This post is updated as of November 5 at 9:32 P.M. PST

In the month preceding America’s 2020 Presidential election, the polls told a clear story: Joe Biden was going to win, and probably by a landslide. Nate Silver, one of America’s leading poll aggregators, predicted an 89% chance of a Biden victory; while many argued that Hillary Clinton had lost despite a polling lead, Silver had only predicted a 71% chance of her victory.

The reason for the widespread belief in Biden’s predicted victory was sensible. Polling in three traditionally Democrat states that went red in 2016 were now showing strong support for Biden. These states, Pennsylvania, Michigan, and Wisconsin, had gone red last election because of their strong dislike of globalist policies that had sent industrial production from the Rust Belt to China. Trump’s prioritization of American isolationism (a Rust Belt belief that also caused some support for Bernie Sanders in the area, who shared similarly protectionist views) turned these three blue states red and won him the 2016 election. This year, with Trump’s perceived mishandling of the pandemic, they were polling strongly for Biden.

These three states were not the sole reason for Biden’s predicted victory. The group of toss-up states with polling barely in favor of one candidate over the other had expanded from traditional swing states like Florida to include previous Republican strongholds like Texas, Ohio, and Georgia. The combination of these two factors – three blue states Trump needed that were polling for Biden and many red states that were polling almost evenly  – made a Biden victory an easy prediction.

But then came election day. While many pundits had warned us that a red surge would happen early, followed by blue push back through mail-ins, the Republican wave was far more significant than expected. Of the so-called “tossup” states – Florida, Georgia, North Carolina, Ohio, Texas, and Iowa – every single one went red, and most by a significant margin. While Georgia and North Carolina have yet to be called, it is very likely that North Carolina will go red, while Georgia remains uncertain (even with 99% of votes reported, the result is not clear. Yes, that’s how close it is.)

Once Trump had gotten through the first obstacle – the “tossup” states – he needed to hit those three key blue states he had turned red last time: Pennsylvania, Wisconsin, and Michigan. In these states, which were strongly in favor of Biden during polling, he took large-margin early leads. However, mail-in ballots turned Milwaukee and Detroit blue enough that Michigan and Wisconsin’s votes have gone to Biden (although Wisconsin was EXTREMELY close, and will probably end up in court, where nothing is likely to change).

That leaves Pennsylvania. Pennsylvania started off massively in favor of Trump. An analysis of their counties show that numerous Pennsylvania counties have flipped in favor of Trump since last election- so it seemed clear that the state would go red. However, the counties that remained blue not only had higher turnout, but that turnout was even more blue than before. Will that adjustment compensate for Trump’s lead in other counties? It’s unclear. I would guess that it will, and that Biden will win the state, winning him the election, but for argument’s sake, let’s say Pennsylvania goes red, because it will likely be quite close.

If Pennsylvania goes red, the election comes down to four states: Nevada, Arizona, Georgia, and North Carolina. If Biden can win any two of these four, one of which was blue last election, he will almost assuredly be our next president. These are the states to watch (alongside PA). North Carolina seems to be going red, and Georgia is so insanely close to call that it is not even worth discussing yet. With a Georgia’s election constituting almost five million voters, the margin is razor-thin: barely two thousand votes form Trump’s lead in the Peach State. There’s only one percent of ballots remaining to be counted, but at this point, Georgia is anyone’s game. That leaves Arizona and Nevada.

While Associated Press has already called Arizona for Biden, few other news outlets have done the same, and it really does seem to be premature. With a small margin and 10% of votes left to be counted, Arizona could go red. Arizona is determined by a single county: Maricopa County. While this county is usually red, it is currently blue. However, many more votes are still coming in that could flip Arizona one way or the other. Trump would need more votes than is likely if he wants to actually turn Arizona red, but it is still possible.

The last state of our four is Nevada. Nevada’s margin is razor-thin (Georgia, Wisconsin, Nevada- notice a trend here? This was not the landslide we were expecting.) and getting even thinner. While Nevada has gone blue in the last couple of elections, our main county, Clark County, is blue by a far less significant margin than expected. There are still a quarter of Nevada’s total votes left to count, and this could change; Trump’s campaigning in the region may have the chance to turn the tide.

In conclusion: the five states to watch are Pennsylvania, North Carolina, Georgia, Nevada, and Arizona. Trump needs Pennsylvania and three other to win, which is a very tall and unlikely order. Still, the margins in both these states and the states Biden has already won was so thin that it leads us to question how, even with the hindsight from 2016, we still consistently underestimate the Republican vote.