Posted on 04/07/2020 2:29:52 AM PDT by Oldeconomybuyer
SEATTLE - A statistical model cited by the White House generated a slightly less grim figure Monday for a first wave of deaths from the coronavirus pandemic in the U.S. - a projection designed to help officials plan for the worst, including having enough hospital staff, beds and ventilators.
The only problem with this bit of relatively good news? It’s almost certainly wrong. All models are wrong. Some are just less wrong than others - and those are the ones that public health officials rely on.
Welcome to the grimace-and-bear-it world of modeling.
“The key thing is that you want to know what’s happening in the future,” said NASA top climate modeler Gavin Schmidt. “Absent a time machine you’re going to have to use a model.”
Weather forecasters use models. Climate scientists use them. Supermarkets use them.
As leaders try to get a handle on the coronavirus outbreak, they are turning to numerous mathematical models to help them figure out what might - key word, might - happen next and what they should try to do now to contain and prepare for the spread.
The model updated this week by the University of Washington - the one most often mentioned by U.S. health officials at White House briefings - predicts daily deaths in the U.S. will hit a peak in mid-April then decline through the summer.
Their latest projection shows that anywhere from 49,431 to 136,401 Americans will die in the first wave, which will last into the summer. That’s a huge range of 87,000. But only a few days earlier the same team had a range of nearly 138,000, with 177,866 as the top number of deaths. Officials credit social distancing.
(Excerpt) Read more at apnews.com ...
I’ll tell you for sure what AIN’T speculation. Millions of people are unemployed.
(1) If you know absolutely nothing about a future event, its probability of happening is usually modeled at 50%. The more you know, the more you can say it is more or less likely to happen than the 50% initial guess. Often times there are sophisticated research or data gathering projects when the value of refining the model probability is high. We need a lot more testing data to have good Covid-19 models.
(2) Modeling the future is like driving your car by only looking at your rear view mirrors. Future modeling (whether time series or cross sectional based) all assume that past relationships will hold in the future. Part of the changes in recent Covid-19 models are due to going from Chinese data to more recent US & EU data. Also changes in medical treatment and technology can change the model results.
(3) Be extremely careful when using models that are based on logarithmic projection. While there are some processes that do grow (for a period of time) in a logarithmic fashion, not many can grow that way for very long periods of time. There are lots of example that range from tree growth projecting trees several kilometers high to bunny rabbits breeding to the point that their physical size exceeds the land mass of the world. Some of the first Covid-19 models forecast out 18 months or so into the future, which was utter nonsense.
Now for the modeling being done at the UW Health Metric group. If you go to their website and read what they say. http://www.healthdata.org/ They are very open about the inaccuracies of their models. They also include error bands based on the statistics they have gathered.
Where things have gone crazy is the people who take the models and their projections (especially the early London Imperial College Covid-19 model) as gospel. In my opinion, the best public policy is to try some of the things the models tell you, then see what happens in a couple weeks and then adjust. This way you don't project too far out into the future with your models and you test the models.
I feel that the problem is that the media and politicians want to desperately believe that they are in control and that they can trust the model results.
Models are better than a sharp stick in the eye, but you have to keep some emotional distance from them, which is hard to do in life and death situations.
The AP never admits this about climate models.
Well put. I especially like the analogy of driving forward using the rear view mirror.
What I learned from about a half-dozen stat classes in grad school was how to make the number say what the boss wants it to say.
AP’s supreme gloBULL warming alarmist, Seth Borenstein, has spoken! backed up by gloBULL warming alarmist, NASA’s Gavin Schmidt! an attempt to cover for the fake models:
1 Oct 2019: WattsUpWithThat: Seth Borenstein, Climate Gloom & Doom: “Déjà vu all over again”
by David Middleton
‘We’re all in big trouble’: Climate panel sees a dire future
by Seth Borenstein
NEW YORK (AP) — Earth is in more hot water than ever before, and so are we, an expert United Nations climate panel warned in a grim new report Wednesday...
https://wattsupwiththat.com/2019/10/01/deja-vu-all-over-again/
13 May 2016: WatsUpWithThat: NASA Admits Substantial Climate Uncertainties
by Eric Worrall
NASA has appealed to the Australian CSIRO not to fire lots of climate scientists, because they are doing “important work”, resolving significant uncertainties in climate modelling...
https://wattsupwiththat.com/2016/05/13/nasa-admits-substantial-climate-uncertainties/
7 Jan 2019: WattsUpWithThat: Anthony Watts: New study attempts to “squeeze out” uncertainty in climate models
From the “we’re gonna need a bigger computer” department.
Climate model uncertainties ripe to be squeezed...
https://wattsupwiththat.com/2019/01/07/new-study-attempts-to-squeeze-out-uncertainty-in-climate-models/
30 Oct 2019: WattsUpWithThat: Escape from model land
Reposted from Dr. Judith Curry’s Climate Etc
https://wattsupwiththat.com/2019/10/30/escape-from-model-land/
pity about the lockdowns!
So why didn't the AP do a piece about all models being wrong when the models were predicting 2 million US and 500,000 UK deaths, but only comes out with this sh_tty piece when the models say there will be less deaths?
AP shilling for more Americans deaths and longer economy destroying lock-downs as usual? You bet!
this one will be reliable! not:
6 Apr: Science Daily: Researchers hope to improve future epidemic predictions
New mathematical model uses information theory to improve epidemiological predictions
Source: U.S. Army Research Laboratory
Summary: As the world grapples with the COVID-19 pandemic, a new mathematical model could offer insights on how to improve future epidemic predictions based on how information mutates as it is transmitted from person to person and group to group.
The U.S. Army funded this model, developed by researchers at Carnegie Mellon University and Princeton University, through the Army Research Laboratory’s Army Research Office, both elements of the Combat Capabilities Development Command...
“These evolutionary changes have a huge impact,” said CyLab faculty member Osman Yagan, an associate research professor in Electrical and Computer Engineering at Carnegie Mellon University and corresponding author of the study. “If you don’t consider the potential changes over time, you will be wrong in predicting the number of people that will get sick or the number of people who are exposed to a piece of information.”
In their study, published March 17 in the Proceedings of the National Academy of Sciences, the researchers developed a mathematical model that takes the evolutionary changes of both disease and information into consideration...
“We showed that our theory works over real-world networks,” said the study’s first author, Rashad Eletreby, who was a Carnegie Mellon doctoral candidate when he wrote the paper...
In addition to the Army, the National Science Foundation and the Office of Naval Research also supported this research. Other researchers co-authored the paper include Yong Zhuang and Kathleen Carley from Carnegie Mellon University.
https://www.sciencedaily.com/releases/2020/04/200406110719.htm
Don't you love it when they come up with these world changing, “revolutionary” models? They work fine in the nice leafy environments of the university. Only problem is when there is a real crisis. Then the models collapse.:)
Good post, thank you.
“Modeling” can’t be done with incomplete, erroneous and intentionally skewed data. The Wuhan virus “models” remind me of another model that’s still being used in attempts to destroy the world economy; New York will be under water any day now...
There will be more adjustments and revisions to the models; the subject matter experts have to try to salvage their reputations.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.