Democratic presidential candidate Bernie Sanders says he has a better chance of beating Republican front-runner Donald Trump than Hillary Clinton. (Reuters)

Denial is powerful. The extent to which the human mind can conjure up scenarios besides the one we don’t want to see is impressive, the result of some weird psychological glitch that probably evolved as a way of allowing us to keep our sanity. I don’t know; I’m not an evolutionary psychologist.

What I am is a dude who spends a lot of time looking at poll numbers, and particularly poll numbers related to the 2016 nomination contests. I am the author of various articles assessing the chances of Sen. Bernie Sanders in the Democratic iteration thereof; those articles come to the conclusion that Hillary Clinton is very, very likely to be the party’s nominee. It’s a simple function of math. Sanders needs to win a lot of states with a lot of delegates by a lot of points -- something that he’s so far shown no ability to do. He needs to win about three-quarters of the remaining delegates. Unless deus emerges from the machina, he will not.

But that’s me looking at things objectively. I suspect that Seth Abramson — a University of New Hampshire English professor and author of “Bernie Sanders Is Currently Winning the Democratic Primary Race, and I’ll Prove It to You” — is not considering the race from the same space.

Abramson’s “proof” consists of the following argument. Actually, Bernie Sanders has more support from Democrats. It’s just that no one knows who he is. So Hillary Clinton banks a lot of early votes. But then they hear about Sanders and prefer him, and that’s why voting on Election Day favors Sanders. So, really, Sanders is preferred.

The only problem with this is all of the parts.

First of all, not every state has early voting, including states that Clinton has won like Mississippi, Alabama and Virginia.

Second of all, Sanders doesn’t always win voting on the day of the election. Abramson points to North Carolina and Tuesday’s contest in Arizona — except that the two were essentially tied in Arizona, by his numbers, and Clinton won on Election Day in North Carolina. (This is waved away as being about the suppression of college-student voters or something.) He ignores, say, Georgia, where Clinton won on Election Day by a 2-to-1 margin.

But that’s beside the point. When someone votes doesn’t tell us when they decided how to vote. For that, we can look to exit polls.

In most of the contests so far, the Edison Research pollsters have asked voters themselves when they made up their minds about who to support. If Abramson’s theory is correct, more people should have made up their minds for Bernie in the last week before the election — when Sanders was unquestionably campaigning — than before that. But a look at the states for which we have data shows that it’s split. In about half the states, Sanders does better with last-week deciders, and in about half he does worse.

In fact, if we apply the percentages of votes from those who decided in the last week to the vote totals, Clinton has earned more votes than Sanders from both those who decided in the last week and those who didn’t.

So more voters who said, “I made up my mind shortly before Election Day” chose Clinton than Sanders.

Not that people who decided earlier are somehow ignorant! The subtext to Abramson’s piece is that Clinton voters are uninformed clods and Sanders voters have had the scales fall away from their eyes. Abramson says his “proof” explains “how Clinton is ‘beating’ Sanders among American voters despite having a -13 favorability rating nationally, as compared to Sanders’s +11 rating.” But that’s among all voters! Among Democrats, Quinnipiac University reported on Wednesday, Clinton's favorability is plus-65! Sanders is plus-69 — but that’s hardly the discrepancy Abramson suggests.

Anyway. Empty theory, unproven. But innovative! An interesting theory conjured up in defense of Bernie Sanders. I wonder why.