Chapter 4 in the R.M. Dolin book, “Truth and Trust in Crisis.” 2021
At the beginning of April 2020, the federal government predicts that in fourteen days the COVID crisis reaches its apex with respect to rate of expansion and begins to decline. According to academic model projections, by the time the apex is reached, 250,000 Americans will have died from COVID. Their assessment of our collective fate is so dire, we decide to better comprehend what’s happening by generating an exponential model to verify if it’s even possible for a natural virus to behave with such virulence and to our dismay discover it is.
We then reviewed data from Italy’s COVID crisis and built a statistical model that indicates it’s more likely 30,500 Americans will die from COVID by Armageddon day. So, here we are on April 3rd, eleven days out, trying to sort out who’s more likely to be right, us or the people we trust with our fate. Given the disparity in predictions, one of us is going to be very wrong, but if our model prevails, it opens a Pandora’s Box of questions that’ll take historians and forensic investigators years to unravel.
Our increasing skepticism in “expert” versions of reality is now coupled with an increasing sense that flawed COVID information is intentionally being generated as a false flag to enact mitigation measures, many of which defy pragmatic logic. While confident that deep in the bowels of the CDC and the White House Coronavirus Task Force good statisticians diligently generate quality information that is well vetted, I’m less certain the COVID experts appreciate the caveats those statisticians must be providing, because at this point it does not seem that a rational person could believe 250,000 Americans are going to die from COVID in the next eleven days. If government policy makers truly believed their predictions, why are crisis response teams not being mobilization across the country and why are death industries not being told to ramp up capabilities and production?
One aspect of the previously discussed Dolin Paradox is that numerical models depend on input parameters to generate output, and those parameters can be tweaked to provide a wide range of results. Based on the growing disparity between “expert” assessments and what the data indicates, I suspect the Brix/Fauci duo don’t understand the complex mathematics underlying the models they rely on, or how sensitive those models are to slight variations in input. I’m less certain they understand a key fundamental tenant of mathematical modeling, namely “Garbage in equals garbage out.”
Consider our simple exponential model that demonstrates to get from the current 3,000 cumulative COVID deaths on April 1st, to 250,000 cumulative deaths by April 14th, the function f(x)=ax, requires a=2.4052, and x=14. However, just as easily we can run this model with a different value for a, say one that’s 11% lower. If parameter a=2.165, the output projection for the number of COVID deaths by April 14th decreases to 61,351
This means an 11% decrease in parameter a, result in a 400% decrease in the output estimate, which demonstrates how sensitive mathematical models can be to slight variations in input. We could just as easily increase the input value of our base variable 11% to a=2.7, which increases the output estimate to 1,101,624. So, making an 11% increase in parameter a, result in a 440% increase in the projected number of COVID deaths.
This demonstrates why I began each semester of my Numerical Methods class telling students a mathematical model always provides an answer, and that the answer is always wrong. It’s wrong because variables like a, cannot be known with certainty, and uncertain input leads to uncertain output. The true value of a cannot be known, only approximated. While we can use past data to infer a future value for a, any such inference is fraught with uncertainty.
There are techniques for minimizing uncertainty. For example, we can perform random scenario simulations using Monte Carlo techniques to run a million scenarios with different values for a bounded by upper and lower limits, to find the most likely value. But even then, whatever value is inferred, has uncertainty. This is a long-winded way of saying that somewhere between a=2.165, and a=2.7, the Brix/Fauci duo decided the “best” value for a, is 2.4052, relative to their model parameters. The question remains, however, why would they choose a number so unrealistic, is it perhaps based on a combination of experience and COVID clairvoyance or are they just that out of touch with how nature behaves, or do they know something about the virus they’re not sharing with the rest of the world.
Given the estimate from our statistical inference model that 30,500 Americans will die from COVID by April 14th, the corresponding value for a in our four-line exponential model is a=2.041. Where science differs from politics is we accept parameter a can never be known with certainty and continually strive to refine it. We utilize existing data to calculate values of a at prior times and apply that knowledge to infer likely values for a going forward. The political constraints of the world where the Brix/Fauci duo operate prevent them from adjusting their initial assertion to match reality for fear of looking incompetent, which ironically, is why they appear incompetent.
April 5, Nine Days to Armageddon: Figure 4.1 shows that yesterday there was an 11% day-to-day increase in the number of infected American’s who died from COVID, which is far below the 243% increase estimated by the government predictions and academic projections. Today’s data indicates that the day-to-day rate of change is 18%, which while higher than yesterday, is still below what’s required to reach 250,000 deaths by April 14th.
We can utilize our exponential model with a slight variation to calculate a running value for a. For this we need a new value for variable x, based the number of days since the first COVID death, which occurred on February 29, or 36 days ago. Today’s cumulative death count is reportedly 7,146. What remains is determining a value for a that satisfies our governing equation.
Figure 4.1. WHO and CDC Reported COVID Data Through April 4th, 2020.
Solving our exponential equation for a, we get that when a=1.224, the area under an exponential curve matching CDC death data equals 7,153. Recall that to compute the area under the curve we integrate our exponential function, making our revised mathematical model
April 7, One Week to Armageddon: We have now reached the midway point in our march toward Armageddon. While the current CDC reported death count is nowhere near government predictions and academic projections, an interesting development has emerged. Rather than revise their assertions downward based on current data as any rational scientist would do, the so-called COVID experts are doubling down on their claims by deploying a cleaver sleight-of-hand. The White House, along with state governors who enacted draconian mitigation measures, are asserting that because of their emergency measures the COVID crisis is being brought under control.
This is an easy assertion to accept because it’s virtually impossible to dispute, however, with time comes the wisdom of clairvoyance and in 2022, a Johns Hopkins study will conclude that COVID mitigation measures had little impact on stopping or slowing the virus spread.[1] This means the new government sleight-of-hand tactic provides a pseudo “get out of jail free” card for making fantastically inaccurate predictions while portraying themselves as heroes for saving the nation by wrestling the COVID beast into full capitulation. Only now, thanks to Johns Hopkins, we know it’s a false flag.
Meanwhile, for the past two weeks we’ve been using data published by the WHO and CDC to chart not only how the COVID crisis is behaving, but also to understand the projections being made by federal and academic models. Utilizing our simple exponential formula, we’ve demonstrated that the academic model projections are increasingly losing alignment, not only with reality but with actual data.
Years ago, I was given the opportunity to enroll in an intensive two-year Six Sigma academy at Bechtel, the world’s largest engineering firm. They taught us that to be a successful change agent you must, “speak truth to power,” and the best way to do that is by trusting your data analyzes wherever they take you. As Americans are increasingly inundated with sensational claims about the number of COVID deaths projected to occur by the mystical apex day of April 14th, our analysis led to an estimate that is not only significantly lower but appears to be far more realistic.
We acknowledge our model relies on a simple mathematical construct, depends on reported CDC data, likely lacks the sophistication of academic models, and is not privy to the kind of insider clairvoyance the Brix/Fauci duo have access to. We concede academic projections and federal predictions, should be more accurate with higher fidelity and precision than our four-line program, but data is as data does and reality is an unyielding taskmaster. Bottomline is that in science, data demands an outcome that matches empirical evidence; an outcome that cannot be massaged to meet a prescribed narrative.
April 12, Two Days to Armageddon: Today we consider CDC data from a fresh perspective to see if we can either validate federal predictions and academic projections or further confirm our assertions. To those who’ll review this analysis with future derision, our goal is to keep the discussion at a high school algebra level so forgive me for not providing Calc-III detail. Thus far, the entirety of our mathematical model is based on the basic exponential function,
Utilizing CDC data, we find the value for parameter a best representing the data to date. Today x=44, since it has been 44 days since the first reported COVID death, and the cumulative death total is 20,608. For these values, a=1.2065, or
Notice that as we approach Armageddon, the value for parameter a decreases, which means the slope of the exponential curve decreases. This is an indication we’re approaching the curve’s apex. That parameter a has steadily declined is a remarkable outcome given all the hysteria being promulgated by the media. The bottom line is that COVID appears to be trending opposite of Brix/Fauci duo predictions and academic projections. Meanwhile, CDC data seems to be validating our simple statistical model generated weeks ago predicting 30,500 COVID deaths by April 14th.
Further verification that the downward trend in parameter a, is correct, UW today revised their model projection downward to 80,000. Why they readjust to a lower but still remarkably unlikely projection remains a mystery but it’s becoming increasingly obvious that empirically grounding their projections in reality does not involve a review of available evidence.
Meanwhile, the doomsday scenario promoted by the Brix/Fauci duo has yet to be revised downward, even though a smattering of official backtracking suggests they know it should be. The challenge for Brix and Fauci is how do they bring themselves to admit the fallacy of their sensationalized assertion without losing credibility. Their rhetoric continues to extol their success by enacting mitigation measures, even though we now know from the Johns Hopkins study that mitigation measures have little effect.
You and I appear to have the only empirically consistent COVID model, yet we must remain vigilant to potential up-ticks. If the world is being besieged by a naturally occurring virus, its ability to dynamically explode is omnipresent, even though current evidence suggests otherwise. While our model relies on past COVID performance to make future inferences, it is likely the federal and academic models rely more heavily on undisclosed clairvoyance. However, as any Wall Street expert or sports bookie can attest, people routinely loose fortunes applying undisclosed clairvoyance to predict future outcomes. As the philosopher Aldous Huxley observed, “facts do not cease to exist because they are ignored.”
What’s difficult to understand is that in the face of mounting evidence, government “experts” and the media continue hyping their doomsday scenario where the COVID death count catapults to levels never experienced in previous pandemics. What’s missing in their hyperbole, is any actual evidence that a dystopian Armageddon is about to darken the nation, but that of course does not tap down the parade of talking heads who emphatically know how deep the darkens gets.
This leads us to Brix/Fauci, their media cohorts, and complicit academics. Using our statistical model coupled with CDC data, it’s becoming painfully clear Americans are being bamboozled by those choosing to scare an already stressed public with a wildly unfounded claim that a quarter of a million Americans will die from COVID by April 14th. Meanwhile, you and I, seeking scientifically grounded perspective, develop a simple statistical model using Italy as a sample population to conclude the COVID death count by April 14th is likely closer to 30,500. While this projection is unsatisfyingly high and demands mitigation, for the government to error by a factor of eight is unconscionable; especially given the widespread public panic it causes.
To rescue credibility, Brix/Fauci need more compliant CDC data and to that end, have directed a complicit medical community to label virtually all deaths in America as caused by COVID. An example given by a Minnesota doctor is that if someone is hit by a bus, the cause of death is ruled COVID. When I taught Measurements Lab, we used the expression “dry-lab” for the act of fabricating data to facilitate a desired outcome. In the scientific community, what Brix and Fauci are doing by skewing COVID death data is tantamount to professional malfeasance, or to use a term they might remember from long-ago when they practiced medicine, malpractice. While we still trust CDC reporting, that trust is diminishing as the CDC has a history being politicized.
Entertainers and social media influencers have yet to realize the extent to which they are being manipulated, and it matters because in the digital age their prominence exceeds their capacity to grasp the detrimental impact they can cause, and it doesn’t take much to put an already panicked public in full tilt.
In times of crisis people rely on news outlets for information. If you’re over fifty you probably remember the trust Walter Cronkite’s voice provided every evening as he reported with stoic seriousness on U.S. troop casualties in Viet Nam. Unfortunately, the trustworthiness of journalism has diminished to the point that news relies less on facts and more on the outrageous opinions of people disguised as experts. In a time when our nation desperately needs truth tellers, clear information vetting, and concise unbiased reporting, we are disappointingly denied. Michigan State Professor Jeff Richards[2] opined best by stating, “There is a huge difference between journalism and advertising. Journalism aspires to truth. Advertising is regulated for truth. I’ll put the accuracy of the average ad in this country up against the average news story any time.”
The inconvenient truth that has been hiding in plain sight since this crisis began is that it’s not hard to ferret through WHO and CDC data, at least for now, to make realistic predictions and projections using simple math. However, we appear to be in the early stages of what’s becoming a deliberate effort by medical professionals to intentionally corrupt CDC data and push an unethical narrative driven by greed. Armed with our simple statistical model and exponential tool, we’ve demonstrated that government predictions and academic projections are ensconced in hyperbole that the media has chosen to sensationalize rather than investigate.
Like a horrific trainwreck we’re compelled to watch, we are witnessing government, academia, and the media scrabble to cover their incompetence even though their motivation remains elusive. What makes this all so fascinating is that we must assume the collaborative skewing of the truth to match a narrative is not new given how quickly and seamlessly they coalesced. What historians will ultimately conclude that makes COVID a Waterloo moment, is that in the digital age science can be used to separate opinion-based hyperbole from data driven truth.
Brix/Fauci maintain that their prediction only seems outrageous because their mitigation measures are working. Keep in mind though that this assertion is later repudiated. The obvious question becomes, if government and academic claims that their efforts to “flatten the curve,” are the reason their forecasts are off by a factor of eight, why don’t their sophisticated models capture the effects of their mitigation measures to provide more reasonable forecasts? After all, our simple statistical model managed to closely match the actual Armageddon death count without caveating for mitigation measures. Given all the intellectual horsepower behind these forecasts, why were we able to accurately model the inevitable COVID outcomes with a simple statistical assessment while their high-tech software running on some of the world’s most powerful supercomputers can’t? As Shakespeare[3] might conclude, “something’s rotten in the state of COVID.”
Suggesting the impact of mitigations cannot be modeled is a false flag, government and academia have been modeling influenza outbreaks for years and they know how mitigations work. Equally poignant is that the WHO/CDC data their models rely on have the effects of mitigation embedded in them, which is why our model captures the impacts. To hear Brix/Fauci attempt to reconcile the difference between their dire predictions and reality as evidence to their effective leadership is a poorly conceived effort to rescue credibility.
This is not to suggest COVID’s not a real crisis or a serious threat to our nation and world. However, when a ship’s taking on water, it’s the responsibility of leaders to correctly triage and properly react. There’s a huge difference between a leak that needs to be managed and a gaping hole in the side of the stern that’s taking on water so fast the ship can’t be saved. In times of crisis, we don’t need leaders shouting in panic to abandon ship when the damage is manageable. Government, academia, media, and medical leaders have decided to go all in on over-sensationalizing the COVID crisis when they have a societal obligation to accurately assess data and make realistic projections, something our four-line program is eminently capable of doing yet eludes the so called “experts.”
April 13th, The Day Before Armageddon: Today, CDC director, Dr. Robert Redfield, announces the virus has stabilized, while cautioning that it’s still expanding, and the apex has not yet been reached. Let’s consider Dr. Redfield’s statement to better understand what he means when he says the virus has stabilized but is expanding.
If the virus is expanding, then the rate of growth for infections and deaths should be increasing. In terms of our simple exponential formula, f(x)=ax, if the virus is expanding, parameter a increases over time. The apex that Dr. Redfield referred to occurs when the rate of change for the virus stops increasing. In absolute terms the apex occurs at the highest point in the rate of expansion after which, contraction begins. The apex can be over a period where the infection/death rate is nearly constant or has small fluctuations. Cynics will be tempted to point out that Dr. Redfield’s declaration seems properly on script as government experts predicted the apex would occur tomorrow.
If COVID follows a “bell-shaped” curve, then the rate of contraction should closely mirror the rate of expansion. In other words, if the number of people in the U.S. who die from COVID ten days before the apex is, say 20,000, then the number of people who die ten days after the apex should be ~20,000. Many things in nature follow bell-shaped curves, but it’s yet to be determined if COVID originated in nature.
Reaching the apex does not mean the crisis is over, it means it’s no longer getting worse. Also, because it took 46 days to reach the apex does not mean it takes 46 days for the crisis to end, however, we’d expect it to happen in a somewhat similar timeframe. It’s also possible, if we’re not vigilant, expansion could restart; something I keep reminding my kids who are increasingly finding ingenious ways to bring normalcy back into their lives.
Tomorrow’s anticipated apex is an important milestone because it signals the worst is behind us and there is an offramp to government’s draconian mitigation measures. Dr. Redfield’s assertions align with our exponential model, which has not only stabilized, but begun to contract.
Some will say both our statistical model used to predict the apex death count and our exponential formula used to model COVID’s behavior are too simplistic, while the more sophisticated federal and academic models consider much more information. Yes, their models are substantially more complicated but in science, that does not imply better outcomes. There’s an important axiom in science called, Occam’s Razor[4], which states “Given a set of explanations for an event, the simplest one is most likely correct.”
Our simple statistical model and exponential formula is what mathematicians refer to as a “first principles,” formulation, which means they rely on basic equations with minimal assumptions or reliance on extraneous data. As validation of Occam’s Razor applied to our first principles formulation, on April 2nd we predicted that the April 14th COVID death count would be 30,500, this has a forecasting error[5] of 30%, while the federal predictions and academic projections have forecasting errors of 964% (determined using a Mean Absolute Percent method of forecast error analysis).
An equally salient point as it pertains to Occam’s Razor is that our first principles formulation relies exclusively on CDC data while federal and academic models rely on that same data plus a plethora of assumptions and extraneous information not at our disposal. The rub, however, is that CDC data has all the information needed about COVID’s behavior embedded in it. In that sense, our formulation has all the relevant information needed to make a credible assessment. I get a sense federal and academic “experts” don’t understand this.
CDC Director Redfield’s statements today are wrong. As we have marched toward Armageddon, the values for parameter a, which we revise daily based on CDC data, show a steady decline since March 24th. In other words, the virus’s rate of expansion has been slowing. I anticipate that tomorrow, federal and academic experts will tout their modeling prowess by proclaiming the apex has been reached, and that the virus is now in decline.
Figure 4.2. WHO and CDC Data as of April 14th, 2020.
Don’t get caught up in the day-to-day variability of the charts. The WHO/CDC have an enormous task of compiling daily data and if one day’s data arrives late or during the weekend, it gets rolled into the next day’s totals. What’s important are the general trends. Also, I am presenting the data in raw form. We could have used regression to smooth out noise, but raw data is less complicated.
The two plots on the left represent the global and U.S. daily infection rates. Notice they generally mirror each other. While the global chart seems to be bimodal (having two peaks), it’s probably more of a plateau with reporting inconsistencies. The important take away from these charts is that the COVID infection rates seemed to have reached their apex around April 10th, however, that both the global and national apexes occurred on the same date is likely just coincidence. From April 10th forward, the infection rates generally decline.
Federal and academic models predicted the apex would be on April 14th, so it appears they missed it by 4 days, which isn’t bad given the magnitude and direness of their predictions. It’s prudent to be reticent about claiming too early that the apex has been reached because what appears to be an apex could later prove to be a momentary dip. Our model confirms what the CDC director reported, namely that there is ample evidence to declare the COVID apex has been reached, and we can all breathe a sigh of relief that infections and deaths rates are in decline.
The two charts on the right, represent the global and U.S. daily death totals. Notice they also generally mirror each other. This can in some sense be attributed to the U.S. currently experiencing the highest death rates in the world, which impacts the global plot even though the U.S. only represents 5% of the world population. This suggests that the U.S. COVID experience mirrors the rest of the world, which is a significant outcome as it justifies our statistical model using Italy as a sample population.
Notice that both the U.S. and global death tallies peak on April 10th. Logically we expect there to be a lag between infection and death rates, however, the data does not bear that out. If there was a lag, the infection apex would occur before the death apex, and they don’t. The steep decline in the global death tallies post apex before a sharp rise probably has more to do with reporting lag. In general, both the U.S. and global death rates seem to be on the decline and while trends can always shift sharply, this is good news for a country and world in worried panic.
Federal predictions and academic projections assert that by April 14th, the U.S. will realize 250,000 deaths, when in fact 23,650 deaths have occurred thus far. This means they were off by a factor of 10, which renders their model’s ability to make projections somewhat akin to blindly throwing darts at a wall. Our statistical model projected the nation would realize 30,500 COVID deaths by April 14th, so we were off by a factor of 1.29, which isn’t bad for a bunch of lock-down detainees using a four-line spreadsheet program and high school math.
The unresolved questions going forward are,
- Do governments acknowledge COVID’s in decline and reopen economies while easing mitigations?
- Will proper postmortems be conducted by trusted experts to get at the truth behind the crisis?
- Can we trust CDC data and if not, is there a way to determine the extent their data’s corrupted?
- Do academics who abdicated their responsibilities have sufficient integrity to step out of shadows to conduct necessary peer reviews?
While some of these issues are beyond the scope of our investigation, the issue of trust in CDC data takes on increased prominence going forward, causing us to utilize unconventional avenues to find alternative sources of truth as we continue to explore how the real science behind the COVID crisis.
[1] https://www.newsweek.com/did-johns-hopkins-study-prove-lockdowns-dont-work-what-we-know-so-far-1676724
[2] https://theysaidso.com/quote/jef-i-richards-there-is-a-huge-difference-between-journalism-and-advertising-jou,
[3] https://nosweatshakespeare.com/quotes/famous/something-rotten-state-denmark/
[4] https://conceptually.org/concepts/occams-razor
[5] https://www.eazystock.com/blog/calculating-forecast-accuracy-forecast-error/
