Modeled To Death

Minnesotans know from experience that computer models are not perfect predictors. Every winter, the weatherman tells us, “We’re tracking a storm out of the Rockies that could bring between 2 inches and 9 feet of snow, depending on which direction the storm tracks.” We don’t shut down schools and churches and businesses Just In Case the worst case cenario might arrive. We wait to see and adjust our plans as better data becomes available.

I wonder if the reason we’re cowed by the COVID computer models is because we’ve been indoctrinated to believe that people in white lab coats know more than we do, so we should suspend critical thinking and trust them implicitly? I suspect that’s why presenters in television commercials and the cosmetic saleswomen at Dayton’s wore lab coats.

When the storm fails to appear, the weatherman doesn’t claim to have saved all our lives with his storm advisory. We know that’s bunk. There was no storm, he was Chicken Little.

If the virus storm fails to appear, I doubt Governor Walz will be as humble.

Joe Doakes

Invoking “Science!” (without the including the data to allow critical thought and analysis by those equipped to do so) or its weasel cousin the “evidence-based” argument is certainly a form of logrolling.

12 thoughts on “Modeled To Death

  1. “All models are wrong, but some are useful”

    A model is neither science nor evidence, it is merely a tool that requires good data and common sense.

  2. The reports that the “experts” data models are flawed, especially when predicting the deaths, are coming out daily, yet the lame stream media isn’t reporting it. Further, it’s come out that the CDC is making healthcare workers report any respiratory related deaths as due to Covid-19, without testing to determine whether or not the person had it.

    It is also concerning that Fauci, a big fan of the promise of hydroxychloroquine in 2013, is suddenly urging caution, despite the facts that the FDA has already approved it for emergency use and doctors across the world are reporting continued success with treating the virus with it. We also have the CDC, several governors and now, some legislators that are even criticizing Trump, claiming that he said people should take it, which I never heard him say. Apparently, in Utah, the control and dispensing of hydroxichloroquine has been taken over by the state. Over reach much? Then, if we look at the drugs themselves, we find that the patents on the most common treatments, the hc and the anitbiotic, expired in 2017. The conspiracy theorists posit that Fauci and the CDC leadership are deep staters, in the pockets of Big Pharma and Big Pharma can’t make as much money, so they want officials to keep killing the economy until they can develop a vaccine for which they can charge whatever they want for it.

  3. I suppose you could simply use your sharpie to get the results you desire.

    Because of the lack of data in the United States, these models are not particularly accurate. Averaging the results may remove some of the technical uncertainty about the virus but lack of data will make the results noisy. Stay home

  4. What now seems like years ago Trump used a map from either NOAA or the National Weather Service. Someone else used the one he didn’t, hence two different predictions. The big brains on the left decided this meant that saying sharpie was a really cutting remark.

    Years later emetic still thinks saying sharpie is clever. I once wondered why people like him troll websites. I now have my answer: he’s too stupid to have friends.

  5. Back on the testing hobby-horse, I see.

    Testing does not matter. Dying matters. Critical illness matters. Serious illness matters.

    The Governor need not lock down the state until every single person has a virus test because (a) people don’t need a test to know if they’re sick, the fever and coughing are sufficient clues to seek medical attention; (b) because a test is a snapshot in time that changes the minute the test patient encounters the next person in line at the grocery store; and (c) eventually, everybody in the entire world will be exposed to the virus and will either die or will develop antibodies, because that’s just how the human body works.

    We are presently testing people who show up with symptoms. There aren’t very many in Minnesota. And very few of them have died, almost all of them old people in nursing homes (remember that expensive end-of-life care you were fretting about last year? Those people.) These facts lead me to conclude that the general societal restrictions which might still be appropriate for New York City may no longer be appropriate for Minnesota.

  6. It’s not just modelling, New Jersey Urgently Needs COBOL Programmers.

    I couldn’t think of a better example of the incompetence of state bureaucracy.

    Let’s envision their thinking. “It’s written in COBOL, so we need COBOL programmers.”

    No, you don’t.

    Any competent programmer can learn COBOL in an afternoon. Proficiency would come in about a week. Learning the editor and control language to compile and link the programs might take a little longer.

    But what is really needed is the experts who understand the business and business systems. It is one thing to read a COBOL program, anyone can do that. It is quite another to understand what the program is trying to accomplish.

    That is what killed MNLARS.

    Every computer system should be budgeted for an overhaul every five years, not because it wears out, like an airplane engine, but because the thinking, the rules and the understanding slowly corrode. By rewriting the system, you train and retain a team who understands it intimately.

    It is really not that expensive. The money and time for doing that can be found in reducing four hour meetings to fifteen minutes, eliminating bullshit initiatives, like diversity training, and not forcing people out of the workplace because of their race, gender and sexuality.

  7. Because of the lack of data in the United States, these models are not particularly accurate. Averaging the results may remove some of the technical uncertainty about the virus but lack of data will make the results noisy. Stay home

    Tell that to the MSM. No, their standard of reporting is “If it makes Orange Man look bad, then we report it”.

  8. No, it’s not. Nobody cares how many people have it but aren’t sick. We’re all going to get it, eventually, that’s how the herd builds herd immunity.

    The key criterion is hospitals can keep up with sick people. In Minnesota, we’ve achieved it. The curve is flat enough.

    Governor Walz needs to declare victory and announce dramatic reductions in restrictions. Protect those at-risk. Liberate the rest.

  9. The same people going out buying a lottery ticket because they love their chances of winning are the same people that think a 2% death rate is really small and won’t hurt the economy.

  10. There are about 320,000,000 people in the United States. A 2% death rate would be 6,400,000.

    So far, we have 10,000.

    Let’s see, 10,000 divided by 320,000,000 gives us a death rate of . . . .00003125.

    Expressed as a percentage: 0.003125.

    Rounded to the nearest whole number: zero percent.

    Add to that the fact the people dying are ALL RETIRED so they don’t add anything to the economy anyway and yes, I’m the kind of guy who is willing to bet against zero percent odds by keeping the economy open.

    One of my areas of study was government, the study of public policy choices, the trade-offs between security and liberty, between safety and expense. This one is a no-brainer.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.