electorate, and 44% of the 2016 voting electorate) have already voted. In Texas, over 80% of the *total* 2016 electorate has voted prior to Election Day. Georgia, Florida and North Carolina (all crucial battlegrounds) have all exceeded 50% of their 2016 totals, and there's still a few more days of early voting to go.
And, even if we assumed a rate of Trump gaining by a point a week, he's only got 8 days left. Only one state is currently within less than a point by polling averages, and that's Georgia (which, as I'll get to, I'm putting in Trump's bucket anyways).
So let's refresh where things stand. Let's make +10 or greater the "safe" marker.
Biden gets:
California (55)
Colorado (9)
Connecticut (7)
Delaware (3)
District of Columbia (3)
Hawaii (4)
Illinois (20)
Maine + Maine CD1 (3)
Maryland (10)
Massachusetts (11)
New Hampshire (4)
New Jersey (14)
New Mexico (5)
New York (29)
Oregon (7)
Rhode Island (4)
Vermont (3)
Virginia (13)
Washington (12)
TOTAL: 216
Trump gets:
Alabama (9)
Arkansas (6)
Idaho (4)
Indiana (11)
Kentucky (
Louisiana (
Mississippi (6)
North Dakota (3)
Oklahoma (7)
South Dakota (3)
Tennessee (11)
Utah (6)
West Virginia (5)
Wyoming (3)
TOTAL: 90
And this is why being the party of rural conservatism is ultimately a losing game in the long-run. Winning the states where nobody lives isn't a great formula, even in a system designed to give the empty states an advantage.
Now let's include the states in the +6-to-+10 range. I'm lowering it from +7 to +6 in this case because +6 is well outside MOE of any reasonable poll, and there's not enough time for a meaningful shift to bring it into MOE. Anyone accusing me of being a "homer" will see that the shift from 7 to 6 benefits Trump more than Biden in the proejction.
Biden gets:
Michigan (+8, 16 EVs)
Minnesota (+7.9, 10 EVs)
Nevada (+6.2, 6 EVs)
Wisconsin (+6.7, 10 EVs)
New Total: 258
Trump gets:
Alaska (+6.2, 3 EVs)
Kansas (+9.4, 6 EVs)
Missouri (+6.6, 10 EVs)
Montana (+6.7, 3 EVs)
Nebraska + CDs 2 & 3 (+6.5, 4 EVs)
South Carolina (+7.6, 9 EVs)
New Total: 125
Of the remaining +/-6% battlegrounds, Biden needs 12 EVs to win, Trump needs 145.
Battlegrounds:
Arizona (D+2.9, 11 EVs)
Florida (D+2.3, 29 EVs)
Georgia (D-0.4, 16 EVs)
Iowa (D+1.2, 6 EVs)
Maine CD2 (D+0.3, 1 EV)
Nebraska CD1 (no polling data, but general consensus is a slight Biden lead, 1 EV)
North Carolina (D+2.4, 15 EVs)*
Ohio (R+1.6, 18 EVs)
Pennsylvania (D+5.6, 20 EVs)
Texas (R+0.4, 38 EVs)
*Should be noted that the North Carolina average is being skewed by heavily contrarian polls from Rasmussen and Trafalgar Group which show a Trump lead, while several other recent polls all show Biden with a 3-5 point lead in the Tarheel State.
Again, Biden needs 12 to win. Over half of the remaining states (not counting CDs) have more than 12 EVs. To put it another way, Trump absolutely cannot win without Florida *and* Georgia *and* North Carolina *and* Ohio *and* Pennsylvania *and* Texas *and* at least one of either (Arizona) or (Iowa plus both CDs). That's a probabilistically tall order, especially with Pennsylvania sitting at a comfortable 5.6 point advantage. (For anyone familiar with 538's tipping-point "snake" graph, this is also why PA has such a high chance to be the tipping point state.)
Put yet another way, if I were to drop the battleground threshold another half-point, to 5.5, Biden has already won and the game is over. At that point it's just about how much can he run up the score to prevent Trump from flipping the table and accusing Democrats, Antifa, BLM, the media, China, the Deep State, Reptiloids, Antarans, and the underpants gnomes from all colluding to steal the election.
FWIW, I think Trump wins Ohio, Georgia, Iowa and Texas. GA and TX have Republican governors who are no friends to voter enfranchisement, Ohio has a lot more "bleak Midwestern farmland" to it than people realize, and Iowa is almost entirely bleak Midwestern farmland. I had hoped that the FirstEnergy scandal might swing Ohio blue, but it appears any push in that direction has faded and receded back to start.
I think Biden wins Pennsylvania, North Carolina, Arizona, Florida and the two CDs. Which would give us a final tally of 335-203. Frankly, I'd love to see Trump fail to crack 200, just because I think psychologically there's a significant difference between 199 and 200. Maybe Iowa hold true to the polls.