“Reply Hazy, Try Again,” Plus the End of the Tom Hanks Death Watch
March 17th, 2020Update on my Amateur Coronavirus Predictions
My second coronavirus death toll prediction model is not doing very well. My total infection prediction model seems to be doing quite well indeed. While it should amaze anyone if either model works, I figured the death model would work better, based on the higher reliability of death numbers, but that hasn’t panned out.
Several days ago, I came up with a differential equation to predict the total number of infections that would occur. It says the total for today (morning, more or less) should be about 178,000. The actual total is 185,000. I call that a home run! Much better than I expected. Of course, the total may diverge wildly from my prediction in the future. It would be remarkable if I even came close.
Later on, I created an equation based on the death numbers, which I thought would be more accurate. The death total for today…well, I don’t know what it was, but it doesn’t matter, because it looks like we will pass the number for March 22 tonight or tomorrow.
So what does this mean?
1. Given my total lack of training in the area of epidemiology, combined with the lack of reliability of the data, maybe there was no way either prediction could work, and the good results we’re seeing from the A equation (total infections) are illusory. Maybe it will turn out to be a bad predictor, just like the D equation (total deaths).
2. Maybe the death figures are actually less reliable for prediction purposes than the total figures, because the death rate varies wildly depending on where the virus is.
These guesses top my list.
I suppose that with a very minor epidemic like this one, where the total numbers are very small (sorry, but it’s true), you end up with a situation in which death rates vary because the tiny infected populations in various countries have characteristics that make them much more or less likely to die.
Example which I just made up: what if the virus hits a big community of retirees in Thailand, and because they’re old, 10% die? In a tiny infection pool of 185,000, that would mess up the death data.
The Italians have a crazy death rate. Maybe they’re not detecting mild cases. That seems to be the only thing that could explain it, since the people in countries around them are doing better.
Scandinavians and Germans have low death rates. Given their rigidity and willingness to follow orders, maybe they’re getting tested more thoroughly.
I wonder if I could go back and apply my model to the flu. I don’t think I could get figures. I would have to apply it early in the season, because it doesn’t take saturation into account.
Do I still think the hysteria is utterly unrealistic? Of course. Don’t forget: we get something like 8 million flu cases per day during a global season, and right now we have 185,000 global cases for coronavirus after several months. There is just no way an honest, rational person can look at those numbers and think COVID-19 is worth the trouble we’re putting ourselves to.
I’m very blessed to live where I do. My poor cousin lives in the Chicago area, where the population density is high and the politicians are corrupt leftist authoritarians. She said the authorities are threatening to tell people they can’t leave their houses except to buy food and so on. Consider this: yesterday, the Chicago Tribune said there were 105 detected cases in Illinois.
I looked at the website for the local paper, and nothing like that is happening here. Events are closing, and schools are shut down, but if they are threatening to imprison people in their houses, the mighty Star-Banner is not saying so online. I would assume the would put that information on their site if it were in the paper.
Unless the people in the press room are just too weak to work their keyboards.
The terrible thing is that we are going to have the economic destruction of a plague, without a plague. Businesses are going to close permanently, all over the place. Stocks are in the toilet. All sorts of jobs will be lost. Jobs are important. Businesses are important. It’s not okay to throw them away over a disease that will probably kill fewer than 20,000 Americans.
We should be focusing on insulating old people and sick people. That’s it. Kids are nearly immune, and the rest of us are very, very unlikely to die even if we get infected. In other words, it’s a lot like the flu, except it affects way fewer people.
If you think 20,000 people is a big hit, you’re bad at math. You don’t understand how many people die here every year. Go look up the number of traffic deaths we suffer in one year. Look at the number of suicides. Look at tobacco deaths. There are over 300,000,000 people here, and you have to take that into account.
Doesn’t every life have infinite value? No. Never did. If a life had infinite value, we wouldn’t build skyscrapers or bridges. Every huge construction project costs lives. We would quarantine people who get colds and norovirus, because these diseases kill a small number of people. We don’t do that. The world has to keep turning. The value of keeping our economic system going, sorry to tell you, is very, very high, and before coronavirus, we accepted the fact that keeping America bustling would cost many thousands of lives every year. We just didn’t talk about it.
If lives had infinite value, we would drive $500,000 cars that provided 100% protection in crashes. We would force kids to be vaccinated. We would ban tobacco and scuba diving. We would build every house from reinforced concrete. We don’t do those things. An acceptable risk of death is something healthy societies accept. The alternative is economic paralysis. And we can’t really control death, anyway.
It’s okay to say money is important. It’s not something to be ashamed of. Money is housing. Money is clothing. Money is medical care for kids with cancer. For the leftists, let me say that money is tarot cards, dope, Molotov cocktail ingredients, and black ski masks. It’s Greenpeace donations.
Guess how many people die in New York City every year. The figure is over 154,000. That’s one city. One year. How many New Yorks can you fit in 7 billion? The global annual death figure is around 60 million, and many, many of those deaths are not from old age.
I just looked up the world’s population. I thought it was 7 billion, but I’m seeing 7.7. Is that really true? Man, there are a lot of us.
We are killing our economy over nothing. When the panic goes away, we will have a real problem: increased poverty.
Here is the total number of cases of coronavirus, expressed as a percentage of the world’s population: 0.0024%. Here is the total number of cases of Spanish flu, expressed as a percentage of the world’s population in 1981: 27%. Here is the number of seasonal flu cases in the US, expressed as a percentage of the nation’s population: 10%.
Sure, the coronavirus total will go up, but if it were going to get anywhere near flu numbers, which don’t upset us, why is it so far behind after several months? The flu gets the job done in 100 days.
There is really no point in my writing these things, except to vent and to commiserate with other people who understand the situation. People are not going to listen to common sense in large numbers. They’re under a mass delusion with a supernatural cause. This is what happens when people don’t get baptized with the Holy Spirit.
I have to go to the grocery store. I can’t believe it. I have to go look for purified water. I’m not buying it because of the virus. I buy it because I’ve had kidney stones. Will I find any water on the shelves? I wonder.
I would like to get sanitizing wipes, because I use them when I go to the dump, and I’m running out. Not sure I’ll be able to buy them.
Last night I wondered if I should buy a sack of rice and some dried beans. Then I snapped out of it. Food isn’t going away. Farmers aren’t going to let their crops rot. People are filling their houses with Pop-Tarts, potato chips, and other boxed food, and when the panic tapers off, they’ll be stuck at home, eating junk. Maybe I’ll have the store to myself for a while.
Time to leave the house. I plan to cover the entire car with toilet paper rolls, with a tiny hole I can look through to drive. This should give me protection from the juju.
God help people who live in cities. I am praying he will wake them up and help them move.
Addendum: in case anyone cares, Tom Hanks and his wife are out of quarantine after a lengthy one-week illness. Here’s the list of terrible symptoms they had: mild fever, chills, and some aches. I have to confess; when I was a kid, I would have chosen that over going to school. Hanks is over 60, and so is his wife. Was he lucky? Did Australian medicine, which was developed mostly on sheep, work a miracle? NO. Their cases are typical, even for old people.
I’m off to brave the panic. If I don’t make it back, my toilet paper stash is up for grabs.
March 17th, 2020 at 5:10 PM
I need distilled water for a medical treatment. I went out in search of it first thing Monday. The first three stores were empty. The forth had lots of distilled water, but no other. The bin label for the distilled read $2.68. I cussed about price gouging but took two gallons. They scanned at 80-cents each, so clearly the bin label was wrong. I think this proves that there’s above which people won’t hoard.
March 18th, 2020 at 8:09 AM
The other possibility is that you’re using the wrong model. In your original post you say you’re repurposing one used for bacterial growth on the assumption that humans will behave like bacteria. However, bacterial growth occurs through binary fission and therefore tops out at the square of the population. Diseases have no such limits, a person could potentially infect hundreds during a single day.
Being off by 7000 isn’t bad. It could certainly be due to underlying noise rather than poor modeling. However, we hit 200k yesterday evening and will likely blow through your Mar. 21st prediction later today. That seems more indicative of the latter.
March 18th, 2020 at 8:32 AM
I thonk the equation should work for bacteria whether they have one offspring or many. The constant would just be bigger.
I’m not saying the model is right, though. It was just a guess. I should probably go back and actually look at a textbook! Maybe I used the wrong equation for bacteria.
March 18th, 2020 at 10:28 AM
You may find this useful: https://mysite.science.uottawa.ca/rsmith43/MAT4996/Epidemic.pdf
One of the tricks to modeling this vs. ordinary flu is how you handle susceptibility. The “novel” in “novel coronavirus” refers to the fact that it’s significantly different from other known viruses. This novelty makes pre-existing immunity much less likely. Susceptibility to seasonal flu, on the other hand, is more heterogeneous. A few strains tend to dominate each year and a certain portion of the population usually has resistance (because they got sick with a similar strain at some point in the past and built antibodies). This actually tends to give older folks a slight advantage that can counterbalance their overall weaker immune systems.
The Spanish flu is a good illustration of this. Mortality was much higher in the young (to whom it was “novel”) whereas older adults likely had partially effective defenses due to previous encounters with related strains (such as the 1880 epidemic).
So when modeling COVID-19 it’s probably okay to assume that each “encounter” has an equal chance of becoming an infection. For seasonal flu you need to account for the fact that a certain unknown proportion of encounters go nowhere due to underlying resistance (from previous infections and/or the semi-rare occasions when the vaccine blends “get it right” for that year)
March 18th, 2020 at 11:41 AM
I would expect very different input from someone who has studied differential equations. Not trying to be condescending, given my status as a rusty former physics grad student who has done absolutely no research into methods, but am I right?
Far as I can tell, the factors you mention should be swallowed up in the exponent, given halfway decent data.