The only caution I’d put on that is that this could be a uniquely US phenomenon where the candidates aren’t even generally decided until six months or so before polling day - so the hypothetical polling before that point is extra hypothetical. In a parliamentary democracy like the UK, where you usually know who the party leaders are years in advance, you can generally put more weight on the voting intention polling (absent some of the unexpected curve balls you note).

The reason I make this point is that might this presidential election be unlike previous ones, in the sense that everyone today already expects a Biden-Trump match up and probably has done for several years?

So whilst I agree not to overweight one poll, I’d also not go right the other way and ignore all polling until next summer.