The polls tend to err as a group in one direction or the other, so a fairly decisive win by either candidate is still possible. With one day left in the 2024 campaign, the polls show one of the closest presidential elections in the history of American elections.

Across the key battlegrounds collectively or nationwide, neither Kamala Harris nor Donald J. Trump leads by more than a single percentage point. Neither candidate holds a meaningful edge in enough states to win 270 electoral votes.

In the history of modern polling, there’s never been a race where the final polls showed such a close contest. If the poll averages are exactly right down to the decimal (they will not be), Harris would barely need to outperform the polls to prevail.

Polling Leader	If Polls miss like they did in …
2022	2020
National	+1 Harris	<1 Harris	+3 Trump
Wisconsin	+1 Harris	+3 Harris	+8 Trump
Michigan	+1 Harris	+6 Harris	+5 Trump
Pennsylvania	Even	+5 Harris	+5 Trump
North Carolina	+1 Trump	<1 Harris	+4 Trump
Nevada	+1 Trump	+1 Harris	+4 Trump
Georgia	+1 Trump	+2 Trump	+3 Trump
Arizona	+3 Trump	<1 Harris	+6 Trump

The 2004 election between John Kerry and George W. Bush might be the next closest election in terms of polling, but it’s hard to put it in quite the same tier. Even then, Bush was a modestly clear favourite. Kerry needed to win states like Florida and Ohio, where Bush held a consistent if narrow lead in the final polls.

This time, neither Trump nor Harris faces such a narrow path to victory. In the key tipping point state in this year’s pre-election polling – Pennsylvania – candidates can point to several high-quality polls showing them tied or ahead. For good measure, they’re both highly competitive in additional states, should Pennsylvania fall through.

Whether the final result will actually be as close as the polls show is another matter. Even in an excellent year for polling, the polls tend to err one way or another, even if only by one or two points. When these subtle errors happen, it mostly goes unnoticed and the pollsters celebrate.

This year, even a relatively accurate year for polling might yield outcomes that feel very different. Do not be surprised if either candidate winds up winning in a manner that feels fairly decisive, such as by sweeping all of the battleground states and winning more than 300 electoral votes.

What if the polls are wrong?

Of course, it’s also possible that the polls will err by more than one or two percentage points. If they do, it might yield an outcome that doesn’t just feel decisive – it could also produce a surprisingly comfortable victory. Indeed, the poll errors from just the last two election cycles might be all it would take.

Even when there aren’t good reasons to expect the polls to be wrong, they can be. This time, there’s no shortage of potential explanations.

If they do err, the range of possible outcomes is enormous. On the one hand, Democrats might still have an outside chance of winning the Senate and reclaiming full control of government. On the other, Republicans could win full control of government with a comfortable majority, including more than 55 seats in the Senate or more.

Where did the polls shift last?

While the race was fundamentally close all of October, it is pretty clear that the polls shifted toward  Trump over the last few weeks.

The national polls have become extremely close, raising the possibility of a Trump victory in the popular vote. And Harris’s once clear-but-modest lead in the so-called blue wall of Pennsylvania, Wisconsin and Michigan has all but evaporated.

If you squinted, Trump narrowly took the lead in our averages in the middle of last week. He pulled ahead in Pennsylvania after seemingly weeks of polls showing the race tied or with him ahead. In the reckoning of the statistical forecasting models, this made Trump an ever-so-slight favourite. But over the last few days, there were signs that the race had begun to swing a bit back toward  Harris:

After a multiweek streak of high-quality polls showing Trump tied or ahead in Pennsylvania, Marist College, YouGov, Muhlenberg College and The Washington Post found  Harris narrowly up there.

Then the Ann Selzer/Des Moines Register poll of Iowa – perhaps the most renowned public poll – found  Harris leading by three points in Iowa, a state that had been considered solidly Republican. ( Trump carried the state handily in both 2016 and 2020.)

Finally, the last wave of New York Times/Siena College polls showed Harris with a narrow edge in the key battlegrounds, including significant gains in the Sun Belt battlegrounds, as late-deciding voters said they were leaning her way.

Harris’s gains came as the political conversation turned toward the former president’s liabilities. This included his rally at Madison Square Garden and his former chief of staff’s warning that he would rule like a dictator.

While the news and the polls have helped give the Harris campaign a sense of momentum heading into Election Day, the overall effect on the averages has been relatively minimal. The race remains deadlocked. Historically, late shifts in the polls do not necessarily carry through to predict how the polls will err on Election Day. In the end, the voters, not the polls, will soon have the final say.

• The New Yorks Times report