A random walk generated from coin tosses would be as reliable as many of the models (see last example, below).
Here is the Latest Projections from December 2021 to Today
December 2021 – projects over 3,000 patients to be hospitalized simultaneously for Omicron, at peak in early February. The state generally has 6,000 +/- staffed beds – and projected a slight majority would be Covid patients.
The actual peak was about the 3rd week of January at about 1,150. The model estimate was completely wrong, as usual.
On February 3, 2022, OHSU’s Peter Graven forecast that Oregon would not reach under 400 hospitalized patients until about March 31st. Oregon announced it would end its useless, mandatory mask mandate, even for triple vaccinated – until March 31st, or when the number fell below 400.
Today, March 4th, OHA reported a total of 375 patients hospitalized for Covid-19.
Even though we hit their own metric, the state will not end it mask mandate until March 12th.
Because Modelstm. Trust Sciencetm.
Models Are Worthless
Throughout the pandemic, disease models were not merely worthless – they were used as cover for public health to cause actual health harm to people.
On March 16th, 2020, Neil Ferguson of Imperial College London released his projection of up to 2.2 million dead in the U.S. by August 2020. In 20 years, Ferguson had never produced an accurate model. His past projections for other diseases were, at times, off by orders of magnitudes. He said we needed to go into “lock downs” to stop his forecast from coming true.
His model also forecast if we did various “NPI” restrictions, cases would remain low. We could then ease restrictions and bring them back only when numbers rose again. In this way, we could, pardon the bull shit, “flatten the curve” permanently until vaccines were available.
How did that work out? It didn’t. We did the NPIs with high compliance and we still had a pandemic lasting over 2 years (typical pandemics last 1-3 years and end on their own), and with case numbers skyrocketing. A month later, Ferguson was caught with his pants down – literally – during lockdown, and just ten days after being diagnosed with Covid, he had sex with his married lover at his flat. He was then thrown off the SAGE committee.
My state relied on model projections from Christopher Murray’s UW IHME random number generation team. Based on their garbage forecast in March of 2020, my state shut all non-life-threatening health care for up to over 3 months (actual duration varied by County). In March, I suffered a foot injury. Not until 13 weeks later was I permitted to get a diagnosis: broken navicular bone, partially torn tendon. Two years later, I experience pain every day, at times. For obvious reasons, I disrespect the arrogant jerks who pretended they could model the disease.
The IHME model’s output was off in space. At the time, they published the model input data – their model input data did not agree with the actual numbers reported by OHA. I spent 3 weeks trying to get IHME to fix the input data. I never received a response. They did not care.
October 2020 to January 31, 2021
OHA acknowledged late in 2020 that the models were garbage. Oregon dropped IHME and later went to IDM models, and then later still, OHSU models.
In late 2020, they went to using only a 3-week model projection. During this time, they produced 13 projections. Their accuracy is shown here – from October 28 to January 31. In 11 of the 13 projections, they had the sign wrong. Tossing a coin would have been more accurate.
The black line shows actual cases, averaged over 2 weeks. The colored lines show the projection at each starting point. This chart was produced by Twitter user “OHA Context & Commentary” (did not appear to be affiliated with OHA).
I will not forget the clods that caused actual harm due to the use of their fraudulent disease models. Disease models were terrible, and the modelers must be held to account for their failures. They can hide behind their spreadsheets in their academic towers – but when their failures are turned into policies that harm people, they must be held to account.
Disease models proved to be so wrong as to be fraudulent. Again, it’s fine for them to play with their spreadsheets while sitting comfortably at their desk. It is not fine when untested, useless models are applied to setting mandatory policies on the people – policies that caused actual harm.
Update: I posted some charts on social media about how bad the forecasts were and received a response that “Of course! It’s a new respiratory disease and the models won’t be accurate!” as if that justifies the use of models that are little better than coin tosses! That’s like saying, “Yes, we know the disease models are really bad, but since we have them, let’s use them anyway!” – even though we know they are wrong and we could just as well set policy by tossing a coin. But tossing a coin does not sound Sciency – better to create a model that is the same as a random coin toss!
Related: What has happened to Doctor Science? Since mid-February or earlier, Dr. Fauci has literally vanished from media appearances. This is the guy who was on TV non-stop, nearly every day. The speculation is the polls have turned so negative for the party in power, the polls are so bad for Dr. Fauci (I admit, I cannot trust him), that has been disappeared.