Daniel W. Drezner is a professor of international politics at the Fletcher School of Law and Diplomacy at Tufts University and a regular contributor to PostEverything.

Exactly one year today, the hard-working staff here at Spoiler Alerts wrote a garbage fire of a column titled, “Why Donald Trump will be good for the Republicans in 2016.” In that lighthearted little post, I offered three counterintuitive-y reasons why this was so:

  1. Trump will lose, and lose big.”
  2. Trump makes the real GOP candidates look good.”
  3. Trump offers an opportunity for GOP contenders to shine in the debates.”

Spoiler Alerts takes great pride in pointing out the myriad policy inanities of the Trump campaign. In the spirit of openness, however, let’s pause for a brief moment to appreciate the mind-blowing retrospective stupidity of that column:

So what did I get wrong? Heck, what didn’t I get wrong!

Looking back, I think I made three mistakes. First, I overestimated the degree to which voters would value (at least some) political experience. In his campaigns and in the debates, Trump did very well mocking the stupidity and fecklessness of standard GOP politicians. For a party that has been primed to reject the very idea of expertise in some issue areas, this proved to be a more potent tactic than I had anticipated.

Second, I badly overvalued the abilities of the other GOP candidates. As some non-Trump campaign operatives told the Huffington Post’s Sam Stein, the other candidates thought Trump would self-destruct because of all the ridiculous things he was saying. To be fair, I thought this, too, but I’m not paid to run a campaign.

Throughout the fall, most of the other GOP candidates were too clever by half, focusing their critical fire and attack ads on each other rather than Trump. None of the “serious” candidates really were able to cut Trump down to size in any of the debates. The one moment I thought another candidate was going to become the consensus anti-Trump candidate was the moment when Sen. Marco Rubio self-destructed. Sen. Ted Cruz ran an extremely smart tactical campaign, and he bandwagoned with Trump until the last possible moment. By the time any of them realized that Trump might actually win, it was too late.

Third, and most important, I overestimated GOP voters. I foolishly thought that the good people of New Hampshire and South Carolina and Nevada, having dealt with presidential candidates for most of their lives, would recognize a total charlatan when they saw one. As it turns out, not so much. Trump didn’t win a majority in any of these states, but there were still way too many GOP voters who bought his shtick.

So here we are, a year later, in a place that social scientists like to call “off the equilibrium path.” When political scientists assume that people are rational, it means that we assume actors will consider all of their options and discard the choices that they know will be suboptimal in the future. When people make an inferior choice, they are off the equilibrium path. And as Matthew Continetti writes today in the Washington Free Beacon, the Republican Party has gone way, way off the equilibrium path:

Every week that Donald Trump remains the Republican nominee, the party comes closer to removing itself from the presidential gene pool. Self-selection is at work here. Trump’s supporters are choosing their party’s demise.

Want proof? Alan Abramowitz, a political scientist at Emory University, built a model that has correctly predicted presidential elections since 1992. The model says the GOP is set to win the presidency this year, 51 percent to 49 percent. But Abramowitz says to ignore his findings, not because they are wrong but because they describe an election that is not actually taking place. “The model is based on the assumption that the parties are going to nominate mainstream candidates who will be able to unite the party, and that the outcome will be similar to a generic vote, a generic presidential vote for a generic Democratic versus a generic Republican.”

Translation: If Republican voters had nominated a typical candidate, a governor or former governor who had won office in a big state by straddling the center and the right, that man would be ahead of Hillary Clinton right now. But instead the voters went for Trump.

Indeed, this time around, factors that political scientists usually discount, such as campaign strategy and negative advertising, might actually play a significant factor because one of the major parties has unilaterally disarmed itself.

Of course, I was badly, horribly wrong about the primary phase of this campaign — a fact that will haunt me for the rest of 2016. I still think that the general election will play out very differently from the primary. Trump’s declining polling numbers in the wake of the Orlando attacks, in contrast with Paris/San Bernardino, Calif., highlights this difference.

Still, this anniversary reminds the hard-working staff here at Spoiler Alerts to focus a wee bit more on world politics and less on the 2016 election. Because when it comes to campaign prognostications, I’m clearly a hack.