I only meant to wait a few days (for Politico's election "calls") to finalize updates on my "Election Night Open Thread w/Prediction vs. Reality Goodness" post. Then I decided to wait on some of the recount/litigation stuff. Then I got distracted. So I just wrapped that up.
Result: For the third presidential election in a row, I correctly predicted the outcomes in 48 states. That's 48 of 50 in 2016 and this year, 48 of 48 in 2012 (I didn't try to predict Florida or North Carolina that year).
Which states did I get wrong? Florida and Georgia.
Why? Well, if I knew for sure, I wouldn't have called them incorrectly, would I? But I do have some thoughts.
First, let's look at Florida on the "legit" side.
Based on my (admittedly limited) travels around rural Alachua and Levy Counties, I thought I detected less enthusiasm for Trump than in 2016. That did change as the election got very close, but I still guessed that Trump wouldn't do as well this year.
I guessed wrong.
In Alachua County in 2016, Trump racked up 46,584 votes (he lost the county to Clinton due to the Gainesville vote). In 2020, 50,972 votes. Biden did beat him by a bigger percentage than Clinton in 2016, but he did get his own turnout up.
In Levy County (no big towns there), Trump whipped Clinton 71% to 26.3%, 13,758 votes to 5,091. This year, he beat Biden 72.4% to 26.8% -- 16,749 votes to 6,205.
In Pasco County, which the political odds-makers seem to consider a good proxy for "suburban/rural GOP-leaning Florida" in general, Trump went from 58.9% in 2016 to 59.5% in 2020.
Now let's look at the "illegitimate" side -- rigging and cheating -- in both Florida and Georgia.
In the absence of good reasons to suspect otherwise, I model that stuff as a wash in each state. That is, the Republicans will get up to their usual games of making it hard for black people to vote, etc., and the Democrats will get up to their usual games of manufacturing votes that weren't actually cast, and that both things will have the same effects they usually do.
So, I assumed that in Georgia the Republicans would continue to do well at making it hard for Democrats to vote, and that the Democrats would continue to not be able to overcome that with urban vote-stuffing. My guess is that I assumed wrong. I don't think that the usual Republican efforts went any worse than expected. I think that the continuing population shift into the urban Atlanta core caused the Democratic efforts to go better than expected (and it was pretty close in 2018).
In Florida, I thought I did have good reasons to expect a change in the rigging/cheating picture.
In 2018, Florida's voters amended the state constitution to allow former felons to vote. The GOP legislature and governor did all they could to reverse the voters' will by e.g. re-defining what "completion of sentence" means, but while they had some success in the courts, I expected that former felons would succeed in registering in the six-figure range anyway, and that their votes would lean Democrat.
I also expected Democratic vote-stuffing in e.g. Miami-Dade County to go on steroids after the 2016 debacle.
I think I got both of those wrong.
In addition, I suspect that the GOP may have turned to Democrat-style vote-stuffing tactics in Cuban-American communities in e.g. Miami-Dade and Broward Counties.
On the "illegitimate" side, I think expected the situation to change in Florida and it didn't; while in Georgia, I expected the usual situation to continue and it changed.
But I'm not sure that the "illegitimate" stuff changed the outcome in either state, and if it did so it more likely did so in Georgia than in Florida.
Either way, I got those two states wrong, and will be looking for ways to not get them wrong next time.