How to explain why Trump hasn't collapsed even with one of the worst Democratic nominees as an opponent. 9,488,789 coronavirus infections and 236,571 deaths and counting, plus the huge economic hits would have easily taken down any other President. Then there's the seemingly erratic pattern to the election polls. Let's first check out the latter using broad categorizations for regions for comparative purposes. The first number is Trump's 2016 margin and the second is Trump's latest RCP average (that includes and gives equal weight to the good, the bad and the ugly).

Southeast Region:

Virginia: -5.3 and -11.8 (swing -6.5)
North Carolina: 3.6 and -1.2 (swing -4.8)
Georgia: 4.0 and –0.8 (swing -4.8)
Florida: 1.2 and -1.3 (swing -2.5)

MidAtlantic Region

Pennsylvania: 0.7 and -4.0 (swing -4.7)
Ohio: 8.1 and -0.2 (swing -8.3)

Upper Midwest

Michigan: 0.2 and -6.6 (swing -6.4)
Minnesota: -1.5 and -4.7 (swing -3.2)
Wisconsin: 0.7 and -6 (swing -6.7)
Iowa: 9.5 and 1.4 (swing -8.1) – (DMR has it at +7 for Trump for a swing of -2.5)

West (plus TX for fun)

Arizona: 4 and -0.9 (swing -4.9)
California: -30.1 and -29.7 (swing 0.4)
Texas: 9 and 1.2 (swing -7.8)

Focus on the direction and magnitude of the swing. As the actual poll numbers are only approximations and few pollsters will nail few numbers, try to ignore that for this review.

The most obvious takeaway is that Trump is down in all of the above states except California where he may be flat. (Trump's margin in AL is down.) However, there's not a consistent pattern to the magnitude of the swings, It's large in states that were close in 2016 and states there weren't close. But it's small in FL. Yes, FL is often very close but the same is true for IA and WI.

How could such a solid and reliably red state like Texas have such a large projected swing? Georgia's is no less confounding.

At the risk of being out-to-lunch, I think there is answer for these seemingly anomalous swings.

Recall the following national age demographics from the 2016 election:
18-29 – Trump 36 Clinton 55
30-44 – Trump 44 Clinton 51
45-64 – Trump 52 Clinton 44
65-110 – Trump 52 Clinton 45

Anecdotally, it has been reported that Trump has lost older and particularly female voters. By older does that mean senior citizens? The demographic particularly hard hit with Covid-19 deaths? If so, why only women when old men are dying at an equal or higher rate? Plugging that into the above states doesn't work. Florida's senior population is considerably larger than the other subject states and the margin swing is the smallest. What does work is the larger the older population, the smaller the margin swing. Say what? It's consistent with what we know/see/experience; the older we get the less likely we are to change or admit a mistake and death is more familiar and anticipated in older populations.

Also consider that the economic impact of Covid-19 has been minimal for seniors. The hardest hit are the youngest and lowest wage earners, but they are also less likely to vote. Plus, Trump lost those under age 45 by a considerable margin in 2016. By a process of elimination, it's a portion of the middle aged voters that are dumping Trump. In 2016 they comprised 40% of the electorate. That also the peak earning age for Americans and MAGA didn't happen for them.

A a low to high rank order of the subject states by percentage of the population over the age of fifty-nine is a pretty good fit for the margin swing:

GA: -6.0
TX: -7.8
VA: -6.5
MN: -3.2
NC: -3.8
MI: -6.4
WI: -5.3
OH: -8.3
AZ: -4.5
IA: -2.5 (DMR poll)
PA: -3.3
FL: -2.5

The large OH swing may be due to poor polling or Trump's large win was an anomaly as presidential elections in Ohio tend to be close.

These aren't stable swings but a response to the horrible person in office. There can be no electoral stability as long as both parties continue delivering different versions of the same screw the people for the benefit of the elites and wealthy and championing overthrow of foreign governments,