A word on the polls, as averaged finally a couple of days ago
here. They're off almost exactly like last time - ie not nearly as far off as some (Trump!) are suggesting.
They appear to have got Biden's overall share of the vote not too far off at 52%, which it probably will be close to when the counting is finally done (currently 50.5%). As last time, the thing they got more wrong was Trump's vote, which will be around 47% rather than 44% as predicted in the average.
Regarding individual states, the only ones called wrongly (on average across polls), if the numbers remain the same, are Georgia and Florida - one each for each side. But they were right that Georgia, Arizona and North Carolina would be really close. They will probably be about spot on for PA and NV. Florida is basically the only state they really misjudged.
One thought occurs to me about why this error has repeated itself. It looks like the polls have again overestimated the 'Other' share of the vote, as in 2016, and like 2016, most 'don't knows/haven't decideds' have voted Trump. But that's easily explained if so-called 'shy' Trump voters don't lie by saying they're going Democrat, but rather they lie by saying 'not decided' when they really have or by saying 'libertarian' or whatever. It would make more sense to me that they'd lie that way rather than lying about being Democrats - it's a softer lie.