This seems like selective math.
The states that you win that are the closest are as much of your electoral college victory as are the states you win by a landslide. The reason most people focused on Pennsylvania, Michigan, and Wisconsin is that they were comfortably blue states in 2012, that all swung for Trump in 2016, and could be won back because of their slim margins. Biden won all three states in 2020 by a larger share than Trump won those 3 states in 2016, AND Biden flipped Arizona, Georgia, and NE-2, and increased the margin in some key states.
- Michigan:
2016: Trump: ~10,000
2020: Biden: ~150,000
- Pennsylvania
2016: Trump ~44,000
2020: Biden ~80,000
- Wisconsin
2016 Trump: ~23,000
2020 Biden: ~20,000
Biden won these three states by 240,000 votes in 2020. Pennsylvania and Michigan individually are more votes in 2020 than Trump in 2016 of all three of these states.
And then if you add in other flips between 2016 and 2020
- Georgia
2016: Trump ~210,000
2020: Biden ~13,000
- Arizona
2016: Trump ~90,000
2020: Biden ~11,000
There are very few states that Trump performed better in 2020 than in 2016, and very few consequential ones. For instance, Maine, New Hampshire, Minnesota, and Virginia were all closer in 2016 than they were in 2020. Maine was 3% for Clinton, New Hampshire 0.3% for Clinton, Virginia 5% for Clinton, Colorado 5% for Clinton, etc... In 2020, Maine 7%, New Hampshire 7%, Virginia 10%, Colorado 13% for Biden. Trump did better in Florida, 3% after narrowly winning by 1% in 2016. Many other swingier states moved back closer, Trump won Iowa by 10% in 2016, but just 8% in 2020, Ohio was about the same in both. Trump won North Carolina by 3.5% in 2016, barely 1.5% in 2020.
I have to look at the results in a table to compare the easily won states.
But, this just seems selectively choosing which states put you over 270. The most obvious states to look at are Wisconsin, Pennsylvania, and Michigan, as those three flipped the E.C. for Trump in 2016 and they were very narrow; adding Georgia and Arizona doesn't
weaken Biden's victory, it makes it stronger.
Using the metric of "the 3 closest states" isn't a great metric for determining how close the election is. Obama in 2008 won North Carolina by just 13,500 votes, Indiana by 25,000, but his EC win was a landslide by modern standards, where this metric would make it look pretty close.
As votes are still being counted, final tallies arranged, it'll be good to look back at 2020 and 2016 and decide where work needs to be done. 3 weeks out, we know that Democrats need to make up losses in the Latino & Hispanic community at a faster pace than they gained votes from white suburbanites. Democrats didn't just lose Latinos in Florida and along the Texas border, but in weird places you wouldn't expect like in Lawrence Massachusetts, a city that votes overwhelmingly for Democrats, but the Latino community swung that about 5% to Trump in 2020 from 2016... The share was still overwhelmingly Democratic, nearly 50%+, but..... it's less than in 2016, and some early prognosticating seems to point at the precinct level about poor performance. Now, there's a question there ... Biden expanded Massachusetts by 5% from Clinton in 2016, so does it matter? Y'know, do you just ignore the areas you lost because you expanded more in other areas (in this cases, rural areas in central and Western MA and near the South Shore voted more for Biden than they did Clinton in 2016).