I hope this election, if it can have at least one positive result, finally puts the lid on the endless noise from know-nothing pundits. Random variation in polls ensures that one poll or another will always be right after the fact.<p>There is no way to reliably determine in advance how people are going to vote, except in the case where one candidate is overwhelmingly more popular, in which case fancy stats are unnecessary.
The USC/LA Times poll was still trash; Trump winning doesn't change that. Their poll was national (rather than broken down by states) and, as such, was meant to predict the popular vote rather than the electoral. They were suggesting a 3 point popular win for Trump and instead it's looking like it'll be a 2 point popular win for Hillary when the final vote tallies are in.