Statistics

Computer Models Aren’t Science — Guest Post by Robert Kurland

This piece ran originally at The American Catholic, but Kurland graciously allowed us to re-run it here. It is a topic of deep interest for us.

INTRODUCTION

We keep reading and hearing “Science (uppercase obligatory) shows…” with respect to politically correct views on the Wuhan flu, anthropic global warming, racism, economic oppression, etc….. “Believe the science.” Horse Hockey!!! Almost invariably it isn’t a scientific analysis that’s being cited but faux science, computer programs using statistical projections that have the desired conclusions built in.

But such computer modeling isn’t science. Science requires empirical tests of hypotheses such that predictions can be tested: the hypotheses are to be falsified or verified by repeated empirical demonstrations. Besides fitting data, the hypotheses must coordinate with general and subsidiary principles of science. The best representation I’ve found for how science works is the “Lakatos Scientific Research Programme,” diagrammed in the featured image above. There is an interplay—predictions, correlations, feedback—between the shells of a sphere, fundamental principles, fundamental theories, auxiliary theories, data. A more detailed description of this is given here.

As with other mathematical tools employed in scientific endeavors—calculus, linear algebra, group theory, topology, Feynman diagrams—computer programming may be accurate, but would not have intrinsic truth value. The truth value comes from measurements, replicable in many labs. That’s how science works. (See here.) It takes more than one “successful” prediction to validate a computer model, and it takes only one unsuccessful prediction to show that it’s worthless.

In this article I’ll first give a very brief background summary of the math involved in epidemiological predictions (models) and point out problem areas. I will then focus on one article that has been cited as support for restrictive measures in the pandemic and show how the predictions put forth in the article were not met, and why, therefore, this was not “science.”

THE MATH OF EPIDEMIOLOGY

(This section can be skipped by mathphobes.) The basic parameter used in epidemiological models is R0, the “reproduction number,” which can be defined as

R0 = τcd

Here,

  • “τ” is the “transmissibility, the probability of infection given contact between infected and non-infected individuals;
  • “c” tells how often infected and non-infected individuals come in contact, i.e. the rate of contact;
  • “d” tells how long the infected individual is infected, i.e. the duration of infectiousness
  • .

It should be clear that there’s a great deal of leeway and ambiguity in assigning values to any of the parameters in R0. Indeed, the possibility of errors in interpreting or estimating R0 is well recognized, as shown in the quote below:

“The interpretation of R0 estimates derived from different models requires an understanding of the models’ structures, inputs, and interactions. Because many researchers using R0 have not been trained in sophisticated mathematical techniques, R0 is easily subject to misrepresentation, misinterpretation, and misapplication.”—Paul L. Delamater et al, “Complexity of the Basic Reproduction Number (R0)

To expect that any of the parameters—τ, c, d—can be represented by just one number is, I believe, an oversimplification. To expect that these numbers will stay constant during the course of an epidemic, given the possibility of virus mutation and mitigation factors, is also an oversimplification. I recommend Delamater’s article for a full account of all the factors that can make R0 variable. Nevertheless, other auxiliary theories in science use simplifying assumptions and make successful predictions, so the outcome of predictions should be the test of whether a theory, implemented in a computer program, is valid.

A final note on the math: epidemiological modeling generally uses coupled linear 1st order differential equations and statistical methods for stochastic processes to generate predictions under various assumptions and for various parameter values. (See here and here for detailed accounts.) If one examines the papers that describe the mathematical techniques, it’s clear that outcomes can vary greatly depending on values assigned to parameters and on the choice of a particular mathematical model.

UNSUCCESSFUL PREDICTIONS FOR THE WUHAN FLU (COVID-19)

Note first that there have been many instances outside of epidemiology where computer models relying on statistical techniques have made predictions that didn’t come true. (The readers of this blog probably know of those dealing with effects of anthropic global warming.) More recently we have seen (here, here, here and here) how statistical analysis can be applied and misapplied to “data” for the Wuhan flu. Matt Briggs, “Statistician to the Stars,” has many articles on the misuse of statistics in this. I recommend his podcast for a general of account of how covid-19 data has been used and misused in statistical analysis.

In this article I want to focus on the “Report 9 from Imperial College Response Team.” Based on this report, politicians and pundits predicted dire consequences from the pandemic: total deaths, 2.2 million in the US, 510 thousand in the UK (Figure 1A, loc.cit.). When critics of the study complained that these extreme mortalities didn’t occur, Ferguson correctly responded that this would have been the totals had no mitigation efforts taken place. However, that isn’t really a prediction; it’s more like a street preacher with his sign that the world is coming to an end in one month. The preacher’s prediction might be true, but there’s no way of verifying it. There never was or would be a situation where no mitigation efforts were applied. That wasn’t even the case for the Spanish Flu epidemic 100 years ago.

Moreover, the deaths per day in Figure 1A go to 0 by September 1st, which has not been the case, mitigation or no mitigation. Further, if one examines Fig. 1B, cases per day for various US states, the curves are totally unrealistic in terms of how they vary from state to state and in their shape. Indeed, this is a general criticism of all the figures: the shapes show symmetric (or essentially symmetric) rise and falls for all mitigation scenarios, and that is not how the actual data, faulty as it may be, is displayed.

Here are some events that computer models have not taken into account. Forty-two percent of all deaths in the U.S attributed to covid-19 (notice “attributed to,” not “due to”) have been in nursing homes and assisted care facilities. In particular, these facilities have been in states where governors have directed nursing homes to accept covid-19 patients who were released from hospitals. Our neighboring county in Pennsylvania, rural and with a low population density, had a sudden spike in covid-19 cases at the end of August and beginning of September; the cause: partying in the local university. This latter spike in covid-19 cases is nationwide, wherever colleges have resumed in-person classes.

So, what are we to conclude? Is such computer modeling an exercise in using “any data input that will enable one to continue playing what is perhaps the ultimate game of solitaire?” I’ll concede that the Imperial College Report is an interesting speculation on how various mitigation programs might affect the transmission of covid-19, and in that sense it is a useful exercise, and the authors are to be commended for that effort. What I object to is the use of these computer projections by pundits and politicians to justify public health measures, because the measures are based on “science.” It isn’t science until the computer projection has made repeated correct predictions.

Here’s what might have been more helpful: applying the model to known data (e.g. South Korea, Taiwan), or if the report had been published later, Sweden, Italy, Spain and Germany, to see how well the modeling fits actual data. This would have been an exercise in retrodiction, which is sometimes useful science (e.g, explaining the anomalous perihelion precession of Mercury using general relativity). There wouldn’t have been as many headlines stemming from the report, but it would have been more of a scientific endeavor.

To support this site and its wholly independent host using credit card or PayPal (in any amount) click here

Categories: Statistics

25 replies »

  1. This is a good article and a good start, but there’s much more to be done…

    These models, of course, are not science, and what isn’t science can never become science for some random happenstance, like predictions fulfilled: logically, that’s the equivalent to put the cart in front of the horses, as we say in my country. Science is “scientia”, knowledge, naked, strict, knowledge. Computer models and predictions are some use of statistics, based on some scientific knowledge and generally accepted hypotheses. That’s not science. I mean, most of what today is called “science”, including the new “human” ones (including what Voegelin called “the new science of politics” is sociology [idelology (sic)]), is NOT science, there is a huge philosophical crisis. So, some random idiot says he proves SCIENTIFICALLY God doesn’t exist or something and people believe it’s actual science, just because he uses some language or jargon or he has made some scientific work (Hawking, for example). Again, not even “theoretical physics” is science, strictly speaking, so, not even relativity: take that. They take hypothetical bodies as science and, then, when those bodies fall, because they are only useful to classifying experimental real science and to design new experiments, consistent with the actual, proven knowledge, so it can be expanded; when those bodies fall, like Ptolomeo and many more, they cave in to postmodern misology… Misology is the cancer, philosophy, as in so many other realms, is the cure. Read Suma Theoligiae, I, q. 32, a. 2, ad 2 (I believe).

  2. Thank you for your comment, CJCG. Unfortunately the featured image in the original article, my depiction of the Lakatos “scientific research programme,” wasn’t included in this article (just a cut). So, my own position of how science works, wasn’t made quite clear. I agree with you. It takes more than successful prediction from a computer model to call a hypothesis “science.” That’s a necessary, but not a sufficient condition. What is required is integration with core scientific principles, basic theories and auxiliary theories, as that image shows.

  3. “Science” is politics now. You know, that redefinition of words thing. Like boys are girls and girls are boys, two same-sex people can marry, right is wrong, wrong is right. If you are to “save” science, you will have to save the country and language. Actually, I think “science” has been usurped by many governments throughout history, plus some tribes with medicine men who were the ultimate “scientists”. Science and religion are very often entwined. Like most everything in life, groups steal definitions and build their own lies and power on those redefinitions. It’s necessary to remind people of this, but it won’t ever really change. People have lied, cheated and redefined for thousands of years. It’s who they are. Sad, but it is what it is.

    I thought it was interesting one could model a NEW virus with which we had zero experience. They used to call that science fiction or a lie. Now we call it policy. Anyway, it obviously was not NEW at all or they would have had zero input parameters for the model. It was kind of different from past viruses of the same name, minus the “2” at the end. Since Sars 1 had a 30% death rate but apparently was less contagious, I guess they did not use that virus as the basis for the model. Or maybe they did and randomly changed a few variables. Sars 1 was also gone in 2 years and there was no “vaccine”, so I’m betting they created their own virus model based on whatever was needed to shut down the world. The idea that a completely new virus can be modeled seems nonsensical to me.

  4. From the featured article: “But in that case Lakatos is gored by the other horn of Feyerabend’s dilemma. For Feyerabend argues that a Demarcation Criterion that cannot tell anyone what to do or not to do is scarcely distinguishable from “Anything goes”. To revert to Feyerabend’s political analogy, what is the difference between an anarchist society and a “state” where the “police” can appraise people for their “criminal” or “law-abiding” behaviour but can never make an arrest or send anyone to jail? That’s a “state” which isn’t a state and a “police force” which isn’t a police force! We have not scientific law-and-order but anarchy, accompanied by uplifting sermons and benedictions posthumously bestowed on the mighty scientific dead.”

    Sounds like the news from this year.

  5. Bob, your article is very good and I learned from it. I told you the thing about predictions (I just re-read it) not in a proper manner. What you answered was spot on. I just wanted to add that the otherthings I said were to expand on your work, onlu that. I’m sorry for my floppy andclumsy ways of expression.

  6. Throughout my life I have watched people repeatedly and mistakenly accept projections or forecasts because they possessed a patina of “science” and accuracy only because they were the product of a computer.

    They are incredibly gullible and easily misled. It’s simply amazing to me.

    According to the1972 Club of Rome computer-derived predictions contained in “The Limits To Growth,” you shouldn’t be reading this— you’re dead.

  7. Not only that, they’re always incomplete and always wrong. Sometimes they give you useful insights. I run computer models for a living for DoD.

  8. A computer model is a fancy mathematical equation. Computers crunch numbers; they solve equations. They do not do experiments, although computers are a valuable tool to experimenters. They do not gather data, although computers can record and store data if so instructed.

    Is a test tube science? No, it is a tool of science, or can be. Similarly computers are tools of science, or can be. It all depends on how they are used.

    Are equations and their solutions science? No, but they can be useful to science. It’s a poor scientist that doesn’t know how to do math, i.e. solve equations.

    The argument that computer models are not science skirts the real issue. The issue is whether the science that created the models was well or poorly done. Some models are based on excellent science and can be very useful. The model that analyzes smoke in your smoke detector might save your life. The model that predicts planetary orbits might not save your life, but it could save the lives of astronauts.

    Other models, such as those cited, are based on speculative or very weak science. Mr. Kurland explains all that; I find no fault in the article. The title is misleading, though.

    Science is the quest for knowledge; sometimes it succeeds and other times it fails. Trusting “science” is like trusting a rope. Some ropes break and then you fall, while others can save your life. Know your rope.

  9. In his excellent little book debunking the putative science of evolution, “The Kingdom of Speech,” Thomas Wolfe observes that evolution ain’t science; Has anyone observed the phenomenon – in this case Evolution – as it occurred and recorded it? Could other scientists replicate it? Could any of them come up with a set of facts that, if true, could contradict the theory (Karl Popper’s falsifiabilty test)? Could scientists make predictions based upon it? Did it illuminate hitherto unknown or baffling areas of science? In the case of evolution..well…no…no…no…no…and no

  10. @ Amatuer Brain Surgeon,

    > In his excellent little book…

    How likely do you think it is that Wolfe knows more about evolution than Darwin? In fact, pretty much everything Wolfe claims in his book is either false or irrelevant. It seems to me that your idea of a good book is one that confirms what you already believe, but if you have the stomach for it, here is an excellent, entertaining takedown of Wolfe’s book:

    https://www.3ammagazine.com/3am/tom-wolfes-reflections-language/

  11. I am not a fan of Dr Karl Popper, too much of his philosophy was embraced carte blanche. Even Karl admitted he had no defense when the Danish philosophers asked Karl why they should believe that only his falsibility axiom was unfalsifiable. As a mathematician, I believe too much of induction was jettisoned and we have suffered some from this. We need to take a more careful look at our epistemology. Insisting that the only truth allowed is falsifiable, thus really not truth, is a rather dark view of reality, in my honest opinion (as a mathematician). The real issue is if knowledge can be “known”. Asserting that only valid knowledge is that “known by science” ignored much non-scientific knowledge. Mathematics is a prime example of such knowledge, very non-physical, but none of us can ignore it, nor its truth content. Karl circumscribed too much. Some of these very real issues are being ignored today.

  12. Swordfish. Confirmation bias? Prolly true 🙂

    Tom Wolfe knows that Darwin stole his ideas from Alfred Russell Wallace and he documents it in the book.

    Dear Swordfish. The book is not expensive. Buy it. Darwinism is dead – DED – dead. The book by Wolfe is a fun evisceration of that antiChristian fantasy

    If you have the liberty to believe for not believe Divine Revelation you think ABS does not have he same liberty to read and take a decision about what he has read based upon the quality of the author, his documented sources etc?

    Why would ABS care to read the opinions of a man who thinks his progenitors were apes?

  13. When science successful sued for divorce from Sacred Theology it began intellectually fornicating with the enlightenment and it produced uncountable numbers of bastids that sensible men of faith are expected to adopt house, feed, clothe and raise as their own.

    Only insane men do that.

  14. Computer models are nothing more or less than theories. A scientific theory is a model, usually mathematical, which attempts to emulate some aspect of the universe and predict unknown events given a set of conditions. In order for a theory to be considered proven, it must successfully do such predictions and never get it wrong (for a given value of “wrong”.

    Models, and even computer models, are an intrinsic part of science and even more so of engineering, so there is nothing inherently wrong with them. However, one must know their limitations and constraints.

    The problem with the climate models currently used is that they have never been validated. To do so, of course, would take several hundred years, since “climate” is generally defined as a 30 year average of the weather.

  15. @ Amateur Brain Surgeon,

    > Tom Wolfe knows that Darwin stole his ideas from Alfred Russell Wallace and he documents it in the book.

    You clearly didn’t read the review I linked to, as it completely destroys Wolfe’s claim: Darwin wrote a 230 page essay explaining his ideas about evolution 13 years before he heard from Wallace. But in any case, even if Wolfe was right about this (he wasn’t), it wouldn’t make evolution false, so Wolfe’s whole point here is irrelevant as well as factually wrong.

    > Why would ABS care to read the opinions of a man who thinks his progenitors were apes?

    To find out the truth? Apparently, that’s not something you’re interested in. Also, what kind of person refers to themselves in the third person?

  16. Re the “spikes in cases” due to colleges being back in session — are these “cases” sick people with symptoms? Or people identified by tests? If both, what percentage actually have symptoms? I am having a hard time finding this out anywhere. Even places that are quarantining students seem to be (some of them, anyway) including students with no symptoms. Can we PLEASE go back to calling only people who are ill “cases”?

  17. what kind of person refers to themselves in the third person?

    Reminds me of Eli Rabett another, er, “deep thinker” who writes in third person.
    https://wattsupwiththat.com/2013/06/29/saturday-silliness-the-only-thing-more-ridiculous-than-eli-rabett-the-required-usda-rabbit-disaster-plan/

    Seems there’s plenty of evidence for common descent. Here’s some:
    o) DNA genetic sequences of organisms: phylogenetically close have a higher degree similarity than those distant.
    o) Animal coloration: camouflage, mimicry, and warning coloration are all explainable by natural selection.
    o) Some birds have seasonal changes (e.g. white in winter and brown in summer) also compelling evidence of selection
    o) appearance of antibiotic resistant bacteria

  18. what kind of person refers to themselves in the third person?

    An Amateur Brain Surgeon.

    Dear DAV, Identify for us the pair of fornicating animals which produced an offspring with one for more different organs than its progenitors.

    If you spent a few bucks for “The Kingdom of speech” you’d know that the idea-their Darwin was completely eviscerated because he could not account for man and his language. He said maybe language developed from the imitation of animal sounds, an insane idea mocked by Max Mueller as th bow-wow theory.

    Mueller connoted to publicly mock the huckster Darwin but not by name- he didn’t have to everybody knew who was being mocked. Muller also labeled as The Ding Dong theory Darwin’s other dubious ideas which Mueller picked as the Ye-be-bo theory, the sing-somng theory the hey you theory etc etc

    In May of 2014 Norm Chomsky, and seven others, admitted there as a reason Darwin had brown eyes when it come to language , There is nothing like it in animal life… and the confession came after forty strenuous years of scientific study trying to account for it.

    ABS laughed when it was disclosed that The liberals lairs of The Linnean Society concocted a sleazy scheme in which Old Darwy could get credit for evolution rather than Wallace who created it.

    Arrange a meeting of the society, don’t invite Wallace and tell Wallace to fake an Abstract that he could present before the public the same day the work of Wallace was made public.

    One problem, no abstract of Darwin existed and so Lyell and Hooker assembled one for him, and the fraud was on and continues to this day.

    Darwy was a low life intellectual thief but ABS knows liberal love him to this day.

  19. One problem, no abstract of Darwin existed and so Lyell and Hooker assembled one for him, and the fraud was on and continues to this day.

    In the long run, it’s irrelevant who came up with the idea. Who invented the radio: Marconi or Tesla? Regardless of the answer, what would it mean in terms of radios working? Did Shakespeare write the plays or did someone else? If it wasn’t Shakespeare would that make them less important?

    Identify for us the pair of fornicating animals which produced an offspring with one for more different organs than its progenitors.

    Not at all clear if animals CAN fornicate as the definition seems to center around unmarried PEOPLE. But then people ARE animals. Things happen during reproduction. For instance, the following. Not sure if it is the result of fornication though. It’s likely that neither parent had more than two arms.

    https://i.dailymail.co.uk/i/pix/scaled/2014/12/29/2450EFC100000578-0-image-a-3_1419863298081.jpg

    You seem to be making the same argument as Intelligent Design proponents. Maybe you are one yourself. They point to the eye and claim a low probability of it coming about. But they are assuming it came about all at once. Selection theory doesn’t say “all at once”. It proposes a latching mechanism: a small but unharmful (and maybe beneficial) change occurs that can be passed on. Over time, the small change mutates gradually adding more and more capability. For the eye this could mean a novel cell that responds to light becoming groups of cells that respond.

    Bacteria mutating to become resistant to antibiotics is a prime example. They can do this in a short period of time because 1) the mechanism needn’t be complex and 2) there are a lot of reproductions thus lots of chances for mutation. Besides bacteria don’t even sexually reproduce let alone fornicate yet they change over time. Assuming the resistance was somehow inherent in the bacteria is assuming a limited number of antibiotic possibilities thus implying there is can be an antibiotic for which there is no defense. So far though, it has not been found.

    The are mathematical systems that demonstrate evolution is possible. The Game of Life was likely the first. Evolutionary Algorithms work through mutation and fitness. Evolutionary Programming can develop novel solutions within a given environment using evolutionary techniques and reinforcement learning. They aren’t restricted to any given set of operations. If they can do this within a limited environment then they can also take advantage of larger ones. The only difference between mathematical models and real life is that the mathematical ones are constrained to a rather small environment and fitness is measured by success in calculation because that’s why they are built. In real life, fitness is determined by suitability to the environment which means survivability and reproduction.

    I’m guessing you think some Sky Spirit actively created all of the diversity we see instead of perhaps setting down rules for it to have occurred over time. If not the former, how do you think it happened?

  20. @ Amateur Brain Surgeon,

    > Dear DAV, Identify for us the pair of fornicating animals which produced an offspring with one for [sic] more different organs than its progenitors.

    This indicates a total lack of understanding of evolution on even the most basic level. Whole organs don’t evolve in one generation. If this is how ignorant you are, it’s not surprising that you can be fooled by Wolfe’s irrelevant lies.

    > One problem, no abstract of Darwin existed and so Lyell and Hooker assembled one for him, and the fraud was on and continues to this day.

    From the review you didn’t read:

    “At this point Wolfe invents (again of whole cloth) an imaginary conversation in which it is not the students, but the naturalist Charles Lyell who must deal with poor simple Charles. He must help dumb Charles concoct a CONSPIRACY!, and it goes like this:

    “Oh, Charlie, Charlie, Charlie,…Who was it who told you you’d better get busy and publish this pet theory of yours? The main thing, Charlie, is to establish your work and Wallace’s. Now that’s fair isn’t it? Even-steven and all that? Well, to be perfectly frank, there is one slight hitch. You’ve never published a line of your work on evolution. Not one line… You don’t even have a paper to present at the meeting…hmmm… Ahh! I know! We can help you create an abstract overnight! An abstract. Get it?”

    And here I guess we are supposed to imagine Lyell giving Darwin a noogie or something as he wonders if Darwin does in fact know what an abstract is and whether he will be up all night helping poor dumb Darwin write it.

    In point of fact, however, in 1844 Darwin had corresponded with Hooker about his idea and by July of that year had written – not an abstract – but a 230 page essay outlining his theory of evolution. As Darwin correctly surmised, it was dangerous stuff and he locked it away with instructions that should he die, it be published with selected research of his. In 1847 Hooker read a copy of the manuscript, and wrote Darwin back, offering helpful feedback but also questioning whether Darwin really was really serious about getting rid of all continuing acts of creation.

    When the Wallace paper arrived and forced the issue, Darwin wrote a paper outlining his own theory AND he helped arrange for Wallace’s paper to be published with his paper in the very same issue of the Proceedings of the Linnaean Society. You might think that was kind of him. Wolfe wants us to know that only sheeple think that way.”

  21. They point to the eye and claim a low probability of it coming about. But they are assuming it came about all at once. Selection theory doesn’t say “all at once”. It proposes a latching mechanism: a small but unharmful (and maybe beneficial) change occurs that can be passed on. Over time, the small change mutates gradually adding more and more capability. For the eye this could mean a novel cell that responds to light becoming groups of cells that respond.

    The subtext of the arguments of everyone who succors this sucker’s hallucination is that there is a teleological component of mutation and natural selection.

    They do it everything…..

    Adios

  22. Teleological?

    You shouldn’t use words you don’t understand or maybe it’s your demonstrated lack of comprehension. Clearly room for improvement.

    Mutation is an event with multiple causes which itself can cause things. I’m unaware of any purpose.

    Don’t blame you for running away though.

Leave a Reply

Your email address will not be published. Required fields are marked *