I think we've been overlooking a really important data point. That is the "Other" party ID voters. Turnout relative to registration shows all three groups of voters converging in recent days. This means that turnout is up evenly considering the change in Reg. 2016 to 2020.
But here's the thing. Registration for Dems grew 17.5%, Reps 21%, and Other grew by 37% between 2016 and 2020. So, if we look at how this affects the totals from each, Other is actually increasing the most between 2016 and 2020 as a share of the overall vote.
Other also tends to be younger. In 2018, I think the increase of this group was suggestive of younger Dem leaning voters. Their voting patterns seem to be consistent with this.
The also seem to prefer mail voting over early in-person. Basically, in a really weird 2020 way, their voting patterns may be tipping us off that they are Biden leaning voters. After all, they prefer mail and are younger.
If this is correct (the Other group is leaning strongly for Biden), then Democrats are in good position to win the election. This is fairly consistent with public polling, but what I am seeing would suggest they are even more pro-Biden than some of the polling would suggest.
Here you can see that the Other group tends to like mail, looking more like the Democrats than the Republicans. In this 2020 environment, voting preference actually does have suggestive meaning about candidate preference.
This is all very consistent with the following information. https://twitter.com/NateSilver538/status/1322188099703758850
The same thing appears to be happening in Florida, but the difference is that in Clark, we don't have a disproportionate representation of Democrats. So, we shouldn't expect this to be strongly reversed on E-day in Clark where it might be in Florida. https://twitter.com/rp_griffin/status/1322145299733893120