You have put together a string of wishful assertions, accepting poll results you like and rejecting those you don’t, in order to conclude Trump will lose in 2020. Then you say McConnell made a big mistake by not switching out Pence for Trump.
As you concede, Trump’s poll numbers are rising. But you turn that on its head and conclude that increased turnout of Blacks and Women will doom his chances. However you ignore polls that show his Black support is rising.
No I don’t see him getting 34% of the Black vote as the polls indicate. But could he go from 8% to 13%? Very possible.
I agree a differential non-response could be at work in his 49% approval rating from Gallup, his highest ever. Head-to-head polls are likely more indicative anyway. But in many states, the 2016 polls significant underpredicted Trump’s actual vote. The polls said Hillary was such a lock in Wisconsin that it wasn’t even necessary for her to campaign there. Are those the polls you are relying on? Further despite all polling to the contrary, Trump won a majority of White women. If you correct for the “shy Trump voter effect" and the existing differential non-response that likely undersamples Trump’s support, you are left with the possibility Trump could cruise to victory. I think it is too soon to say much either way, but rising approval numbers are not a bad sign for Trump. Certainly it is impossible to conclude a Pence substitution would be much better.