President

The Democrats have nominated Joe Biden, who at the outset seemed to be on the stronger end of a candidate for Trump to face. However, his clear cognitive decline and strange statements have made him a weak candidate, possibly even weaker than Sanders. His running mate, Kamala Harris, is a mixed bag at best, and at worst a liability for Biden. We also have seen more lid days than any candidate in modern times from Biden, which is usually a sign that the candidate is losing with grace.

The three questions of this election I have are:

- Will there be a significant "buyer's remorse" effect?

- What is the final partisan breakdown of VBM/IPEV/Election Day voting

- Is there going to be a larger enthusiasm differential than 2016 (1.1%)

Remember how I said on my 2020 page that these three elections are a Jenga tower? This one is particularly so. I say we have approximately a 7 in 9 chance (about 78%) that we know the winner by 10pm ET on November 4th. However, certain scenarios could lead us to infer the winner of the election before we even get to November 4th. This election may be close, but it won't be indecisive (ahem, 2000). POST NOTE Nov 30 2020: BOY WAS I WRONG!!!

Here, I will compare the national RCP aggregate figures from now to today 4 years ago. I’m only using two-party figures, so if the aggregate is “49.0 to 40.5” these two numbers are added together, and the lead is taken from the proportion of this total.

The live polls only model is now active.

Final Ratings

Trump has a slight edge (56.7%) to win the election

Complete Ratings:

Analysis

My biggest four errors in my prediction are: GA, NE-2, AZ, and MI (I had bigger errors but they were in states I got right)

Nebraska's 2nd District was down to Suburban bleed-in from Omaha. While the national environment shift would have predicted this 1 vote to be won by Biden, it would still be off by 6%.

Out of the three states Trump won in 2016, Michigan was the only one I predicted would swing right, but only by 0.5%. It did not. What appears to have happened is that while Trump did improve in the very rural parts of the state, as well as very urban and heavily minority areas, he slipped in the suburbs of the cities. My guess is that he slipped slightly more in these areas than he gained back.

Arizona was very marginal this year across the board. Most areas here swung according to the national shift, and that already makes the state narrow (within 0.5%), but not quite enough to flip. Breaking the result down by congressional district shows that the 2nd and 6th were the only two to trend notably towards the Democrats, and based on the raw vote totals, this was just enough to explain for the 11k vote win.

Georgia has no clean explanation. Applying a national environment shift to GA brings it to R+1.8%, or with the high turnout scenario we saw, a Trump win by 89k votes, a 101k vote error. Breaking the results down by congressional district, the 2nd, 5th, and 8th had notable trends towards trump, and the 4th, 6th, and 7th had notable trends towards Biden.

Compared to national shift:

In the 6th District, Biden netted 16.2k votes. He also netted a similar amount in the 7th District (17.5k votes). He netted a smaller, 12.4k votes, from the 4th district.

In the 2nd District, Trump netted 3.5k votes. The 5th District I do not have raw vote data yet, but it will likely be in the same region of 3.5-5k votes. The 8th District was also a small netting of 1.2k votes.

Combining these: 89k/2 - 16.2k - 17.5k - 12.4k + 3.5k + (0 for sake of argument) + 1.2k = 3.2k (double for margin) = 6.4k vote win for Trump.

Yes, 6.4k votes is close, but three states are decided by ~12-15k votes each, and it is still hard to overturn in a recount.

As for the other major states I had listed:

PA was a shot in the dark looking at VBM data. I saw dems having a "near insurmountable" lead going in, and the GOP turned out with their own copy of it. The result is that I struck a bullseye for this state.

WI I had made every single call for how the state would trend and move correctly, except for two areas: the WOW counties (where Trump failed to improve there), and the two main cities in WI having higher turnout than expected. But the miss in the WOW counties would have been just enough for the state to tip ever so slightly blue.