The bad Electoral Vote predictions in the U.S. Presidential Election in 2016 had me swearing I wouldn’t write a blog about polls showing Trump losing AGAIN four years later. But here we are.

This week in 2016, Hillary Clinton was ahead and I blogged that I thought Trump simply didn’t have enough time to catch up before Election Day, but that it could end up being very close. While that ended up to be true for the Popular Vote, Clinton’s lead narrowed considerably after the October Surprise, and Trump pulled out an unprecedented, targeted and unexpected win, in the Electoral Vote.

Clinton was up by 7.1 points on October 18, 2016 in the the RealClearPolitics poll average — a lead which fell to 1.5 points days before the election after Comey’s “October Surprise.” This year, the Democrat candidate is up by 9.0 (down from last week’s +10.3 points, Biden’s maximum lead over the last year).

One big difference in the history of the graph, however, is that the race has never been less than 8 points in Biden’s favor over the last four months. In 2016, it was close to deadlocked several times, and you can’t even pick out a major sine-wave pattern that the 2016 polls had.

Assuming that there’s no October Surprise this year and that the margin doesn’t narrow beyond 6 points, Biden’s wider margin all this year probably means that he wins key battleground states, taking the Electoral vote easily. Since the first two Presidential Debates helped Biden more than Trump, it’s unlikely that this week’s will do much to narrow his lead). As of today, I think Biden is in a much better position to be our next President than Clinton was.

Electorally speaking, it certainly leans in Biden’s direction. The only tricky thing is that there are nearly as many unpredictable electoral votes (179) as Biden would win (216), if the election was held today. However, if you flip the tossups to where they are leading today, Biden wins with a handy victory of 357-181, a wider margin than both the prediction, and the reality of 2016.

That said, the crucial thing to remember this year, is that the RealClearPolitics Electoral Map the day of the election in 2016 failed miserably, predicting 203 -164 in Clinton’s favor, with tossups. With no tossups, the RCP 2016 Electoral prediction was 272-266 in Clinton’s favor, and if one state’s electoral votes went in a different direction, that would have been forgivable. But the reality ended up being 232-306, something the pollsters still can’t figure out. Even though the popular vote went very close to their prediction, it was the first time RCP was wrong about who would become President. For the rest of 2016, even the media stopped believing polls, yet here we are again.

The other really big thing this year is early voting; as of this writing, more than 25 million have cast their ballots, and that will likely double in another week, meaning it’s too late to make up the ground Trump may need to.

Also interesting to look at is the betting poll at PredictIt.com. This time 4 years ago, Clinton was trouncing Trump 82 to 18, but boy did that change suddenly on election night! That essentially proved that these gambling polls are useless.

I don’t think the gamblers were taking the Electoral College into account in 2016. This year, they probably aren’t making that mistake, as Biden only leads Trump by 60/40.

I just hope that the newspapers remember what happened in 2016 this time around, and don’t jump the gun on Election Night.