This is an opinion piece. I say that to distinguish it from an article that is based on science and stuff, like the recent poll from Quinnipiac University that has the Mayor of South Bend, Indiana defeating the incumbent President of the United States in his re-election bid. I confess that I did not conduct my own poll to refute that result. I simply laughed out loud when I read it. That was not very scientific of me.

The same poll had other colorful figures beating the incumbent President:

• Vermont Sen. Bernie Sanders over President Trump 51 – 42 percent
• California Sen. Kamala Harris ahead of Trump 49 – 41 percent
• New Jersey Sen. Cory Booker over Trump 47 – 42 percent.
• Massachusetts Sen. Elizabeth Warren tops Trump 49 – 42 percent;
• Former Vice President Joseph Biden ahead 53 – 40 percent

I’ve been around a while. I think I have a pretty good sense of what is possible in American politics. These results are not possible. None of them. Three of these results — Harris, Booker, and Buttigieg — qualify as “preposterous.”

The Biden result sounds like it ought to be possible, but it isn’t. Biden is Bob Dole in 1996, or John McCain in 2008: the revered party warhorse who has patiently awaited his turn, and will now get it as his last hurrah. Nobody really expects him to win… and he won’t.

A couple of days later Fox News produced another pile of science: Bernie Sanders up by 9 over Trump. Biden was leading by 10.

Either your gut tells you that something is really wrong here… or it doesn’t. If it doesn’t, these results sound plausible and you are either overjoyed or terrified, depending.

I don’t think that either Fox or Quinnipiac made this stuff up. I assume people are really telling them that they are going to vote for Bernie, or Biden, or Cory Booker. My question is how many of them actually intend to vote for Bernie, Biden, or Cory Booker… and how many of them are having some fun?

We have other kinds of election-prediction science, and few of them are anywhere near as sanguine about the Democrats’ chances as these polls. There are several economic models out there that have proven to be excellent at predicting the results of presidential elections. Virtually all of them point to a Trump win. So does history: it’s always a lot tougher to knock off an incumbent President than it looked. The last guy to get tossed out after one term had 17% inflation, odd-even gas rationing with hour-long lines at the gas stations, an overseas hostage crisis… and when he tried to fix that he crashed a bunch of helicopters in the desert. The usual hotheads who hate Trump claim that Trump is worse, but no one outside their little club seriously believes that.

We also have the phenomenon that no one can find these “disaffected Trump voters” who supported him in 2016 but are unhappy now. Try as they might, journalists cannot find large numbers of such people. Trump’s support appears to be rock-solid. In addition, turnout among the Dems was not especially low in 2016; the NYT concluded that “low turnout” cannot be blamed for Clinton’s loss. It is therefore not clear where these Magic Voters are coming from who would put Pete Buttigieg in the White House. We only know they’re showing up in polls.

It’s early yet, and polls are historically goofy this far out. Still, we do have something new going on, and it might be worth watching. Trump voters are genuinely angry with the press, and not just a little bit. They are also likely to view pollsters as an arm of the press. Trump voters have it in their power to make polling worthless if they want to. This is what it would look like if they decided to do it.