There is a world that President Trump sees when he closes his eyes, and it is to him a better world than this one.

It is a world in which more than half the country consistently views his job performance with approval, though polls understate his approval by 10 or 20 points. Where the only news from Wall Street is when stocks are going up, and one in which voters constantly gush over Trump because their retirement accounts are up by “60, 70, 80, 90 and 100 percent.” Where the jobs numbers under President Barack Obama were horrible and fake and the same numbers during Trump’s presidency a sign of his success. A world in which the proper point at which to measure economic expansion is before taking office and in which bad economic news can be blamed on his possible successor.

As far as facts go, numbers are the stubbornest of things, hard measures of reality. But the thing about numbers is that there are a lot of them, and Trump has learned that sprinkling numbers on top of other numbers can create a mist of uncertainty and, by extension, some political breathing room. He uses numbers the way others use emoji, as approximations of how he feels or he wants you to feel.

In the examples above, Trump’s willingness to misrepresent numbers to present his more perfect world is rather innocuous. His incessant misrepresentations of reality have yielded a well-earned reputation for unreliability, but for the most part, the damage done by the above claims is to American politics.

This week, Trump took this propensity into a dangerous new realm.

On Tuesday, a World Health Organization official stated that the mortality rate for covid-19, the disease caused by the virus, is at 3.4 percent globally. Asked about it during an interview Wednesday night with his friend Sean Hannity on Fox News, Trump disagreed with that number.

“I think the 3.4 percent is really a false number,” Trump said. “Now, and this is just my hunch, and — but based on a lot of conversations with a lot of people that do this. Because a lot of people will have this and it’s very mild. They’ll get better very rapidly. They don’t even see a doctor. They don’t even call a doctor. You never hear about those people.”

“So you can’t put that down in the category in the overall population in terms of this corona-flu or virus,” he said. “So you just can’t do that. So if, you know, we have thousands or hundreds of thousands of people that get better just by, you know, by sitting around and even going to work — some of them go to work, but they get better. And then when you do have a death ... all of a sudden it seems like 3 or 4 percent, which is a very high number, as opposed to ... a fraction of 1 percent. But again, they don’t —they don’t know about the easy cases because the easy cases don’t go to the hospital. They don’t report to doctors or the hospital in many cases.”

He concluded: “So I think that that number is very high. I think the number — personally, I would say the number is way under 1 percent.”

Let’s set aside the stunningly dangerous suggestion that people can simply go to work with the highly contagious virus without issue and instead focus on Trump’s math. He twice admits that he’s simply making up the percentage he’s talking about, calling it a “hunch” and saying that it’s his personal assessment. Yes, he has access to more experts on the subject than your average American, which may inform that personal estimate, but his access to experts didn’t prevent him from reiterating obviously inaccurate information at an event with drug companies earlier this week.

What Trump is arguing is that because some illness related to the coronavirus is mild, which it is, some cases may not be tracked. If you have a population of 100 people who are infected with the virus, half of whom go to the hospital and two of whom die, you may think that the mortality rate is 4 percent — two out of 50 — when in reality it’s 2 percent.

That the virus is newly emergent means that there is, in fact, uncertainty about the mortality rate. The WHO, though, was offering a new estimate of mortality that was higher than the rate that had been presumed previously. With more information about the illness and its spread, the organization now thinks that the mortality rate is higher than what it previously assumed, Trump’s hunch notwithstanding.

Of course, the best way to track the spread of the illness and the threat it poses is to test people to see if they have it. The Centers for Disease Control and Prevention was delayed in rolling out broad testing by a faulty test, something it now says is resolved. States including Oregon and New York are developing tests or doing their own testing. In at least one case, someone who later tested positive for the virus was released by a San Antonio hospital without the virus being detected. Experts believe it had been spreading for weeks in Washington state without many cases being detected.

That would seem to bolster Trump’s point about the mortality rate being lower than estimated: The bigger the pool of sick people, the smaller that percentage of deaths becomes. But there’s a difference between knowing the spread of the illness and subsequently scaling back mortality estimates and simply declaring that the rate is lower than reported based on a hunch.

To Hannity, Trump revealed his own thinking. You hear about a death and you assume the rate of death is high when it’s actually low. He’s using statistics as a measure of how people should feel about the coronavirus, not to accurately convey the threat it poses. It’s like when he once admitted that he evaluated his own net worth based on how he felt on a given day; to Trump, the mortality rate is really more of a subjective Fear Metric and not something related to the virus’s effects.

Trump has good reason to want people not to worry about the virus. He’s repeatedly considered the spread of the virus in the context of the economy, an issue important to him as a businessman and as a candidate for reelection in November. It’s important that people not freak out about the coronavirus unnecessarily, of course, but waving away the threat — and telling people that it’s not a big deal to go to work if infected with it! — is not responsible management of the issue.

We don’t know the mortality rate of the coronavirus in the United States in part because we don’t know the spread of the virus thanks to the government’s slow, faulty start in measuring it. We do know, though, that, by themselves, numbers offered by Trump aren’t trustworthy. That the world he presents is often not the real one.