“Predictive models” rarely are predictive

Guest “Yogi Berra’ism ” by David Middleton

Source: First Coast Advisers

Predictive models dominate our lives — not always for the better
BY MERRILL MATTHEWS, OPINION CONTRIBUTOR — 05/06/20

The vast majority of Americans are completely unaware of how much models that predict the future of the economy and the climate – and now disease – dominate our lives. There’s no escaping their reach. Those models drive many of our public policy debates and much of the major legislation passed by Congress. 

That might not be so bad if the models’ predictions were generally accurate. But they aren’t. Indeed, they are often wildly wrong.

As the late statistician Prof. George E. P. Box warned us: “All models are wrong, but some are useful.”

[…]

All models are built on a multitude of assumptions, and many of those assumptions increasingly reflect the ideological and political views of the modelers. If the assumptions are skewed, so will be a model’s predictions.

[…]

Environmentalists, the left and most of the media accuse skeptics of being “climate deniers.” But what many on the right are skeptical of isn’t actual scientific data, but some climate models’ predictions of temperatures and sea level rise 50 or 100 years in the future. 

And yet the media regularly conflate the two. If you don’t believe the predictions of a climate model then you are denying the science, when what’s actually being questioned is many of the assumptions built into the model.

Here’s an example. Nearly all climate models in the late 1990s and early 2000s greatly overestimated rising temperatures because they didn’t take into account what’s now known as the “warming hiatus” that lasted about 14 years – from about 1998 to 2012 – when global temperatures remained relatively flat. 

In other words, the actual data did not match the models’ predictions, which left climate modelers and environmentalists scrambling to explain the discrepancies.

[…]

And yet leftists and environmentalists want us to dramatically alter the economy and our way of life – e.g., through the Green New Deal – based on predictions that might, but probably won’t, be correct.

And speaking of predictions that aren’t correct, can we talk about those coronavirus pandemic models? A new National Bureau of Economic Research (NBER) working paper highlights just how influential – and wrong – some of the pandemic models have been.

Both U.S. and UK leaders were advocating a measured response to the coronavirus pandemic until the UK’s Imperial College-London released its model’s results predicting 500,000 deaths in the UK and 2.2 million deaths in the U.S. 

[…]

Within a couple of weeks, the Imperial College scaled back its predictions, to no more than 20,000 UK deaths. And most pandemic modelers have been revising their worst-case scenarios.

[…]

Even so, the media have been obsessed with the worst-case numbers. And anyone who raised doubts about those predictions was pilloried by the media and the left as denying the “science.” 

[…]

And yet models increasingly control our lives because policymakers use them to justify their actions and their votes.

As Dr. Anthony Fauci, the lead U.S. epidemiologist in this pandemic, recently warned, “I know my modeling colleagues are going to not be happy with me, but models are as good as the assumptions you put into them.” He’s right. 

[…]

Merrill Matthews is a resident scholar with the Institute for Policy Innovation in Dallas, Texas. Follow him on Twitter @MerrillMatthews.

The Hill

The IHME (Institute for Health Metrics and Evaluation at the University of Washington) is anther model policymakers have relied upon.

Influential Covid-19 model uses flawed methods and shouldn’t guide U.S. policies, critics say
By SHARON BEGLEY @sxbegleAPRIL 17, 2020

A widely followed model for projecting Covid-19 deaths in the U.S. is producing results that have been bouncing up and down like an unpredictable fever, and now epidemiologists are criticizing it as flawed and misleading for both the public and policy makers. In particular, they warn against relying on it as the basis for government decision-making, including on “re-opening America.”

“It’s not a model that most of us in the infectious disease epidemiology field think is well suited” to projecting Covid-19 deaths, epidemiologist Marc Lipsitch of the Harvard T.H. Chan School of Public Health told reporters this week, referring to projections by the Institute for Health Metrics and Evaluation at the University of Washington.

Others experts, including some colleagues of the model-makers, are even harsher. “That the IHME model keeps changing is evidence of its lack of reliability as a predictive tool,” said epidemiologist Ruth Etzioni of the Fred Hutchinson Cancer Center, who has served on a search committee for IHME. “That it is being used for policy decisions and its results interpreted wrongly is a travesty unfolding before our eyes.”

[…]

The chief reason the IHME projections worry some experts, Etzioni said, is that “the fact that they overshot will be used to suggest that the government response prevented an even greater catastrophe, when in fact the predictions were shaky in the first place.” IHME initially projected 38,000 to 162,000 U.S. deaths. The White House combined those estimates with others to warn of 100,000 to 240,000 potential deaths.

[…]

IHME uses neither a SEIR nor an agent-based approach. It doesn’t even try to model the transmission of disease, or the incubation period, or other features of Covid-19, as SEIR and agent-based models at Imperial College London and others do. It doesn’t try to account for how many infected people interact with how many others, how many additional cases each earlier case causes, or other facts of disease transmission that have been the foundation of epidemiology models for decades.

Instead, IHME starts with data from cities where Covid-19 struck before it hit the U.S., first Wuhan and now 19 cities in Italy and Spain. It then produces a graph showing the number of deaths rising and falling as the epidemic exploded and then dissipated in those cities, resulting in a bell curve. Then (to oversimplify somewhat) it finds where U.S. data fits on that curve. The death curves in cities outside the U.S. are assumed to describe the U.S., too, with no attempt to judge whether countermeasures —lockdowns and other social-distancing strategies — in the U.S. are and will be as effective as elsewhere, especially Wuhan.

[…]

While other epidemiologists disagree on whether IHME’s deaths projections are too high or too low, there is consensus that their volatility has confused policy makers and the public:

— Last week IHME projected that Covid-19 deaths in the U.S. would total about 60,000 by August 4; this week that was revised to 68,000, with 95% certainty that the actual toll would be between 30,188 and 175,965.

— On March 27, it projected that New York would see 10,243 deaths (and that the total had a 95% chance of falling between 5,167 to 26,444) by early August. Three days later, the New York projection was 15,546, and on April 3 it was 16,262, Jewell and her colleagues pointed out in another analysis, published in JAMA on Thursday.

— On April 8, IHME projected 5,625 deaths for Massachusetts by August; on April 13, it was 8,219.

[…]

A different, data-driven model from researchers at the University of Washington predicts “about 1 million cases in the U.S. by the end of the epidemic, around the first week in June, with new cases peaking in mid-April,” said UW applied mathematician Ka-Kit Tung, who led the work. “By the first week of June, we project that the number of new cases will be close to zero if current social distancing policies are maintained.” That model predicted two weeks ago that the number of new daily cases would peak around now, as seems to be the case.

Stat News

One of the major pitfalls in using predictive models to drive policy decisions, is that no matter what happens, it always would have been worse, if we hadn’t followed the model-driven opinions of “experts”…

The chief reason the IHME projections worry some experts, Etzioni said, is that “the fact that they overshot will be used to suggest that the government response prevented an even greater catastrophe, when in fact the predictions were shaky in the first place.”

Stat News

Rarely do we ever have a way to determine whether or not “the government response prevented an even greater catastrophe.” One of the clearest examples of being able to demonstrate that “the predictions were shaky in the first place.” was the 2009 economic stimulus bill.

The Grand Obama Illusion: Major Promises Never Delivered

Kyle Smith

Summarizing his wonderful, magisterial book, The Discoverers, in an interview, historian Daniel Boorstin said, “The great obstacle to progress is not ignorance, but the illusion of knowledge.”

Fast forward to one of the great illusions of our time, the infamous chart that precisely laid out exactly what the unemployment rate would be at each stage of the recovery, with and without the Obama stimulus package. Today the chart is a monument to folly. It is not merely incorrect; it is stunningly off. It might as well have been produced by a witch doctor or by random guessing.

[…]

The chart by Obama economic elves Jared Bernstein and Christina Romer — it would be granting them far too much dignity to call them “economists” — tells us that, were it not for the miracle of the stimulus, we would be stuck with unemployment of about 5.7 percent today, but with the stimulus we were told to expect a jobless rate of about 5.2 percent. Instead, unemployment is at 7.8 percent, and the $800 billion we spent on snake oil stimulus has vanished as the disease it purported to cure continues to ravage us.

There is a direct line between the arrogance of the chart and the personality of the Commander in Chief. Obama is a frightening combination: He possesses both a proudly non-empirical mind — he admitted on The Tonight Show this week that his math skills began to fail him as early as seventh grade and that homework in the subject done by his ninth-grader daughter baffles him — and an absolute faith in those who call themselves scientists. Like the most devout churchgoers, he admits to no understanding of how those he worships works, yet is prepared to defend everything they do. The difference is that people of faith don’t get to redirect hundreds of billions of dollars of other people’s money to their belief system.

[…]

Forbes, October 31, 2012

“It’s tough to make predictions, especially about the future”… But it’s easy to get away with bogus predictions, if there’s no way determine what would have happened under different conditions. In the case of the worst COVID-19 model, Covid Act Now, they just keep shifting doomsday to the day after people are allowed to go back to work.

Texas began lifting restrictions on May 1. The Covid Act Now model is particularly useless, because it has two options 1) current trend and 2) no restrictions at all. In the no restrictions scenario, Texas would have had about 25,000-30,000 COVID-19 hospitalizations by now…

Covid Act Now, Texas, April 18, 2020

As of May 7, Texas has 1,750 COVID-19 hospitalizations…

TXDSHS

The latest Covid Act Now model, still shows Texas at the brink of doomsday…

Covid Act Now, Texas, May 7, 2020

To hammer home, the danger to our liberty and prosperity, that bureaucrats armed with models present… Dallas County government officials have steadfastly relied on the Covid Act Now models in imposing restrictions on Dallas County residents.

Dallas County Judge Clay Jenkins (Fire Marshal Gump) and Covid Act Now model. https://thefederalist.com/2020/03/25/inaccurate-virus-models-are-panicking-officials-into-ill-advised-lockdowns/
Dallas County Health and Human Services Director Dr. Phillip Huang and Covid Act Now model. https://thefederalist.com/2020/03/25/inaccurate-virus-models-are-panicking-officials-into-ill-advised-lockdowns/

They recently extended the shelter in place order to mid-May, defying the state’s decision to end shelter in place at the end of April. Dallas County Judge Clay Jenkins (Fire Marshal Gump) and Dallas County Health and Human Services Director Dr. Phillip Huang cited a sudden spike in COVID-19 cases that mysteriously began on May 1, 2020 and claimed this was not due to increased testing… But Dallas County will not release data on the number of tests being performed or recoveries.

Record coronavirus cases for 2nd consecutive day in Dallas County despite no test change
BY STEFAN STEVENSON
MAY 01, 2020

Dallas County reported a single-day high for the second consecutive day on Friday with 187 new coronavirus cases and two more deaths.

[…]

“This increase in positive cases has occurred without any significant increase in testing capacity,” Dallas County Judge Clay Jenkins said in a release. “We have seen younger people dying from COVID-19 this week and today’s victims add to that list. All this illustrates why we all must make smart decisions and follow the science to flatten the curve.”

[…]

Fort Worth Startlegram

Another COVID-19 Record for Dallas County; Masks Required, Enforcement Unclear
County judge amends county order to make Gov. Greg Abbott’s recommendations for reopened services a requirement in Dallas County
By Frank Heinz • Published May 4, 2020

Dallas County Judge Clay Jenkins amended his Safer at Home order Monday to make Gov. Greg Abbott’s recommendations for businesses reopening in Dallas County to now be requirements.

[…]

“For instance, when the governor says to the fullest extent possible wear a mask as a recommendation, we would say that’s a requirement,” Jenkins said. “When the governor says in a movie theater, let’s close every other row and put two seats of separation, that’s recommended, we say in Dallas county that’s required.”

He told NBC 5 Monday that the county doesn’t plan to fine individuals, but that code inspectors could fine businesses. The amended order can be seen below.

[…]

Dallas County is now reporting a total of 4,370 positive COVID-19 cases. Dallas County has not been releasing statistics on the number of recoveries in the county saying it’s not a surveillance variable being used nationally by the Centers for Disease Control and Prevention or state health departments.

[…]

NBC5DFW

Since Dallas County won’t report daily testing data, only reporting positive tests and deaths, the claim that “This increase in positive cases has occurred without any significant increase in testing capacity,” can’t be verified. And the Dallas County claim sticks out like a sore thumb. The State of Texas publishes statewide testing data…

Texas and Dallas County COVID-19 testing data.

The rate of statewide testing clearly has been increasing, while the rate of positive tests has steadily been decreasing. Yet Dallas County claims that the positive tests spiked right when the state began reopening because the infection rate was still increasing, therefore, restrictions must be tightened. However, Jenkins claims that “the county doesn’t plan to fine individuals”… Except…

Dallas salon owner gets 7 days in jail for reopening in defiance of countywide restrictions
Published 2 days agoDallas

DALLAS – A Far North Dallas salon owner will spend 7 days in jail after she refused to apologize for opening her business in defiance of countywide restrictions.

A Dallas County judge offered Shelley Luther, the owner of Salon a la Mode, a deal: apologize for being selfish for having her salon open while everyone else’s were closed, pay a fine, shut down until Friday and she could avoid jail time.

[…]

Fox 4 Texas

Shelley Luther wasn’t jailed for violating “countywide restrictions,” she was jailed for refusing to apologize to Dallas County officials. Texas Lieutenant Governor Dan Patrick paid her fine, the Texas Supreme Court ordered her immediate release from jail and Governor Abbott has ensured that this sort of travesty doesn’t happen again.

134 thoughts on ““Predictive models” rarely are predictive

      • From Pg 771: ” Rather the focus must be upon the prediction of the probability distribution of the system’s future possible states by the generation of ensembles of model solutions”

        This perfectly highlights the idiocy of the so-called climate scientists.

        An ensemble of models, all of which are wrong, is as useless as a single model that is wrong! The probablity distribution of the ensemble will be as wrong as the members of the ensemble.

        It would be far better for everyone if we had *one* model that was accurate!

        • It’s all based on the discovery that the average of guesses of the number of jelly beans in a jar by is pretty close to correct. Pretty sure that has more to do with experience and statistics than actual science.

        • Climate science is the only field where averaging a bunch of wrong answers is capable of generating the right answer.

          • …and just start out with the IPCC”s worst most implausible no way in hell scenario

            which is exactly why the IPCC did it in the first place….so they could use it

        • Tim, these folks are all playing fast and loose with the term “model,” as though physical models, engineering models, and statistical models were all the same thing because they’re all called models.

          As you well know, they’re not the same thing.

          Statistical models (wrong but useful) include epidemiological models. They are not physically causal. They are inductive probabilistic conjectures based on guesstimates (parameters with wide uncertainty bounds).

          If the modeler is lucky, the eventual reality will conform to some scenario within the output probability distribution. Oftentimes, as in the Imperial covid model, reality turns out to be so far from the model projection that it’s on another planet.

          Engineering models include as much causal physics as is known about the system under study. This already puts them in a different class from statistical (epidemiological) models. Engineering models include parameters and phenomenological expressions where the physics is unknown. The values of the parameters are derived from experimental results, which are translated across the entire engineering specification range of the system under development.

          The engineering model then interpolates between the experimental points to predict the behavior of the system between the experimental calibration points, over the specification range. This is how airfoils are developed, for example. The results from wind tunnel experiments parameterize the engineering model. Experiments are also done to test whether the model interpolations are accurate.

          Then the model is able to predict all important behaviors of the system, within the experimentally derived bounds and across its specification range. They cannot be used to predict behavior beyond the calibration bounds; except perhaps immediately outside them, and only with serious caution and close attention to predictive uncertainty.

          Climate models are engineering models. They include known physics. They are heavily parameterized where climate physics is unknown, in order to reproduce past observables. Past observables are the calibration bounds. Climate models with parameters adjusted using the observables of the past climate, can reproduce the behavior of the past climate (though models tuned to reproduce past air temperature cannot reproduce past precipitation, and vice versa).

          But climate models are then very improperly used to predict physical behavior of the climate very far beyond their calibration bounds. This fatally violates engineering practice. The predictions are entirely unreliable.

          Finally, physical models are fully causal. Their output is a prediction fully deduced from a falsifiable physical theory. Some behavior must be parameterized because no physical theory is complete. But parameters are subject to measurement and experiment, and their uncertainty bounds are made as tight as possible.

          Molecular Dynamics (MD) is an example of a physical model. MD uses physical theory to describe the behavior and interactions of atoms, ions, and molecules in motion in solution. Gas-phase MD is very accurate.

          Solution-phase MD is a bear, because of the emergent behavior of molecules in close and dynamical contact. Water may be the most difficult medium of all.

          MD is an advanced physical theory. The physics of charge-charge and charge-dipole interactions is well understood. The atomic/ionic/molecular descriptions and interactions are based in Quantum Mechanics. Nevertheless, MD is parameterized, because the physics is not complete. Some parameters, such as ion polarizability, are not well constrained. Changing their value can strongly impact predicted behavior.

          The dissolved systems are so complex that exact predictions of behavior in solution always turn out to be wrong when compared to experiment.

          The interplay of theory and corrective experiment is how science proceeds.

          That interplay is exactly what has been torpedoed in climate modeling, by climate modelers.

          Will Happer said that of all the scientists he interviewed when head of the DOE office of science, only the climate modelers were defensive and hostile.

          The climate modelers were defensive and hostile exactly because they had torpedoed the scientific method (falsifiable theory, mortally challenging experiment) that used to reign in their field.

          They weren’t (aren’t) doing science, and didn’t want Will Happer to find out. But of course, he did find out. Will Happer now speaks of climate modeling as cargo-cult science.

          And we all know why, and why that label is fully deserved.

          • Pat,

            good post.

            It’s like if all the guesses about how many beans are in a jar are too high, and not by a little bit but by a lot.

            People would soon become suspicious that perhaps the jar didn’t have a flat bottom.

            Yet no one becomes suspicious when all the climate guesses give answers that are too high. Why is that?

            As far as climate models being engineering models, I don’t believe that for a minute. Engineering models are typically built up from the ground. For instance, a performance model for a race car could start with engine performance based on fuel/air mixture models followed by things like cylinder performance models based on valve open/closed duration, cam lobe lift, and exhaust back pressure. The performance model would then have transmission performance (e.g. slip, friction, etc) added in. And then drive train performance. And on and on.

            The climate models should include things like a model for cloud cover that actually works and is proven against empirical observation. But we already know the models don’t handle clouds very well if at all. This should then be followed with a model that accurately predicts the enthalpy that exists at all altitudes and latitudes. But we know that doesn’t happen since the models use temperature instead of enthalpy and temperature is not a good proxy for enthalpy. And so on.

            This all leads to your statement “Finally, physical models are fully causal.”. Climate models are not causal at all. A causal model would accurately predict the future from the present but would also predict the past from the present. Thus the climate models, if fully causal, would be able to accurately predict the MWP and LIA from present conditions. I’ve never heard of anyone being able to do this with a current climate model. If the models cannot accurately predict the past over long periods of time then how can they accurately predict the future over long periods of time? A fully causal study could do both.

            Current climate models are WAGS, opinions, and biases encoded into a computer program.

          • “Statistical models (wrong but useful) include epidemiological models. They are not physically causal. They are inductive probabilistic conjectures based on guesstimates (parameters with wide uncertainty bounds).”

            wrong. agent based models in epidemiology are physically based.

          • Tim, good point.

            Maybe I should have described climate models as an abuse of engineering models and their methods.

            Climate models include physics, such as the dynamics of an atmosphere on a rotating sphere. They do have a physical base. But the critical physics that can describe the workings of the terrestrial climate is either highly deficient or missing.

            For example, the atmosphere is made to be hyperviscous in order to suppress enstrophy. The simulations blow up, otherwise.

            Jerry Browning is outspokenly critical of that failing. He says the modelers are solving the wrong hydrodynamic equations.

            A proper engineering approach would be to work on handling the behavior of the real atmosphere before claiming the model is representative of it, even within the calibration regime. That achievement would take considerable time and effort.

            Modelers do none of that. They don’t seem interested in doing the really hard work of getting the physics right. The whole enterprise looks like an exercise in corner-cutting, so as to immediately jump to a global picture.

        • They can never be accurate. The future is non-linear. Black Swans popping up everywhere.
          And as the lead-in article stated, “the illusion of knowledge is our biggest enemy.”

          Case in point, the current economic Depression setting hold of the world’s economies.
          No economic model could have forecast that 6 months ago. You could have fed in all the business metrics and readings on national and regional economies, money supply metrics, interest rates, energy supplies, drilling stock market indices and running averages, rig counts, base metal supply stockpile levels, warehouse inventories, consumer sentiments, building activity and construction loans made by lenders, and perfect long-range weather projections from WeatherBell you wanted and still nothing could have predicted that come 1 June 2020 the world’s industrialized economies would be falling head-long into a major economic depression that’s about to make 1930-36 look quaint.

          And as the lead-in article stated, “the illusion of knowledge is our biggest enemy.”

          One of social media memes popular right now with the Leftist elite snob set is this one on the
          Dunning-Kruger Effect out of the social pseudo-science called Psychology.


          In the field of psychology, the Dunning–Kruger effect is a cognitive bias in which people with low ability at a task overestimate their ability. It is related to the cognitive bias of illusory superiority and comes from the inability of people to recognize their lack of ability.
          source wiki: https://en.wikipedia.org/wiki/Dunning–Kruger_effect

          This is of course used in an arrogant, Appeal to Authority fallacy, that says lesser mortals without requisite PhDs and academic credentials can’t tell bullshit coming out of our betters (the experts). It’s the “stay in your lane” mindset from “experts” who don’t like their gross errors pointed out by those outside their particular field, even much of science and statistics in multidisciplinary.

          We know this is the case from medicine. How many times have you heard of the horror stories of medical mis-diagnosis by “experts” (MDs’)? It is because we know that (misdiagnosis happens regularly in medicine), many people will wisely seek 2nd or 3rd medical opinions and specialist diagnoses in hopes to resolving what they sense is their original doctor(s) getting it wrong. We see this vividly with climate pseudo scientists like Mikey Mann who attack people who show definitively the many and frequently fatal flaws in their published papers through the years. They employ the “we’re the experts, shut up” and now they like to cite the Dunning-Kruger effect.

          How can we dismiss the Dunning Kruger claims thrown in our face? Welll we start with the obvious, that experts frequently get things badly wrong in their discipline. It was probably the basis of Richard Feynman’s famous quote of that the one who recognizes that “ignorance of the experts” is how science advances.
          My retort to a Liberal elitist on social media when I’m told how do I know better than the experts is as follows:
          “- I don’t have to be an expert to observe that something is wrong medically with a person, and maybe if they aren’t getting better when they should be, that maybe the experts’s diagnosis/projections/prescription is wrong.
          – I can look at a person who is obviously sick to the casual observer and know that they need to seek a medical professional and licensed care.
          – I can look at a model projection compared to reality and see the model is badly wrong if the model doesn’t agree with observations of reality by which it must be judged. Do I know how to fix it? No.
          – What I am not qualified to do it make diagnosis of what is wrong with someone who is obviously ill/injured, or to declare someone “healthy”.
          – This is how we identify Quackery. And that is how we also know that folks like Mann, Hayhoe, Dessler, Schmidt, Overpeck are all “quacks” in every sense of the word.”

          • Joel
            One does not have to have a PhD in ichthyology to know when a dead fish is rotting.

        • Averaging the predictions works because 1) people are pretty good guesses and 2) the number of beans doesn’t vary between guesses.

          I suppose you might test the models by starting them at, say 1900, and after each 10 year interval throw out the high and low outliers compared to the actual temperatures over the 10 years. After 10 years go to a yearly pruning. By 1940 you just might find 1 climate model that comes fairly close to working.

          • Philo,

            What good does it do to throw out the high and low if they are *all* high? If the “low” is still wrong by being too high then you still wind up with an average that is too high.

            If even one model today made predictions that were too low to match reality then perhaps some kind of weighted average could give a more accurate prediction. But that’s not likely to happen today, those associated with the low model would quickly loose their funding!

    • Thankyou, Steve. I have that one “on speed dial” in my “Climastrology Quotes” file! I firmly believe we ought to find a reason to use it at least once a week. Shouldn’t be too hard!

      • To be honest… I kept looking until Cindy Crawford went into acting… 😉

  1. If the situation is obvious, you do not need a model. If you have historic data, you can often use that in place of a model. Models are needed when you have something novel – untested, untried, and complex. Predictive models are ONLY useful if they can be tested and adjusted towards a more accurate outcome.

    Climate is not obvious. They want to simply ignore historic data because it has a different outcome from what they need. They are making predictions about far in the future which cannot be tested and adjusted except over great periods of time – but meanwhile they demand action as if their models were real outcomes.

    I would argue that modeling a pandemic is possible – it can be tested over realistic periods of time and changed to give more accurate results. The models may be inaccurate now – given they are so simplistic and we know so little about most viruses, but they can (and likely will) be a useful tool in the future.

    Climate models predictive of 100 years into the future are not useful, and likely will never be useful – they are misleading at best and sheer lies at worst. We would need hundreds of years of future data to compare against the predictions, and at that point will likely find out that climate acts chaotically and therefore cannot be accurately modeled.

    It is very sad that so many so-called scientists cannot understand the limitations of a tool. You can try to install that big pane glass window using a sledge-hammer all you want, but the results are never going to be what you had hoped for.

  2. Interesting posting, David, and my take-away will be, that in Texas as the number of tests went up the percent of positives went down. So, two comments, 1. this Chicom-19 virus, actually the second variety of the SARS virus, will have an infection profile similar to normal flu, but its mark is that it is contagious before symptoms show, and 2. as a person with 74 years of accumulated data stored in my brain, in a retrivable form, I am going to wear a face covering, use alcohol gel, and practice social distancing, partly as reasonable self-protection and partly to avoid interacting with the thought police. The issue the left ignores is that a culture must maintain sufficient wealth-generation to take care of itself, there is no money tree. Stay sane and safe (because of document number I am on lockdown today, will golf with dog caddies in back yard, malbec in good supply, nowhere near any panic).

    • Apart from the face covering, we already did those sorts of things. My wife (53, with lupus & lymphoma) and I (61, with asthma & high BP) prefer dogs over people, so social distancing was easy for us. We have 11 dogs, so we have hand sanitizer dispensers all over the house (including a 2 liter one) and in our vehicles. We had a couple of masks for sanding and painting and my wife made us a couple of cloth masks from fabric she had (mine has the Avengers on it, hers has Pomeranians). I wear the sanding mask on top of by baseball cap and pull it over my face in grocery stores… Mostly to avoid being annoyed by safety Nazis… 😉

      • Using hand sanitizers or soap, except after returning from a potentially infected environment, is probably detrimental to health. See the video from CARTA on UCSD TV below for some of the issues.

        1. Skin, a Window Into the Evolution of the Human Super-Organism:

      • Boss Corgi keeping the Poms safe ?
        😉 😉
        (had a Corgi that bossed the Doberman , big time )

    • “…as the number of tests went up the percent of positives went down.” If testing were random, that would be an odd finding. But the selection of people to be tested is not random. If you are limited in the number of tests you can give, you give them to the people who need them the most, namely the people who have symptoms, so you can choose their treatment based on the result. And many of the people who have covid-like symptoms will indeed have the disease. But if you can give more tests, then you start giving them to people who show fewer/ different/ no symptoms, and fewer of them will turn out to have the disease.

      So the shift in the % of positives is completely expected. The # of positives, otoh, will almost certainly go up, particularly since one can have it but be symptomless.

      BTW, I’m nearly your age, and trying to avoid close contact (I go out on trail runs). But do you suppose you could send some of that malbec this way?

  3. If models actually worked even marginally >50% – bookmakers would be out of business.

    If you can’t reliably model a match between 2 sides of 11 men, or a race of 8 horses, what chance the global climate!

      • That quote was not invented by Yogi Berra, he did not claim to say it, and it does not resemble REAL Yogi Berra quotes.

        Therefore we can not trust ANYTHING you write, Middleman.

        Yogi Berra fan for over 60 years here.

        From New York, when the Yankees were great.

        • Reminds me of a Casey Stengel quote…

          You have to have a catcher because if you don’t you’re likely to have a lot of passed balls.

          –Casey Stengel

    • If you can’t reliably model a match between 2 sides of 11 men, or a race of 8 horses, what chance the global climate!

      Excellent analogy (-:

      • But a very simple model got men to the moon and back. f=ma and its derivatives. It would appear that for a model to work (even though not totally correct) there should be no more than 2 constants and 1 variable.

    • In order to protect the healthcare system from being overwhelmed, hospitals and doctor’s offices were closed and healthcare workers were furloughed or fired. Cancer and other health screenings are not being done along with elective procedures and patients are dying for lack of care.

      Good job.

    • You can make money on the horses! They tend to be consistent enough -good or bad. Jockeys, annual earnings, conditions of the track … fit into the equation too. It is a full time job for the small percent that do it for a living. An uncle of mine traveled a lot to tracks in Canada and the US and used bookies to cover particular horses wherever they ran.

      We need epidemeologists. But they need better statisticians and the right data. We need weather forecasters but think about it, there is absolutely no use whatsoever for a climatologist. There are hundreds of thousands of them. What service do they provide, really? Something awful is going to happen in 10 or a hundred years? All we need is a lonely guy with a sandwich board in each large city. We used to have them and they did this for nothing. Now it’s been made into an expensive industry.

      • You can…but you need an edge to overcome the house and state takeouts. That’s typically 12-15% for straight bets and 20-27% for exotics. You are wagering vs the betting public for the rest, and they have access to the same data you do generally. In the 1970s, the edge was home-grown speed figures. And guys like Andrew Beyer realized it was better to get paid to write books and have his published in the Daily Racing Form than to use it as his edge in wagering.

        There are people who use algorithms to find inconsistencies between straight odds and exotics, but that has been increasingly more challenging because “late money” coming in from simulcast locations can be significant.

    • Oh come on. It depends entirely on what you’re modeling. Trying to model the flip of a fair coin in a single flip, not so well; trying to model it over 100 flips, pretty darn good. (Which speaks to the prediction of weather vs. climate, btw.)

      • So, like the IPCC, you argue that weather is chaotic but climate is not.

        That just demonstrates you and the IPCC do not understand what chaos is, mathematically speaking.

        It’s the mistake made in Box 10.1 (Page 894 WG1AR5 link below). They compare the natural variation and observations and then say that he difference is GHGs. But the natural variation is assumed to be low and consistent. Chaotic systems are not like that. Most of the time a butterfly wing does not cause a hurricane – but sometime they do.

        You assume that climate variability averages out and that, like a fair coin, there are no unknown forcings.
        That’s simplistic junk science.

        Reference to junk science:
        https://www.ipcc.ch/site/assets/uploads/2018/02/WG1AR5_Chapter10_FINAL.pdf

      • That only works because we know a priori that the coin flip has exactly 2 possible outcomes and that, so, assuming a standard distribution, it is simple statistics to predict the result of 100 flips. We don’t know what the range of possible future climate outcomes are, nor the probability distribution. Your analogy fails.

      • >>
        Trying to model the flip of a fair coin in a single flip, not so well; trying to model it over 100 flips, pretty darn good.
        <<

        Nonsense, we can model any number of flips of a fair coin. The function is

        \displaystyle {{\left( \frac{1}{2}H+\frac{1}{2}T \right)}^{n}}

        where n is the number of coin flips. The total number of possible arrangements of 100 heads-tails is 2^100 which equals about 1.26765e+30. To compare, the total number of seconds since the Big Bang (assuming 13.5 billion years exactly) is about 4.2602e+17. One-hundred coin flips has a lot of possible arrangements. The reason you don’t see, say all heads, is that it’s a 1/2^100 probability of occurring–very, very small chance.

        I’m not sure what you mean by modeling 100 flips. Do you think it’s always going to be 50-50? We can calculate that probability by using the binomial theorem. The probability of exactly 50 heads-50 tails is 100 choose 50/2^100 = 0.079589 or a little less than 8%.

        The probability of 51 heads-49 tails is 0.0780. The probability of being off from 50-50 by one, i.e., 51 heads-49 tails or 49 heads-51 tails is 0.156. That’s greater than exactly 50-50. To be off by 5, i.e., 55 heads-45 tails or 45 heads-55 tails is 0.0969–still greater than just 50-50. No law of averages was ever passed–fair coins don’t have memory of previous flips.

        If we sum the probabilities of 55 heads-45 tails through 50 heads-50 tails on to 45 heads-55 tails we get 0.728384. The probability of being within 5 of 50-50 is roughly three-fourths. That’s why you tend to see a more even distribution of head-tails–there are many more of those (three-fourths more) than the outliers.

        Jim

    • Mydad warned me against gambling with the memorable “You’ll never see a bookie on a Bike”. That was at the time when my transport was a bike.

  4. “But it’s easy to get away with bogus predictions, if there’s no way determine what would have happened under different conditions.”

    Reliable antibody tests should be able to tell us the extent of the virus and how well the models did.

    The big question, how many people had the virus but little in the way of symptoms(if any)

  5. Did Prof. “Lockdown” use a YAD061 moment in his, computer, prediction, like an “up-tick”? Because now, no surprise, pollies are “Wondering what can be done to fix the economy.”

  6. Here in Melbourne, Australia we are waiting for our premier Dan Andrews to make a Decision.
    But it’s too late, the people have made a decision. People have had enough and are ignoring what little lockdown we had.

  7. As for climate BS, the end is nigh but always for a little more than tomorrow and if possible after I retire, and meanwhile, I should permanently and very professionally move the goalpost to avoid being fired while my scam is exposed.

  8. Shelley Luther has become a folk hero for standing up to the ignorant judge. People contributed over $500K to a gofundme campaign set up for her by one of her fans.

    • District Judge Eric Moye should be impeached for that travesty of justice he presided over.

      • The court video was infuriating. The judge put her in jail and increased the fine because she refused to apologize for being “selfish.”

        https://www.youtube.com/watch?v=i8ktM-GuZgs

        Her business was closed for a month. She fell behind on her home mortgage, in order to pay the lease for her business. The 18 stylists who worked in her shop were contractors, not employees. They approached her about doing “house calls” for clients. Ms. Luther realized that her people needed to work and it was safer to work in the controlled environment of the shop rather than in people’s home. She decided that the right to work was more important than the “law.” When she reopened the shop, customers began arriving in droves. She immediately became a folk hero here.
        https://www.nbcdfw.com/news/coronavirus/gov-abbott-removes-jail-as-punishment-for-coronavirus-restriction-violations/2365261/

        Businesses are usually fined for these sorts of violations. Clay Jenkins and Eric Moyé imply wanted to make an example out of her because these pompous tyrants are high on power trips.

        • If Ms. Luther had tried to defy a judge here in UK, she’d have been in jail by now with a baying horde of the general public outside (keeping 6ft apart, of course) shouting for her blood. We are a cowed and obedient people here. God bless America!

        • Operating while be selfie-ish? Wow, the judge is operating outside the popular consensus.

        • The Supreme Court unanimously over turned a 9th court decision (surprise, surprise).
          Seems that a case involving a charge of enticing someone to enter the US illegally. That’s all I know about the original case.
          Seems the 9th circuit hired contacted a group of activists and asked them to make a presentation on the grounds that the law in question violated the 1st amendment. They gave the group they invited 20 minutes to make their case. They then gave the defendents 10 minutes to respond. They didn’t spend a single minute actually reviewing any of the arguments that were actually made during the lower court hearing.

          The SC spanked the 9th circuit telling them that their job was to review the arguments as presented in the lower court, not to hold a new hearing altogether.

          • Amicus curiae briefs are very common in appeal courts , even Scotus has them. Mostly activists from all sides.
            It seems in this case they went too far

    • Wasn’t the jail for contempt of court.
      She got a cease and desist letter from the court and her response was to rip it up for the cameras.
      This is the only thing that makes sense, as you would normally need a court hearing with a lawyer to be jailed for any other offence .
      Judges contempt powers can be very strong

      • She was jailed for refusing to apologize to county officials.

        Dallas County Judge Eric Moye said during the hearing that he would consider levying a fine instead of jail time if Luther would apologize and not reopen until she was allowed to do so, but Luther refused.

        https://www.nbcdfw.com/news/coronavirus/gov-abbott-removes-jail-as-punishment-for-coronavirus-restriction-violations/2365261/

        The @$$hole specifically demanded that she apologize for being “selfish.”

        • Clearly she wanted to continue the contempt and still open.
          No one is jailed for not apologizing..its just not how the legal system works
          “Moyé said during Tuesday’s hearing that he would consider levying a fine instead of jail time if Luther would apologize AND not reopen until she was allowed to do so. Luther refused.
          Refused to obey the court order to NOT REOPEN
          https://texasscorecard.com/metroplex/dallas-salon-owner-shelley-luther-fined-and-jailed/

          “Texas businesswoman who opened her Dallas salon during shelter-in-place orders found guilty of civil and criminal contempt of court despite Abbott’s intervention.
          Have you mentioned contempt of court as the ‘charge’ at all ?

          “In that time, the salon’s owner, Shelley Luther, has been given
          a citation from local officials (1)
          and a cease-and-desist order (2) from Dallas County Judge Clay Jenkins,
          as well as a temporary restraining order from the City of Dallas. (3)
          Throughout, Luther REPEATEDLY said she would not close her business or pay the citation issued against her,

          Great on her for taking a stand …. this way to the jail house. Even the President cant defy the law or a direct order from a Judge without consequences

          • She wasn’t “found guilty” of anything.

            Dallas County Judge Eric Moye said during the hearing that he would consider levying a fine instead of jail time if Luther would apologize and not reopen until she was allowed to do so, but Luther refused.

            https://www.nbcdfw.com/news/coronavirus/gov-abbott-removes-jail-as-punishment-for-coronavirus-restriction-violations/2365261/

            Dallas County Judge Eric Moyé had offered to only impose a fine if Luther apologized, but she refused. Moyé called her defiance “selfish,” saying she was “putting your own interest ahead of those in the community in which you live.”

            https://www.houstonchronicle.com/politics/texas/article/Texas-Supreme-Court-frees-jailed-Dallas-salon-15253974.php

            The @$$hole specifically demanded that she apologize for being “selfish,” to avoid jail time, without an actual trial, or an actual conviction.

            When the governor realized that liberal judges were releasing criminals from jail because of the virus, and locking up business owners, he amended his order to expressly prohibit jail time for violating the order. The Texas Supreme Court ordered her immediate release under a personal bond.

            Any further actions the county takes against her will play out in court, where it will be decided if the government has the constitutional authority to arbitrarily and capriciously decide who gets to work and who doesn’t.

          • David? You can bet your bottom dollar this judge is already going after her again. She will be driven from business and bankrupted because she publicly refused to submit. Bet on it.

  9. modelling:
    earthquake proof buildings.
    bridges.
    multi story buildings.
    just about everything nuclear power.
    moon landings
    sling shot rocketry round the solar system
    sling shot rocketry through and beyond the solar system
    Mars landings
    design of vaccines
    vehicle steering systems
    vehicle braking systems
    engines
    fuel load for planes
    plane design
    electronic circuit simulation
    pcb design
    microchip design
    thermal design
    structural integrity modelling
    prediction of electric power requirements
    prediction of road requirements
    prediction of water requirements
    where to drill for oil/gas
    where to find minerals
    etc

    these require models. those predicting future events will be less accurate but can still show the way forward.

    there is nothing you can do to get 100% accurate future prediction but that should not make you ignore them. The tsunami in japan is an example of a failed prediction – sea wall height did not take into account the fall in land height. The millennium bridge in uk did not take into account the people walking in step when the bridge moved.

    When it comes to trying to model a potential global disaster (climate change) it becomes more difficult because of the variables involved. BUT we know climate will only slowly change decades / centuries and the models have been duplicated by others.
    No one else has come up with modelling that has been verified that shows that no climate zone will be adversely affected when CO2 increases in these time scales. Why not?

    • Models are useful tools in the hands of scientists and engineers. They are weapons of mass destruction in the hands of politicians and bureaucrats.

      • To be fair, the Obama administration’s unemployment models may have been more accurate than they appear. They addressed the effect of the economic stimulus package, but did not incorporate the effects of regulatory strangulation on the economy.

      • “They are weapons of mass destruction in the hands of politicians and bureaucrats.”

        And in the hands of scientists scrounging for grant money.

      • There is a very simple a priori test that can be applied to models in general – any model whose output unequivocally supports or favors a greater role for government in society should be considered suspect, and must be met with the highest degree of skepticism. Specific examples include Keynesian economic, GCM climate forecasts, and now epidemiological projections.

    • ghalfrunt,

      What models have been duplicated based on independent assumptions? All of the models (except perhaps the Russian model) are based on the same assumptions about CO2, i.e. the same rate of introduction into the atmosphere, the same sensitivity of temperature to CO2, and the same psuedo-feedback reactions. It’s why none of the models have generated accurate predictions for 20 years. If they can’t get the short-term predictions correct then how can they possibly get long-term predictions correct?

      I have been collecting temperature data of my own here in the central US since 2002 every 5 minutes of every day. My data shows the average temp going up while Tmax is going down and Tmin is going up. Yet every study I have read for the past 20 years predict climate disaster based on Tmax going up – causing food shortages, wide areas of the globe becoming uninhabitable, and widespread loss of species ranging from insects to birds to animals.

      I have pulled cooling degree-day data from widespread stations around the globe, Brazil, China, Siberia, Africa, and the US. They all show either a decline in cooling degree-days over the past 36 months or a stagnation of the number of cooling degree-days. This would seem to indicate, to an objective observer at least, that many regions of the globe are not seeing increasing Tmax temperatures.

      A 6th grader can tell you that an average can go up from either the maximum values going up or the minimum values going up. Can anyone tell me what the *bad* impacts of Tmin going up over time might be? All I can see is longer growing seasons, higher food production because of increased plant growth at night, and fewer people dying from cold temps.

      Let me advocate one more time for the so-called “climate scientists” to move away from global “average” temperature in their models to regional cooling and heating degree day models. Impacts on the climate don’t occur at the “average temperature”, they occur at the edges of the tempeature envelope. Cooling and heating degree-day values wuld describe climate impacts far better than a “global average temp”.

    • Once again the troll claims that just because some models are useful, that the climate models must be right.

      As to models being duplicated, that’s sorta right. All the models predict warming, not surprising because all of them make the same general assumptions. Any way, the amount of warming being predicted by these different models is all over the map, and all of them predict way more warming than actually occurred.

      BTW, there are no climate models that have been verified. Period.

      • MarkW
        How do you define “verified?” I would be inclined to say that the Russian model(s) have been verified because it/they reasonably track the historical temperatures.

    • ghalfrunt:

      No one else has come up with modelling that has been verified that shows that no climate zone will be adversely affected when CO2 increases in these time scales.

      In addition to your argumentum ad verecundiam logic above (and I’m not even addressing the presupposition of “modelling that has been verified”), we can also confirm the following concerns highlighted above regarding model believalists:

      The chief reason the IHME projections worry some experts, Etzioni said, is that “the fact that they overshot will be used to suggest that the government response prevented an even greater catastrophe, when in fact the predictions were shaky in the first place.”

      From an exchange between a layman model denialist and a professional model believalist (emphasis added):

      Q: “What if we turn our focus to evaluating how successful (read: “accurate”) our models predicting the, e.g., hospital burden turned out to be?”

      A: “For IHME I would say it was damn useful for New York. They predicted more beds than needed and more vents than needed.”

      Note that for model believalists, it would appear “accuracy” means the same thing as “usefulness,” given the model believalists answer here. Further, that’s all that seems to matter where model believalists are concerned, that more x was predicted, whether it’s +1, +10, +1000, +10,000 or +10,000,000.

      But it’s never +1, +10, or +1000 is it?

      That doesn’t matter, however, because doomsayers get to push the non-falsifiable hypothesis, “well, well, it WOULDA BEEN even worse if WE hadn’t been here to model the crisis you MORON!!!”

      But if “useful” is the goal, rather than accuracy, how does anyone know that the next time we use the model it won’t underestimate what’s required (unless of course, the models are first coded to OVER PREDICT in the first place, but that’s just my coniurare doctrina)?

      As for evaluating models for accuracy . . . oh no-no-no silly boy, thou shalt not question the Lord thy Model:

      Q: “What say you? Should we use the various historical model results from past models to model the usefulness of any future models?”

      A: “probably not.”

      And why? Because of argumentum falsum dilemma:

      “The issue is you basically have two types of models.
      A) agent based models ( discrete time step event models) or mechanistic models
      B) Differential models (continuous equations ) non mechanistic

      And to make the problem harder you don’t have a lot of historical data to improve them.
      Unlike, say, weather models. You have what you have. and decisions will be made.
      Decisions to do nothing or do something. And absent a time machine that means modelling.

      And then by his own admission, the professional model believalist provides an honest assessment of modeling certain scenarios where you DON’T have any real business modeling them at all, (e.g., climate, pandemics, war) but “the experts” want you to model them anyway:

      “Now, no one whines when war modellers got it wrong. How do I know? well I did combat modelling in the 80’s. we got a ton of shit wrong. There was no choice but to give our best assessment, which we knew would be wrong and which we knew would offer no lessons for the future.”

      https://wattsupwiththat.com/2020/04/28/key-indicator-analysis-the-chinese-virus-and-the-climate-scam/#comment-2980590

      And there you have it. You end up with this:

      https://wattsupwiththat.com/2020/04/30/model-madness-parallels-between-failed-climate-models-and-failed-coronavirus-models/#comment-2982431

      • sycomputing
        I would say that if a model prediction comes within +/-10% of actual measurements that it has utility. If it gets to within +/-1% is it reasonably accurate and demonstrates skill in predicting within the constraints of its calibration limits; if it is always high (or low) then there is a systemic bias that has to be accounted for.

    • ghalfrunt, pulling a bit from memory here but Fukishima is a bad example. At the time it was built the seawall was built high enough for the known tsunami danger. Years later scientist discovered the potential for a much bigger tsunami than originally predicted, as such the seawall needed to be upgraded to meet the new threat level. Japanese government could of made them upgrade that wall but they didn’t. This wasn’t a fault of modeling but of politicians and business.

  10. My calendar model says it’s spring, specifically May 8. My thermometer outside says it’s more like February or March.

    Is this proof that warming causes cooling? Perhaps it’s just weather and global warming is just weather variability.

  11. We know that tossing a coin makes better predictions than those generated by experts. link

    We have to quit trusting experts and their models. Christopher Monckton has dealt with the problem recently but I can’t find the link.

  12. One of the better courses I had in undergraduate, was statistical mechanics. The professor started the course using weather as the introductory example for using ensembles. That more or less told us that predictability wouldn’t be easy in the physical world. For social systems, almost impossible, particularly in chaotic times like today.

  13. Without a doubt, yogi was a great in all the word, his teachings, lessons and phrases like “this does not end until it ends”, without a doubt one of the Yankees for history

  14. — Last week IHME projected that Covid-19 deaths in the U.S. would total about 60,000 by August 4; this week that was revised to 68,000, with 95% certainty that the actual toll would be between 30,188 and 175,965.

    How much does the certainty change if it’s between 30,189 and 175,964? The point being that those exact numbers sure look impressive to the rubes. It’s just gotta be all scientificky and correct.

    But I can do better than that. I predict with 99.9999999999% confidence that the deaths will be between zero and 380 million.

    So… what are our policy options?

  15. Look at the small margin that casinos have on the various games yet they make billions of dollars. If anyone could make a model that could predict the future climate with the same accuracy they also could make billions. Where are these billionaires? Now make a model of the stock market or even one industry or easier yet one company. If that could be done you could make trillions of dollars. Where these billionaires?

    • Casinos make money from betting not because they win the bets, but because they skim a bit out of each pot. Odds makers set the odds so that the amount of money bet on each side is the same.

    • Look at the small margin that casinos have on the various games yet they make billions of dollars

      the margin maybe small, but the volume is large (do your realize just how many hands of blackjack, how many spins of the roulette wheel, how many pulls on the one-armed bandits, etc a casino does in a single day? month? or year?). Also, I’d argue, the margin isn’t quite as small as it first appears. The idea it’s very small is based on the odds (while the house has the edge, it’s only a small edge statistically speaking), but the odds assumes all the players are skilled and know what they’re doing. There’s a lot of idiot gamblers that gamble without much skill at the games they gamble on, or they have the skills but don’t play sober (there’s a reason casinos like to ply their patrons with drinks, beyond the large markups on the alcohol).

  16. All knowledge about anything is, in fact, a model.

    We have no other means to predict the future.

    That is not the problem. The problem occurs if the model in use is wildly inaccurate.

  17. It’s definitely time to readjust the model assumptions.

    We now have data that was not available earlier.

    Surprising new data include:
    • Almost no transmissions happen outside.
    • NYC infection mortality data show that 2/3 of infections occurred in the home with a large % of those being shut-in elderly adults. Governor Cuomo was aghast at that finding…expecting to see grand results from the lockdown…got just the opposite.
    • Nearly half of fatalities in many States are in Nursing Homes…and not improving.
    • Those under 18 years of age rarely have any symptoms and very rarely transmit the disease.
    • Covid-19 mortality and morbidity correlates very strongly with Vitamin D levels (possibly spurious but cheap to remedy).
    • Protecting the Elderly Works IF IT’S DONE. Florida performed strick isolation of vulnerable people. The prediction by the experts (and screamed loudly for weeks in the MSM) was that the large retirement communities in Florida would be death traps. But, do to the rational strategy of focusing resources on protecting the elderly (where almost all fatalities occur) there was no severe infection hot spots in retirement centers. A few sparks were detected and put out. IN CONTRAST, the idiotic idiots in NY actually moved sick Covid-19 patients INTO nursing homes. The torrent of death that resulted would have been no surprise (if they had consulted some 4th graders beforehand).

    The new data is now showing exactly what common sense indicated from the very beginning:
    • Aggressively protect the vulnerable segments of the population. Duh ! Most States are STILL not doing this. That is just plain common sense to concentrate where all the fatalities are.
    • Send the kids back to school. Keep older teachers that don’t have immunity out of the classrooms.
    • Get people out of the house (safely). Do not all stay inside together all the time. Get lots of sunshine (Vit. D).
    • Safely stop the lockdowns ASAP.

  18. Nothing new about model absurdity. Royama, T. 1977. Population persistence and density independence. Ecological Monographs. 47:1-34. This one said– “There is no need to test its (theory) validity against observations… .” Discussion was a little better, however. At least he was honest, many assume this without proper discussion or bury it somewhere outside of abstract.

  19. In a way, the Green New Deal is a predictive model.

    One estimate shows it would destroy 75,000,000 jobs in the US!

    Some models are far worse than the problem!

  20. Model results can be wrong for several reasons.

    (1) The abstract model may not capture all the relevant behavior of the actual phenomenon being modeled. If we can’t validate the model against actual observations it is more likely something has been missed. In the case of epidemics, there is never a control case to use as a reference because there is always some policy response, or collective change in behavior. “All other things” never remain equal during a real epidemic.

    (2) The data used as model input may not be fit for purpose. Many posters and commentators at WUWT have complained this the case with COVID-19 data — inconsistent reporting criteria, incomplete data, delayed reporting, etc. The most perfect model in the world will not produce usable results when fed bad data.

    (3) Defects in model implementation. With all due respect to epidemiologists and mathematicians, computer programming is a different skill set. If this critique of the Imperial College model is even halfway correct, nobody should pay any attention to its predictions. I’ve observed that computer programmers are generally snarky about someone else’s code and I haven’t reviewed it myself, but several specific “black box” results ring alarm bells.

    (a) Multiple runs using same initial conditions and data gave different results. The researchers covered this up by describing the model as “stochastic”.

    (b) This behavior went away when run in single-threaded mode. The ability to understand and manage concurrency is one of the marks of a competent systems programmer. That the model behaved differently when run multi-threaded indicates that someone less than competent had a hand in modifications that never worked as intended and yet were still left in the code. One wonders how many papers were published before this foible was noticed.

    (c) It was reported that different results are obtained when run on different kinds of computers. Another common bug in code written by inexperienced programmers.

    These are all “black box” test results and don’t depend on anyone actually looking at the source code. As such they are much less likely to be tainted by programmer bias.

    The real issue here is we can’t be making trillion-dollar decisions based on unverified models created and maintained by inexperienced programmers and fed with questionable data. That’s not a failure of the model; that’s a failure of public institutions and elected officials.

    If an engineering firm builds a model to design a bridge which then fails under conditions within the design specs., they are professionally and perhaps even criminally liable. So when a university researcher trots out a model and demands the government bet trillions of dollars in economic consequences on it, the government should be asking the university: “Are you willing to bet the next 20 years of government grants to your institution on these results?” The people and institutions promoting these models have no skin in the game and we need to be appropriately skeptical of their claims.

    The academic mindset is simply different: “success” means you got published and your work gets cited. I remember well one time working in the computer center we had found a bug in the serial input driver such that sometimes on a clock interrupt the floating point registers would not get saved and restored properly. So one application would resume processing with the floating point register values that belonged to another application. When we told one student who had been working for months with running data through a series of fortran programs that it was possible some of his results were wrong his response was “I don’t care; my thesis got accepted”. Anyone who has worked IT in an academic environment has similar stories. You get what you reward and academics are not primarily rewarded for getting things right.

    In some industries there is a category of “man rated” software. If someone has to bet their life on it, a whole bunch of much more stringent standards are applied. This is why Boeing is in such hot water over the 737-MAX8 failure — proper review should have prevented it.

    We should not be applying one standard to Boeing and a different one to Neil Ferguson and Imperial College when the scale of potential harm is even greater.

    • “the 737-MAX8 failure — proper review should have prevented it.”

      The Max issue is special case of the practically unmanageable issue of runaway trim on 737. But it was rare, and the automatic erratic behavior is runaway trim-like, and much more common.

  21. “‘It’s tough to make predictions, especially about the future’… But it’s easy to get away with bogus predictions, if there’s no way determine what would have happened under different conditions.”

    Eloquent. And the essence of the paternal left’s political dogma.

  22. If predictive models worked these clowns would be billionaire stock traders, instead they are just lie spewing clowns.

  23. A model is valuable if it provides something you can use – pretty much every warmist AGW virtual reality I’ve seen pretty much boils down to ‘We don’t know where, we don’t know when – but something AWFUL is going to happen.’

  24. In the IPCC based models the assumption that water provides a positive feedback to the Greenhouse Effect is WRONG . This due to the omission of the thermodynamics of water evaporation which occurs at constant temperature giving a value of zero to the Planck sensitivity coefficient.
    The inclusion of this in the models would result in a net negative feedback situation.

    • Non minuscule positive feedback was also quite unlikely to exist in a system that is empirically not unstable toward warmth over very long time.

      Positive ice feedback is another thing…

  25. I don’t pay too much attention to numbers quoted on BBC News programs, but in one report that interrupted music it was stated that 40% of “normal” presumably non ICU, Non Isolation ward beds were empty in NHS hospitals. If true that is a worrying statistic, it means one of two things either there are a lot of very sick people at home, or in normal times nearly half hospital beds are filled with people who don’t need to be there. The continual scare messages designed to keep people at home will probably end up killing many thousands.

    • I wonder how many people will die of preventable things because they missed annual appointments, screenings, etc. Nobody ever talks about that, and indeed, even quantifying them will be difficult if not impossible. So we’ll probably never know the true cost of the shutdowns, nor how many were saved. That’s a huge gamble made on very shaky data.

    • Both can be true: many people get a lot of health care that never needed any (thanks costly medical insurance, thanks Medicare…) and many other people that really needed treatments don’t get it.

      Either way, the solution is rationality in medicine. Not slogans like “EBM”.

  26. Are the models themselves at fault?
    Or is it the evaluation of the models that is the real problem.
    All of us who programmed or tried to programme learned G.I.G.O and the sure knowledge a programme will carry out it the instructions given,no matter how wrongheaded.
    The ignorance of politicians and bureaucrats is legend,Garbage In garbage Out has evolved into Garbage In Gospel Out amongst our effete elites,who proudly boast of their numerical illiteracy..
    Kind of frightening,can’t add,subtract ,multiply nor divide but they control our budget and billions of stolen tax dollars.
    And now this willful ignorance has devolved into policy based evidence manufacturing,where the desired panic has “evidence” created to serve the desired ends.

    This current pandemic has those earmarks.
    Logically we should now have real world evidence of worst case.
    Where has this virus spread unchecked and untreated?
    How do the pandemic models compare?
    And where did it meet complete lockdown,with minimal spread and maximum treatment.?

    In the long run the pernicious refusal to think,endangers civilization far more than bad models.
    The magical thinking of deferring to a computer, is a boon to the fool and bandit
    Leadership has become a vanity project and those who might be capable do not want the job.

    • John
      One of the mystifying things about this pandemic is the wide variation in deaths per capita. They range from absolutely terrifying, such as in Lombardy and NYC, to few if any deaths in some countries. It is easy to claim that early and stringent lockdowns are responsible for the low infection rates in NZ and Oz, but much more difficult to explain why countries that has done little out of the ordinary are doing about as well. So far, India seems to be doing better than most countries that have employed lockdowns, when it appeared to be “an accident waiting to happen” with the population density and poverty.

    • The even greater problem with Ferguson “model” is even if the conflicting data that represents our knowledge of the disease was input as “gospel” into his model, we would still get garbage out.

  27. With a slide rule or back of envelope,the first question of ones assumptions had to be “Is it reasonable”.
    With “Simm World” the colours are so “lifelike” all reason flees?

  28. “Sheltering” at home is the best way to get infected :
    https://www.cnbc.com/2020/05/06/ny-gov-cuomo-says-its-shocking-most-new-coronavirus-hospitalizations-are-people-staying-home.html

    The fact is that the best way to suppress a virus like the SARS-COV2, the best strategy to adopt, is to do nothing except, washing your hands, going to work, to a restaurant, outside in a park, in the woods, running, jogging, biking, etc.

    Applying a strict lockdown, staying at home and closing most of the stores and restaurants has been the best way to increase the infection and the death toll.

    Those apocalyptic bullshit models are complete nonsense laid by a bunch of dangerous crooks :
    – they have to be exposed and held accountable for this tragedy and the economical and social disaster they caused.

  29. The common seasonal flues were present long before we had the ability to characterize them at a genetic-code level. Our response to these traditional viruses, which we have come to know and ‘love,’ has been to try to guess what the next season’s strain will be, and make a vaccine that addresses it. Sometimes the guess is good, sometimes it is not. My perception is that some samples are taken randomly to understand why more people are getting sick than expected. The novel COVID-19 flu is different in that, world-wide, the testing is more extensive than ever before and, as poor as the data are, we know more about the prevalence of antibodies and asymptomatic carriers than ever before.

    We are basing our epidemiological response on how we think previous epidemics behaved. What if that mental model is wrong? What if it is common for past flues to have a large percentage of asymptomatic carriers, and the typical observation of a rapid ramp up of cases peaking in late-Winter, because of rapid establishment of ‘herd immunity,’ shapes the curve and it starts to drop off even before people start building up their natural vitamin D levels? This infection-curve behavior seems to be present whether the vaccine is considered effective or not.

    The point of my musing is that there are unstated assumptions built into the epidemiological models that we are working under, which may not be justified. Just as with climate models, there appears to be a lack of transparency in what the foundational assumptions are.

    • We know the flu vaccine has zero visible macro effect all the time, but does create antibodies for most people and does match the virus some of the time. Why is that?

      One possibility is that the vaccine protects the healthy individuals that aren’t too old (very weakly, with extreme risks for some people, so that it’s crazy to vaccinate healthy individuals that aren’t too old and the vaccine doesn’t work on old people either), weakens some people so they are at risk of the flu when they are vaccinated, and increases the rate of propagation, because there will be a lot more carriers with a mild “flu” (“rhume” like, some coronavirus-like) and not with “la grippe” feeling (strong influenza syndrome).

      IOW, when you vaccinate yourself, you might put the community at risk while protecting yourself if you are in the group that tolerates the vaccine well.

  30. A real public health campaign would empower the citizenry.

    What we have now is a regime of ordering us where to stand, where we can move, where we can go.

    That’s not a public health campaign.

    Literally, beyond the orders to social distance, wear masks, and stay inside, what have we been told?

    WAIT for a vaccine… wait… just be like sitting ducks.

    That’s not America.

    We, Americans control our destiny… Americans can successfully build our health. Health science has given us the knowledge to do so.

    A real public health campaign would tell us to take our vitamins & minerals:

    Vitamin A, C, D, and minerals zinc and magnesium.

    Serious discussion of improving & building peoples’s immune system would take center stage from the White House, each governor’s executive office, and the front page of every newspaper.

    How long have we been at it with the shutdown? Two months?

    How much could each citizen build their immune system in that amount of time?

    But no, nothing like that has happened, even easy, available drugs have been discouraged by the media and many in government. And no discussion of the power of vitamins & minerals to build the immune system… to build our health.

    Have you heard one governor talk “vitamins & minerals”?

    Something about this whole situation stinks.

    • People who will die from this flu will, those who won’t, well, won’t. Shutting people in and suppressing their immune systems is stupid, that is why Democrats are all for it. Blocking DRs from prescribing pharma that works against Chinese Disease is stupid, that is why,,,,,well, you get it by now.

    • James
      Speaking for Americans, we have a history of wanting to take a pill to do something. Everything from losing weight, to helping with sex, comes out of a bottle. Instead of focusing on having a healthy diet, getting adequate exercise, and getting sunshine and fresh air, we want to take a pill; it is like paying indulgences to be allowed to sin. Humans evolved without access to vitamin pills and mineral supplements. If one truly believes that evolution fine tunes a species for survival, then it should be obvious that only under unusual circumstances should a person need medicine or supplemental vitamins and minerals, such as with the manifestation of a deficiency such as scurvy.

      The problem with recommending a regimen of vitamins and minerals for whatever ails you is that there are as many opinions on the right combinations and dose as there are people offering them. The permutations are mind boggling. It is not unlike religion where even those in the same sect have differences of opinions, let alone between sects or religions!

      Nature figured out that if one had a diverse diet, offering up opportunities for all the things that are necessary, such as trace elements, that the body would absorb what it needed, and excrete what it didn’t. Funny how that works! There have been instances where a place, such as in Australia, has soils that are deficient in cobalt, and thus the entire food chain was deficient in an essential trace element. But, those situations are rather rare, especially now that most people, who aren’t subsistence farmers, get food from all over the world.

      Do we eat foods that are over-processed? Possibly. But, instead of trying to compensate by buying a bottle of pills, one could change their diet or cooking habits. Humans are interesting creatures! They are only capable of rational behavior for short periods of time in order to achieve their irrational goals.

  31. Modelers should at least know what they are modeling. I strongly suspect that COVID-19 modelers don’t know. Please model what happens when an infection occurs in
    a) a cruise ship, with a crew of 1500 in a good health, and 2000 mostly elderly passengers
    b) an aircraft carrier, with a crew of 4500 in a good health.
    How many deaths do models predict for each?

    • Mosher
      With a hat-tip to John Conway for his game of Life, does the simulation remove a dot when a person dies, or becomes green?

    • What did the models predict about shelter in place?

      ALBANY — The majority of recently hospitalized coronavirus patients in New York are people who have followed the precaution of staying home, Gov. Cuomo said Wednesday.

      The governor said it was “shocking” that 66% of new coronavirus hospitalizations are people who are either retired or unemployed and not commuting to work on a regular basis.

      https://www.nydailynews.com/coronavirus/ny-coronavirus-cuomo-coronavirus-stats-20200506-eyqui4b5lfdn7g6cqswkf6otly-story.html

      • David

        I have asked Christopher Monckton (in a different post) why he defends strict lockdowns so strongly when Cuomo has evidence that two-thirds of NY’s new hospitalizations are people who aren’t impacted by the ‘Cease Non-Essential Work’ decree. He hasn’t responded.

        I suspect that keeping people under ‘house arrest’ may be counterproductive by keeping family members close together with uncirculating air — like the conditions found in nursing and retirement homes! The very worst hot spots (and the first in the country) have been these closed quarters. But, when someone gets an idea in their head that they think right, it takes a crowbar up alongside of the head to dislodge it. That was one of the points that Chamberlain made in defending his Method of Multiple Working Hypotheses. Unfortunately, few but geologists have ever read the paper.

        • Social distancing measures probably saved lives… The lockdowns will probably cause more harm than good.

    • If you actively tear down your immune system by isolating yourself a mask is going to be your only defence, and it will fail.

  32. From the article: “The White House combined those estimates with others to warn of 100,000 to 240,000 potential deaths.”

    Incorrect. The White House uses the figures of 100,000 to 140,000 deaths if mitigation is practiced. Sowing confusion.

    From the article: “While other epidemiologists disagree on whether IHME’s deaths projections are too high or too low, there is consensus that their volatility has confused policy makers and the public”

    So there’s no agreement on whether this model is too high or too low. Sowing confusion. I can understand why there is a consensus about confusion because confusion is everywhere as we get all sorts of opinions about these models. Most of which are wrong. So yes, there is confusion about the models but it’s not because the models are confusing, it is because people are confused about the models. Including the so-called experts. They are confused and they spread confusion as they go. Kind of like this article.

    From the article: “Last week IHME projected that Covid-19 deaths in the U.S. would total about 60,000 by August 4; this week that was revised to 68,000, with 95% certainty that the actual toll would be between 30,188 and 175,965.”

    Incorrect. Last week the actual death count was 60,000, and this week the death toll was up to 68,000. These are not predictions they are facts. Look at the numbers quoted in the above paragraph and then compare them to the actual numbers predicted by IHME (talk about confusion): The IHME initial prediction (without any information on the Wuhan virus) for mitigated deaths from the Wuhan virus was between 100,000 and 140,000. Last week they revised their number up to 134,000 deaths by Aug. 4, 2020 (after three months of Wuhan virus data). Note that the 134,000 figure falls between the two initlal estimates of the number of mitigated deaths. It looks to me like this IHME model is right on the money so far.

    From the article: “The chief reason the IHME projections worry some experts, Etzioni said, is that “the fact that they overshot will be used to suggest that the government response prevented an even greater catastrophe, when in fact the predictions were shaky in the first place.”

    Overshot? According to whom? How do they calculate this “overshot”? Are they using the mitigated deaths as an example? The mitigated death count seems to be tracking reality. How is this an overshot?

    Are they talking about the 2.2 million unmitigated deaths? How can they calculate this figure since we do not have a situation where no mitigation is taking place? This is all BS. Unsubstantiated BS, and as far as the unmitigated numbers are concerned, they don’t know the numbers are wrong. The only time we will know these numbers is after we figure out just how infectious and deadly this Wuhan virus is, and we don’t know that yet, so claiming you got it all down pat, and the unmitigated numbers are wrong, is just blowing smoke and causing more confusion among the public.

    I certainly hope that after this Wuhan virus model scientfic debacle (a debacle because of all the misinformation being put out), that the virus model criticizers turn their attention to the Human-caused climate change models and go after them with even half the vigor they are going after the virus models. And hopefully, a little more accurately.

  33. From the article: “Shelley Luther wasn’t jailed for violating “countywide restrictions,” she was jailed for refusing to apologize to Dallas County officials. Texas Lieutenant Governor Dan Patrick paid her fine, the Texas Supreme Court ordered her immediate release from jail and Governor Abbott has ensured that this sort of travesty doesn’t happen again.

    Throwing Texans in jail whose biz’s shut down through no fault of their own is wrong.”

    She was jailed for defying an order that applied to everyone in her profession, although I agree jailing her was not necessary.

    Shelley Luther claimed she was defying the order because she could not feed her children. I don’t believe her.

    She doesn’t look like she has missed many meals to me. Why don’t those who are praising her like Sean Hannity and Tucker Carlson and Laura Ingraham ask her if she received her $1200 check from President Trump, and since she has two kids, that would total $2,200 from the president. She and the two kids ought to be able to eat on that. And she has what looks like a big, strong boyfriend who got $1200, too, so I don’t believe that her kids are going hungry.

    Of course, she, like me, might have been suffering from the two week delay Nancy Pelosi and the Democrats imposed on all the payments, and hasn’t gotten a payment yet (I actually recieved mine yesterday). But if that is the case, she has relatives and State food stamp programs and many charitable organizations who are giving food away to anyone who shows up.

    I’m sure Shelley could walk into any Salvation Army facility with her two kids and tell them she was hungry and they would give her more food than she could eat. I don’t buy the “feed my children” line. It was said to gain sympathy. It worked.

    Sean, and Tucker and Laura and all the other righwing talking heads who are praising Shelley, should ask her if she got paid by the president, and what about the Paycheck Protection Plan?

    Isn’t Shelley Luther a small business owner and doesn’t the government pay small business owners and give them money so they can keep their employees paid? Why hasn’t Sean and Tucker and Laura asked Shelley Luther about these things? Answer: Because that doesn’t fit the narrative. They want to lionize her not undermine her.

    The Right suddenly wants to promote insurrection. But what about in the future when we really need people to socially isolate because we have a very deadly virus on the loose, yet we are setting this kind of an example where we put doubts in people’s heads about the government and how the situation is handled and promote people who go against the grain.

    Shelley gets let out of jail over the Wuhan virus. If the Wuhan virus were as deadly as the Ebola virus, Shelley would be thrown in jail and the key would be thrown away, and no rightwing host would be praising her, they would be condeming her for being so reckless. It’s just a matter of degree. But in the meantime these actions cast doubt during a crisis and are teaching bad lessons.

    Shelley Luther should have done what sll the other people in her profession did: Wait one more week for her business to be allowed to open. Instead she has to grandstand, and act like she is special, and claim her kids won’t eat if she doens’t work, and people eat it up.

    She’s lucky I’m not interviewing her. I would ask her how long she has not worked (I think she said 40 days). And I would ask her all about her finances, especially since she has apparently benefited greatly on a Go Fund Me page. Don’t get me wrong, I don’t think she is a fraud, I think she just said her kids were hunger as a convenient, very plausible excuse, probably the first thing that entered her mind, and I certainly don’t think she was aiming for a Go Fund Me bonanza.

    Some people are easily fooled, especially when they want to be fooled because of emotions or an agenda. I think that is the case here.

    If its OK for Shelley to defy the rules meant to make society safe, then everyone should be able to defy the rules. Anarchy anyone? That’s what is being promoted in some quarters.

    • So, all that blahblah simply to cast your vote nazism in America. Could have done that in a single sentence and moved on. Oh, and to be perfectly clear NONE of this bullsh*t was needed, its just the flu.

  34. Here’s all you need to know about the Wuhan virus models with regard to the United States:

    The IHME (Institute for Health Metrics and Evaluation at the University of Washington)

    Initial estimate of unmitigated deaths from Wuhan virus: 1 million to 2.2 million.

    Initial estimate of mitigated deaths from Wuhan virus: 100,000 to 140,000.

    Today, after three months of evaluating data on the Wuhan virus crisis, the number of unmitigated deaths is unchanged at 1 million to 2.2 million, whereas the lastest estimate for mitigated deaths is now at 134,000.

    Initial estimate: 100,000 to 140,000. Current updated estimate: 134,000

    It looks pretty consistent to me. It looks like the people doing this model made a good initial estimate, considering they had very little data to work with at the time, and were relying mainly on past experience of how previous infections unfolded.

    People trashing this virus model have an agenda, but the facts don’t back them up, at least so far. They have no error they can point to. As time goes along, it favors the model, as the latest estimate appears to affirm the initial estimate. Not easy to debunk. Just the opposite of the climate models.

  35. In the absence of verified data, the accuracy of any prediction has a probability of zero of being correct. We do not know for certain: a) how the virus is transmitted from person to person [is it aerosols, clumps of water vapor, fog, other}; b) what standard is being used to label the death as due to CV-19 [many fatalities had other major illnesses]; c) what the overall infection rate in the population in general is [there still are not enough tests for 100% coverage].; d) what the most effective treatment is [see elsewhere on this blog concerning remesdivir versus hydroquinone].
    Not knowing either the numerator or the denominator, the likelihood of death given infection is not computable.
    We have public policy being made by fear, not science. That is bad.

Comments are closed.