“Trust is like glass: shatters in an instant, with a single blow, and takes a long time to restore.” — Kathy Sullivan, chief of the National Oceanic and Atmospheric Administration under President Obama

The world was riveted last week at the ruin wreaked during Hurricane Dorian. The storm blew the Bahamas to Kingdom Come. Officially 50 people are dead, but with at least 1,300 missing after almost two weeks, that count is sure to skyrocket. Seventy thousand people are without homes. Now a new storm has smashed into Grand Bahama, ravaging the island further. Through all of this, the world could not turn away.

But not from the storm damage. The world was riveted by the spectacle of the president of the United States doubling, tripling, quadrupling, and quintupling down on an outright lie that Dorian was likely to hit Alabama. He (charitably) made a mistake about the direction of the storm, causing the Birmingham branch of the National Weather Service to tweet

That should have been that. A sane human would have said, “Welp, you’re the weather people. Good job, folks!” Alabamians would’ve inched out from under couches to a summer sky and sighed in relief. Instead, Trump claimed he was right all along, tried to get FOX News to back him up, then took out a Sharpie and drew his own weather forecast on a map.

The picture of Dorian, degraded.

Bafflingly, it got worse. White House Chief of Staff Mick Mulvaney asked Commerce Secretary Wilbur Ross to fix the problem. Ross threatened to fire employees who disagreed with the president’s forecast. The National Oceanic and Atmospheric Administration issued an unsigned statement that threw NWS Birmingham under the bus. With investigations underway, the weather report was now a political document.

But why did the president care so much about covering up so obvious a lie? Well, he’s nuts. But also, he lies all the time. Specifically, he lies an average of 13 times a day, for a whopping whopperfest of more than 12,000 lies in 28 months in office. It’s unthinkable he could be that efficient.

We talk about believers in Trump living in their own reality, disconnected from any notion of the truth. That’s easy to understand, but hard to explain. Who benefits from divorcing themselves from reality? What value could there be in believing that a hurricane would hit somewhere it wouldn’t?

Game theory makes it easier to explain. In multiple studies, players were given the chance to lie about an amount of money they’d give another player. Participants were more sensitive to changes in their gain from lying than to their opponent’s loss. In general, people find it difficult to lie. But they’re more willing to get over it as incentives to lie go up. Meanwhile, as the consequences to their opponents go up, people don’t care as much.

Trump has insanely high benefits for lying. Since those that don’t like Trump really hate him, his power depends on him keeping those who like him happy. Telling them the truth regularly—admitting error, especially—probably won’t get that done. Lying to them about the magnitude of his successes makes it more likely he will get their vote again, or get them to commit chaos in his name. It might make some members of his base a bit freaked out on occasion, but their loss is low. Even when members of his base suffer—say, soybean farmers who get told China will pay for his idiotic tariffs—Trump’s payoff for caring about their suffering is lower than the value of continuing to lie. Even if the lies are blatant.

There’s a very good reason he does this: It’s a loyalty test. Just ask chess master and dissident Garry Kasparov:

Loyalty is a deeply studied concept in game theory. It’s put through millions of non-academic tests every minute. You’re probably serving as a test subject in dozens of those tests right now. Just look in your wallet. How many loyalty cards you got? Three? Ten? Twenty? You might have them for coffee shops, hair salons, airlines, grocery stores, pharmacists, hotels, and any number of other programs. They all bet on your willingness to restrict your information and make suboptimal choices. They’ll pay you to do so.

What’s in your wallet?

Here’s how loyalty cards work. Every business is incentivized to retain customers. It’s commonly (though not universally) believed that it’s more effective to keep customers than to try to get new ones. So a company will offer you an extra payoff (“a free haircut…”) if you make several payments (“…after 10 haircuts”). Now, if you like your hair stylist, you’ll sit through 10 haircuts to get the free one. I do. I love my stylist; she’s a gamer who makes me look good. Since I get about six haircuts a year, it’ll be a couple years before I get that free one. Who cares? I’m in it to win it. My hair stylist makes me like the game I’m playing.

But boy, if I didn’t. There’s no reason to believe my stylist’s shop is the best economic choice on any given day; in fact, it’s almost certainly not. If I were willing to look around, I’d save more money than the value of that free haircut, and I’d have that money right away. The use of my loyalty card is a terrible economic investment, but I don’t care. As a customer, I’m aware of the switching costs, the price of changing my behavior. I could get a bad haircut if I shopped around. Screw that. I like the service I get, I think the payoff is okay, and I don’t want to do the research. My stylist makes the switching costs high, even if every choice I make is bad.

Behaving like this can have costs I can’t see. Extreme loyalty is fundamentally a bad idea for the person being loyal. If I restrict my choices—if I don’t think—I will make bad decision after bad decision. If I have a limited amount of time to consume news, I could either look at different sources on different days, or the same source every day. If I choose only one, and if that one is the commentary shows on FOX News, I’ll only know what Hannity and Carlson and Dobbs tell me. If they lie to me, I’m only going to know lies.

In politics, a base is a group of loyal people who have intentionally restricted their choices to one set of politicians. I’m part of the Democratic base: in the last 20 years, I have only voted for a few Republicans in primaries, and only when I had no meaningful choices as a Democrat. But I’m not part of the progressive base or the mainstream Democrat base; I vary between those groups because I’m looking for good choices among (mildly) different sets of political views. Being part of a highly restricted base can hurt. For example, if you’re so wedded to progressivism that you’ll tear down a moderate candidate of your party, or so wedded to moderateness that you’ll tear down a progressive, you may cripple that candidate when you need them.

Trump’s base is built on lies: immigrants are all criminals (they’re not), voter fraud is rife (it isn’t), the media is the enemy (they aren’t). His believers are motivated to tell him they believe his lies, because they need him to represent their views. So when he says Alabama is in a hurricane path, and says over and over “No, no, it’s really true, I really mean that,” what he is saying is “You should always believe me.” They might just. They’ll make suboptimal choices to show they continue to be loyal.

Trump continues to lie because he needs to profit off future lies. He lies about immigrants because he needs people to fear them when he steals money from the Pentagon to build his wall. He lies about voter fraud because he needs to sow doubt in the election system in case he comes up short in the polls. He lies about the media because he needs people not to believe them when they say he’s unduly influenced by Vladimir Putin. He lies about Alabama because he needs his base to believe he is infallible.

This week, we saw how dangerous these lies could be. The Houthi rebels in Yemen claimed responsibility for a drone strike that disrupted Saudi Arabian oil production. Without producing any evidence, Secretary of State Mike Pompeo blamed it on Iran. Even with John Bolton pounding the pavement, the administration knows it might need a war with Iran to remain in power. Wars aren’t just things you can launch willy nilly; you need a justification. If you lie about who did something awful—say, you conflate Saddam Hussein with 9/11—you can get the justification to attack whoever you want to. But people aren’t going to believe you in a vacuum; you need to have prepped them by inundating them with lies and seeing who remains loyal. That’s how you keep your base. That is, if you build your base on lies.

Thankfully, there’s an upside to Trump lying all the time. When his people make an outlandish claim such as that Iran bombed the Saudis, Trump has already told so many lies that people outside the base don’t believe them. In essence, there’s an anti-base, one that won’t believe the administration even if it appears to be telling the truth. So Trump’s bar for war with Iran is higher in the face of lie after lie after lie. We may no longer have truth as a bedrock, but we don’t follow liars easily either. If his anemic approval polls are to be believed, more of us are in Trump’s anti-base than the base.

I can’t guarantee that Trump’s loyalty gambit won’t work out for him. The electoral college is weighted toward Republicans and also incumbents. But my hunch is that it won’t, at least in 2020. Most of the Democratic candidates are building their campaigns on truth, and the truth is more compelling than Trump’s continued dispersal of lies. If we can get out of our own way—if we can focus on the goal of defeating Trump first and enacting potentially divisive policies later—we will probably win this thing. We’ve chosen truth over blind loyalty. That’s a base I’m happy to join.

This is the 42nd installment of a series on politics and game theory. It has covered impeachment of Trump, Russian collusion, white supremacy, abortion, guns, nuclear war, debt, the NFL, sexual harassment, the Mueller probe, taxes, Trump’s first year, the Clinton Foundation, immigration, parades, the Democrats, hope, family separation, trade wars, Trump’s endgame, the New York Times op-ed, Justice Kavanaugh, Speaker Pelosi, lame ducks, the GOP legacy, the stock market, the Democratic field, shutdowns, third party candidates, the Virginia scandals, in-party impeachment, Trump’s mafia code, college admissions, William Barr, Brexit, Iran, the Mueller Report, Joe Biden, Oregon’s standoff, the environment, and Jeffrey Epstein. The first 21 of these essays are in my book Game Theory in the Age of Chaos, which you can order by clicking the link.