While Trump yammers on about HRC again, let's look at the race in 2016 v. today. It's 51 days to the election. Sept 18 was 51 days from the 2016 election. 2016 forecast final PV was 46.7-44.3 (HRC+2.4). Pretty close to final result (HRC+2.1). Today? Biden up 52.8-45.9 (+6.9).

How about 538's forecast of EC victory? 2016 it was HRC with a 61.1% chance, having dropped from 89% a month earlier. Today? Biden has a 76% of victory, stable all along, though a bit higher than two weeks ago (it was 67% on Aug 31).

I've said all along that this election is about the pandemic and not "law and order." But that's not really it either. This election is about Trump and his actual record and behavior as President. So let's look at his job approvals v. past Presidents up for re-election.

Obama and W were +2.7 and +4.7 net job approval at this point respectively. Trump's is -10.4. Trump's net approval is most like George H. W. Bush in 1992, which was -14.0. Trump is also closer to Jimmy Carter's -21.7 than he is to Obama's.

The only way to win if you have low job approvals is to destroy the opposing candidate. So let's look at RCP's average of favorability ratings for Trump and Biden. Biden is in net positive territory after both conventions and 15.8 points higher than Trump.

So here's the thing: Plenty of things can still alter the course of this campaign. But something big will have to change for Trump to have a chance to win. The PV lead is high enough that an EV/PV split is unlikely (esp. w/ Biden focusing on states that caught HRC off guard).

And if things DO change, we'll see it in the national popular vote polls first since those polls are conducted so regularly. If Biden's average PV lead drops below 4, this race will be wide open. The race DID drop in 2016. Check out the "Now Cast" on Sept 20, 2016.

Look at the national polls leading up to that near-tied "Now Cast" on Sept 20, 2016. Lots of 1 or 2 point margins either way. And very high undecideds. Compare it to recent national polls for 2020. Biden up by much more, and at or over 50.

The swing state polls are certainly useful, especially to see if the EC coheres with the popular vote. Right now it looks like a ~3-point bias toward Trump. And they tell us about potential weak points. But the national polls tell us the big picture.

To sum up: If Trump is going to win, something big has to change with this race. Every day the polls stay in place, Trump's chance to change the race drops. If all you hear is "Trump supporters are lying to pollsters!" or "but Joe is senile!", you know that Trump is still losing.

ADDENDUM: Thanks to @tedbrogan6901, here is Nate's chart of likelihood of EC victory given certain PV win thresholds. If Biden wins the PV by over 4 points, he has an 89% chance of winning the EC. Goes to 98% if he wins over 5%.

Follow-up on Fox News poll. It has Biden up 5 overall (v. up 7 in August). Both sides gained a bit after conventions. But Biden up 6 in battlegrounds (and they DID use an oversample to get a decent MOE of 4.0%), means chance of EV/PV split goes down. (Hispanic MOE 8.5% too high).