Time Nov 4, 2020 at 8:00 AM.
Here is the most likely scenario at this point. Details:
Nevada (6) - Biden 49.2%, Trump 48.6% (75%)
Arizona (11) - Biden 51.0%, Trump 47.6% (84%)
Wisconsin (10) - Biden 49.6%, Trump 48.9%, (100%)
Michigan (16) - Biden 49.4%, Trump 49.1% (91%)
Pennsylvania (20) - Biden 44.8%, Trump 54.1% (80%)
North Carolina (15) - Biden 48.7%, Trump 50.1% (100%)
Georgia (16) - Biden 48.3%, Trump 50.5% (100%)
It looks like it will all depend on the final count (and recount) in Michigan. When is Michigan supposed to finish counting? By later tonight, say 7pm EST. So we will have a tentative winner by tonight.
When will we have an official winner? North Carolina has to count absentee ballots until Nov. 12, so it won't declare a winner until then. Pennsylvania won't declare a winner until "later this week", so it should be by Friday night. But Biden doesn't need those states to win..
Will this be challenged? Absolutely, since there are allegations of voter fraud, and it will go up to the Supreme Court. So we won't know the actual winner until that case is decided, probably in early December.
===================
How does this compare to my previous predictions?
1. My initial prediction on June 18, called "Sleepy Joe is going to win", predicted that Biden would win 318-220 overall and winning Florida and Pennsylvania. So I was way off.
2. My second prediction on Oct. 9, called "Trump could still win", predicted a Trump win 274-264 with Trump winning Arizona and Michigan and Biden winning Pennsylvania. I said that Trump had to win both Arizona and Michigan and I gave him 30% odds of winning. So this was still projecting a Biden win. I said "It could come down to a few thousand people in Michigan."
3. My third prediction on Oct 16, called "Revised election odds" presented the same scenario (of Trump winning Arizona and Michigan and Biden winning Pennsylvania) and gave it a 35% change. So this was also projecting a Biden win.
4. On Oct 21, I wrote a post called "Florida is the key state", and said that Trump had a 30% chance of winning Florida, but a 77% chance of winning the election if he did win Florida. The 77% chance included Trump winning Arizona, Pennsylvania, Wisconsin and Nevada. I said there was a 23% overall chance of Trump doing so, so this was projecting a Biden win.
5. On Oct 30, I wrote a post called "Arizona will decide the election". This saw Biden winning Pennsylvania and Trump winning Michigan, in which case Arizona was need to break the tie. I said Trump had a 12.5% overall of winning, so this was projecting a Biden win.
6. On Oct 31, I wrote a post called "Contested Election Very Likely". I said that if Biden didn't sweep on election night then Trump had a 60% chance of winning.
So at this point I have to say that I was wrong on every single prediction, because I never seriously thought Trump would win Pennsylvania or that Biden would win Arizona.
But I was much better than the delusional pollsters at FiveThirtyEight.com.
It's crazy that I would have to give odds of winning, the day after the election, but my confidence level is only 51% that Biden will win Michigan, and thus the election. This is turning into a repeat of 2000, however, Michigan is the state where you have to count and recount every vote instead of Florida.
========================
Update: Trump's campaign manager believes that Trump could still win Nevada and Arizona, because not all the votes have been counted. And Trump will ask for a recount in Wisconsin and Michigan. So it's not over yet. So it's too early to declare a tentative winner until we get a final count of Nevada and Arizona.