It may protect the reputation of pollsters if they sound defensive in saying the race is too close to call, but it creates a separate danger: Muddyness makes it easier for election deniers to challenge the outcome.

Could pollsters mess things up again this year? Let’s count the ways. There is something called non-response bias; Pollsters try to adjust for the fact that they are not reaching enough respondents in a particular demographic group (voters under 40, for example) by overweighting the responses they receive from people under 40. There is the “shy voter” syndrome, in which voters who perceive a social stigma in their choices lie to pollsters. Or perhaps it’s the other way around: Voters get excited when they surprise pollsters with their unconventional ideas, but sober up in the voting booth. Or another possibility: Voters who don’t want to appear racist or sexist will speak positively about Harris when they actually lean toward Trump.

There could be a secret women’s ballot, where moderate or Republican women who Trump turned down secretly vote for Harris. They don’t even tell their husbands who they voted for, let alone a stranger on the phone. There’s also “recency bias,” where pollsters manipulate data to reflect how respondents behaved in the last election. But voters may misremember their votes and claim they only sided with the winner, like the millions of baby boomers who insist they were at Woodstock.

Survey professionals are responding more harshly than ever to these distortions, massaging the data in a way that allows for one possibility or another. Many embraced Silver’s innovation of aggregating poll averages from several different sources; this probably helps filter out statistical “noise”. But now there’s a new problem: partisan polls deliberately designed to favor one candidate elbowing their way bring the averages into the mix by skewing them. As Rick Perlstein points out, all of these variables—let’s not forget the weather—influence the polls wrote recently “No more scientific than shooting at a dartboard with a blindfold on,” wrote The American Prospect.

And need I add the complicity of the media, which has polluted the campaign dialogue with breathtaking poll coverage as if it were the real thing?

All of this makes your head pound. Youth Survey Harvard University’s survey of voters under 30 revealed results for male voters and found a 17-point lead in favor of Harris among young men who said they would definitely vote. But the poll found an 11-point preference for Trump among young men who were less confident they would vote; This rate is a change of 28 points. As always, the only poll that matters is the one voters take on Election Day.

As a species, humans hate uncertainty. Right now we’re either gorging on Halloween candy or doing deep breathing exercises to ease the anxiety of not knowing. This explains why we stick with the polls even though we are wrong again and again.

But here’s the thing: Every poll is, by its very nature, speculating about something that exists. Actually it doesn’t happen. Emphasizing an unknown future is the antithesis of living in the moment.

It’s better to wait a few days (or maybe weeks). Polls try to give confidence in an outcome. But like most trust games, you’ll probably be fooled.

Renée Loth’s column appears regularly in the Globe.