Just about a week until the election. The polls are getting... strange...
Sketchy Republican Polls
I'm not one to discount a right leaning poll just because it's backed by right leaning people, but we're seeing results that are incredibly suspect.
Rasmussen has Trump leading in the national polls by 1 point. This is a four point swing from their last poll, conducted about a week ago. That in itself isn't too strange considering the results are within the margin of error, but the eyebrow raising thing here is that according to the poll, Biden only has 77% of the democrat vote. Which... is strange. Trump by comparison has 84% of the republican vote. By comparison, Pew Research shows that 94% of Democrats voter for Clinton and 92% of Republicans for Trump.
The poll shows Biden leading by 7% among independents. Which again is strange. I believe Biden tends to do better among independents in most polls, but that's besides the point. More questionable is why Biden would be polling so well among independents while struggling among his own base. Rasmussen also claims that black approval of Donald Trump has just about doubled in the last week or so. That's a pretty drastic change in a short time, with nothing really to explain it. Biden didn't do great in defending the 94 crime bill at the debate, and 50 Cent endorsed him before denouncing him (50 cent btw is known to use his Twitter to get a rise out of people). Can these really explain such a rapid increase in Trump approval?
Meanwhile Trafalgar had some... out there polls, which we discussed before. Basically three polls bearing the Trafalgar name were released showing Trump up by 3 in NC, Michigan, and Arizona. Rather odd that all of them would have the same margin. More troubling was that the Crosstabs (demographic breakdown) was released for two of the polls. For Michigan, it showed Trump winning independents by about 30 points, Trump doing about as well with Hispanics as with Whites, and Biden only capturing 68% of the Democratic votes. Biden somehow did almost as well with Republicans (25%) as he did with independents. (28%) The head of Trafalgar spent the morning defending these polls and then later denied that they were his.
Lastly, Inside elections has Pennsylvania voting for Biden by 3 points. This is a 6 point swing from their last poll. Again, the polls show Biden doing relatively poorly among Democrats (80%) while crushing it among Independents (60%). They have Biden winning among seniors while losing among voters 18-44. Perhaps most surprisingly, they have Jo Jorgenson equaling Donald Trump with 26.3 of the Hispanic vote... O_o..?
So, those are obviously some very suspect numbers. There are 3 possible explanations for this.
a) These pollsters, who were right in key states in 2016, have done it again, and somehow picked up on trends everyone else has missed. The electorate has indeed changed massively since 2016.
b) These pollsters are intentionally going against the grain. If the results come out as expected and Biden wins Michigan, nobody will remember or care, but if Trump wins they seem like geniuses.
c) They are intentionally producing bullshit for those who hire them to promote a narrative.
My money is on a combination of B and C.
Other Polls
In regards to other polls, the news is mainly the same old thing. Biden has dropped slightly from his peak, but is still averaging a 9.4% lead on fivethirtyeight.
In Texas Biden briefly held a lead, but a couple of recent polls have shifted things back. Still, they show a very narrow race in the state. NY Times has Trump up by 4, and Yougov by 5. Trump is the clear favorite, but Biden has a larger chance of winning Texas than Trump does at Pennsylvania.
Speaking of which, IPSOS has Biden up by 5 in Pennsylvania. Considering how likely this state is to determine the election, that's a bit too close for comfort. Yougov and Muhlenberg college have Biden at a more comfortable 7 point lead. Trump really can't win without PA, because Michigan, Wisconsin, and Minnesota seem to be further to the left than PA. If those 4 states vote for Biden, Trump has to either flip New Hampshire and Nevada or flip Virginia, in addition to winning every swing state. That doesn't seem feasible.
Biden continues to do well in North Carolina and the last 3 polls of Georgia show a dead heat. If Biden wins either state, Trump would need to counter by winning Pennsylvania, Arizona, and Wisconsin, in addition to every other swing state. If Biden won both, then Trump would need all of those states, plus Michigan.
Last note for those still unable to let go of 2016 (myself included), Biden is now running ahead of where Clinton was at this point in Battleground states. He's up 4.0%, and she was up 3.5%. So, a larger error would be needed for a repeat.
Winning Scenarios for Biden
With all that in mind, what are the most likely win scenarios for each candidate? We will assume Michigan, Minnesota, Wisconsin, Pennsylvania, Texas, Georgia, Iowa, Arizona, North Carolina, Iowa, Nevada, and New Hampshire are all in play.
Biden's best path is keeping the states Clinton won (Nevada, New Hampshire, and Minnesota are the only ones that are questionable) and winning back PA, MI, and WI. Considering Clinton lost those by less than 1% each, that seems very likely. Of those states, PA is most at risk. If Biden loses there, Arizona or North Carolina could make up the difference.
If Biden keeps the Clinton states and wins Florida, America's penis, then he's at 252. Either Pennsylvania, or any combination of two states out of WI, MI, and MN would take him over the top.
If Trump keeps the three rust belt states he won, Biden has a tough path. He would have to win Florida and either Arizona or North Carolina. Without Florida, Biden would need Georgia, but that seems less likely.
Lastly, Biden could score a TKO in Texas. This is one of the less likely outcomes, but Texas is different enough that he may win it without doing well in the rust belt or Florida. If Biden wins in Texas, he'd need one more state to push him over the top. The likeliest candidate would be Arizona which is somewhat similar in demographics.
Of course these scenarios are not mutually exclusive, and their could be some mix and matching. For example, it's feasible that Biden loses PA and Wisconsin but wins in Michigan. In which case Arizona and North Carolina could carry him to victory. Michigan, and Florida are also a winning combination. So, if the rust belt winds up as close as in 2016, and the states wind up getting split, then there are other options for Biden.
Winning Combinations for Trump
Trump has a narrow path. Which actually makes this section a bit easier.
Texas, Ohio, and Florida are essentially must wins. Unless Ohio is strangely more left leaning than PA and WI this time around, losing Ohio is losing the election. Obviously losing Texas ends the election for all intents and purposes.
Losing Florida still leaves Trump with a path to victory, albeit a very narrow one. He would have to win the three rust belt states he carried in 2016, and win every other swing state. Possible, but the odds of Trump winning without Florida are less than 1%.
Basically, Trump's path to victory is basically every swing state outside the rustbelt, Ohio, Iowa, and one of the 3 states that were within 1% in 2016.
How likely is this to happen? Well, according to fivethirtyeight there's about a 13% chance. The polls would have to be off in the rust belt to a similar but slightly greater degree than in 2016. The polls would also have to be off in Arizona (where Trump was overrated by half a percent) by 2-3%, Florida by 1.8% (Trump was underrated by about .8%), and in North Carolina by 1.3% (Trump was underrated by about 2%), and in Iowa by about 1%.
None of those things are particularly unlikely in themselves. Aside from the rust belt states, all of those results would be within the margin of error. But, all of those happening, plus Biden not outperforming polls in Texas, Georgia, Iowa, or Ohio, are not all that likely.
Welp. Just 8 days until we hopefully find out.