The World-O-Meter Thread

[Elixir]

5,000 sounds too much, 3,000 would be more linear. The trouble is, people are going to suffer because they aren't going to go to the hospital soon. I feel like we are playing Wolfenstein, and going through a gauntlet to get to the final target (get the vaccine approved) where the game just stops and you go to the next level (presumption of immediate distribution of said vaccine). We are taking serious damage, but we will get there with about 10% life left. Victory!


Another way to look at it, about a total of 3 million people (according to antibody study) got it in the initial outbreak in NY/NJ/MA/CT, we had 7 day averages peak at 2,100 or so people dying a day. We currently have 4 million active cases.

Yeah, I don't expect the 5k/day rate to maintain for long. I guess that would be the "good" news. But it will almost certainly top out near that level, given the limited capacity of the health care system to cope with the kind of load that 20,000+ cases/day will represent. The mortality rate is almost certainly going to rise significantly while the case load exceeds capacity.

 

[Derec]

You are better than that at math, Derec.

I don't think there is a problem with my math.

At 10m cases it's around 3% of the population.

About 3.3%, but yeah, to one sig fig 3% is right when it coms to confirmed cases.

So you are assuming that the actual number of infected is 8-15 times what is reported. Possible, but unevidenced.

I think overall it's probably more like ~5x, with early 1a wave that hit NY and North East in general having a greater share of hidden cases, probably more like 8-10.
Take NY. Official cases are still only 3.0% of the population, but an antibody survey in April found almost 14% of New York state residents were infected by then.

(To have any significant effect according to epidemiologists, it needs to be over 20%. I don't pretend to know all the formulae, but I know that such formulae exist and are grounded in fact)

A 5x multiplier would put US at an average rate of ~16.5%. That would mean many regions in the country are at >20%, and others are getting there.

Further, the amount and duration of immunity conferred by getting infected and recovering is yet to be understood. So I consider that line of reasoning very shaky.

That is true, but the virus has been going for almost a year now, and we still have not seen widespread reinfection. That is a hopeful sign.

Rallies and election-related activities have shown to be effective at spreading the virus, no question. But now we have holidays... and a lot of really stupid people who think that a public announcement of a vaccine trial that was 90% effective constitutes permission to ignore the fact that there is a raging pandemic.

That's idiotic. A vaccine is 0% effective before you take it, and the Prizer vaccine is not yet approved, much less available to the public.

I agree that we are probably near peak. I just don't see a rapid decline from the peak happening soon, vaccine or no vaccine.

The down-slope will probably be comparable to the up-slope. Maybe a bit slower, but probably not so much as to take until inauguration day for new cases (7d average) to drop below 80k.

In the short term I still believe saturation testing is the answer. If we know exactly who has it, we can isolate and control the whole thing in a matter of 8 -10 weeks. If antibody tests are ubiquitous we will also quickly learn about conferred immunity - it's efficacy and duration. I'm hoping testing turns out to be the low-hanging fruit for the next administration.

I agree with that. We need to test many more people so asymptomatic/mild spreaders can be identified.

 

[Derec]

Antibody testing indicated that 1/5 to 1/4 of New Yorkers got it. So, this "saturation" bullshit needs to stop. How many times do we need to relive history in the past year?

And their rates are much, much lower than cities such as Chicago that did not have the large Spring wave.
Thanks for proving my point.

 

[Derec]

Merica 330,000,000 citizens. Latest Covid cases day 140,000. Hmmmm.

At that rate in 10 days 1.4 million cases accumulate. In 100 days 14 million cases accumulate. In 1000 days 140 million cases accumulate. So in about 2500 days 330 million cases accumulate. But in seven years US population will be 28 million persons larger.

First of all, the known cases are a fraction of all cases. Add a factor or 3-4x to all new cases.
Second, the infection rates will not be constant. The more people get infected, the more are no longer susceptible and you get something resembling the logistic curve.

Does anyone other than Derec see that, even with long immunity, herd immunity will save us?

It will definitely help us. Not as much as proper strategies would have, but it's the cards we've been dealt because of Trump.

Anyway, infection rates drop as "effective R" drops due to fewer susceptible people. And since some people (those working with the public and those who like to go out despite the pandemic) are much more likely to get infected than others, you do not need to infect nearly as big a fraction of the population than if everybody was equally likely to get infected. That's why you can see slowdown's in tranmission rates when 20-25% of the population have been infected.

 

[Derec]

Those numbers imply a linear increase not the exponential rise that uncontrolled pandemics would follow.

An uncontrolled pandemic with a constant R₀ would follow the logistic curve. That curve resembles an exponential function early on, a linear function in the middle (maximum rate of change, aka new daily cases) and a sort of inverted exponential function in the late stages.

Red: logistic curve
Blue: derivative (rate of change, aka slope) of the logistic function
Green: early asymptotic (exponential)
Purple: middle approximation (linear)
Orange: late asymptotic (exponential)

 

[Derec]

By Thanksgiving, if the 4% daily increase continues, it'll be 225,000 new cases a day. At some point, the states are going to have to close down again, as the hospitals won't be able to handle this. The widespread nature of this third outbreak is filling up areas where there aren't lots of hospitals.

Yeah, 1c has hit rural areas hard. It has also since spilled into Midwestern urban centers like Milwaukee (103/100,000/d) and Chicago (Cook County: 88/100,000/d).

Man, the GOP will have a party over that, as McConnell says the US can't afford to fund states closing down 'needlessly' and refuses a second Covid-19 bill. McConnell is going to try to sabotage the United States for 2024... assuming Biden becomes President.

He has been quite the obstructionist.

 

[Shadowy Man]

Those numbers imply a linear increase not the exponential rise that uncontrolled pandemics would follow.

An uncontrolled pandemic with a constant R₀ would follow the logistic curve. That curve resembles an exponential function early on, a linear function in the middle (maximum rate of change, aka new daily cases) and a sort of inverted exponential function in the late stages.
View attachment 30314
Red: logistic curve
Blue: derivative (rate of change, aka slope) of the logistic function
Green: early asymptotic (exponential)
Purple: middle approximation (linear)
Orange: late asymptotic (exponential)

Yes. Hence why I used “uncontrolled”. But you are right that it will slow and be more linear across the middle.just don’t know where the middle exactly is yet.

 

[Derec]

Looking like it might top out between 200,000 and 250,000 new cases per day, and around 5 thousand deaths per day at the peak.

Are you talking about 7 day averages or raw daily numbers? You are right about raw numbers, but we should look at 7 day averages and I think those will peak below the 200k/d mark.
5k/d deaths is not likely at all because deaths are just above 1k/d. So you'd need a 4x increase to get to that number.

 

[Derec]

Yes. Hence why I used “uncontrolled”.

No, this model is uncontrolled.

But you are right that it will slow and be more linear across the middle.just don’t know where the middle exactly is yet.

We already had two "middles". One in Spring, one is Summer. Partly due to control efforts, but mostly because we have a geographically large country where different regions had the main outbreak at different times.

I would say we are close to the peak (middle) of wavelet 1c. I think we will peak by the end of November, and probably by the 25th.

 

[Derec]

Another way to look at it, about a total of 3 million people (according to antibody study) got it in the initial outbreak in NY/NJ/MA/CT, we had 7 day averages peak at 2,100 or so people dying a day. We currently have 4 million active cases.

The hidden numbers early on were much bigger than now, because there was very little testing. We still have too little testing, but much better than March/April.

 

[bilby]

I don't think there is a problem with my math.


About 3.3%, but yeah, to one sig fig 3% is right when it coms to confirmed cases.

So you are assuming that the actual number of infected is 8-15 times what is reported. Possible, but unevidenced.

I think overall it's probably more like ~5x, with early 1a wave that hit NY and North East in general having a greater share of hidden cases, probably more like 8-10.
Take NY. Official cases are still only 3.0% of the population, but an antibody survey in April found almost 14% of New York state residents were infected by then.

(To have any significant effect according to epidemiologists, it needs to be over 20%. I don't pretend to know all the formulae, but I know that such formulae exist and are grounded in fact)

A 5x multiplier would put US at an average rate of ~16.5%. That would mean many regions in the country are at >20%, and others are getting there.

Further, the amount and duration of immunity conferred by getting infected and recovering is yet to be understood. So I consider that line of reasoning very shaky.

That is true, but the virus has been going for almost a year now, and we still have not seen widespread reinfection. That is a hopeful sign.

Rallies and election-related activities have shown to be effective at spreading the virus, no question. But now we have holidays... and a lot of really stupid people who think that a public announcement of a vaccine trial that was 90% effective constitutes permission to ignore the fact that there is a raging pandemic.

That's idiotic. A vaccine is 0% effective before you take it, and the Prizer vaccine is not yet approved, much less available to the public.

I agree that we are probably near peak. I just don't see a rapid decline from the peak happening soon, vaccine or no vaccine.

The down-slope will probably be comparable to the up-slope. Maybe a bit slower, but probably not so much as to take until inauguration day for new cases (7d average) to drop below 80k.

In the short term I still believe saturation testing is the answer. If we know exactly who has it, we can isolate and control the whole thing in a matter of 8 -10 weeks. If antibody tests are ubiquitous we will also quickly learn about conferred immunity - it's efficacy and duration. I'm hoping testing turns out to be the low-hanging fruit for the next administration.

I agree with that. We need to test many more people so asymptomatic/mild spreaders can be identified.

Sorry, but did you really just imply that stupid people would refrain from doing something because it's idiotic?

I would suggest that doing idiotic things was something of a hallmark of the stupid person.

 

[Elixir]

Sorry, but did you really just imply that stupid people would refrain from doing something because it's idiotic?

That's what I read. I guess stupid people doing stupid things is inconsistent with the Happy Talk.

I would suggest that doing idiotic things was something of a hallmark of the stupid person.

QFT.

 

[blastula]

I think we will peak before the end of November, so I definitely think we will drop below 80k well before inauguration day. Probably even by New Year.

With the holidays coming up, no way.

 

[Shadowy Man]

I think we will peak before the end of November, so I definitely think we will drop below 80k well before inauguration day. Probably even by New Year.

With the holidays coming up, no way.

Late January then?

 

[Jimmy Higgins]

I think we will peak before the end of November, so I definitely think we will drop below 80k well before inauguration day. Probably even by New Year.

With the holidays coming up, no way.

It is impossible to tell. Right now, states are re-enacting limitations and closures... so this will impact transmission. IE "THE MODELS WERE WRONG!!!! I'D DON'T UNDERSTAND SCIENCE!!!! WUHAN FLU!"

There is every reason to believe cases will increase, but states can influence this. Right now, Covid-19 cases might be slowing down again both as state policies impact spread and the human beings with compassion or a brain left in the US, are reacting to the increase in cases. Of course, with the Red Hats, they'll probably try to throw the biggest Thanksgiving super spreader even they can.

 

[blastula]

I think we will peak before the end of November, so I definitely think we will drop below 80k well before inauguration day. Probably even by New Year.

With the holidays coming up, no way.

Late January then?

Unless our behavior changes dramatically, I would guess it would be like the flu, and peak in February. Maybe the vaccines could have an impact before then?

 

[Jimmy Higgins]

Late January then?

Unless our behavior changes dramatically, I would guess it would be like the flu, and peak in February. Maybe the vaccines could have an impact before then?

I'd imagine the vaccine will help prevent death, not cases, as most of the people spreading the disease won't qualify for the vaccine this early in deployment. Granted, we need to have the vaccine deployed too, and the Trump Admin doesn't have a plan for that, other than CVS and Walgreens I think, which won't work.

 

[Loren Pechtel]

Late January then?

Unless our behavior changes dramatically, I would guess it would be like the flu, and peak in February. Maybe the vaccines could have an impact before then?

I'd imagine the vaccine will help prevent death, not cases, as most of the people spreading the disease won't qualify for the vaccine this early in deployment. Granted, we need to have the vaccine deployed too, and the Trump Admin doesn't have a plan for that, other than CVS and Walgreens I think, which won't work.

The vaccine won't be preventing that many deaths at first as the first line will be the medical people--who are mostly young enough to have a good survival rate.

 

[Jimmy Higgins]

I'd imagine the vaccine will help prevent death, not cases, as most of the people spreading the disease won't qualify for the vaccine this early in deployment. Granted, we need to have the vaccine deployed too, and the Trump Admin doesn't have a plan for that, other than CVS and Walgreens I think, which won't work.

The vaccine won't be preventing that many deaths at first as the first line will be the medical people--who are mostly young enough to have a good survival rate.

Will it be health workers first? I would presume yes in hotter spots, no where it isn't as hot.

 

[lpetrich]

Coronavirus second wave in Europe may be easing, but someone is still dying every 17 seconds - The Washington Post

Europe’s painful second coronavirus wave may be starting to ease, a top World Health Organization official said Thursday, though its toll continues to be staggering, with someone on the continent dying every 17 seconds from the virus this past week.

The cautious assessment came after new diagnoses of the novel coronavirus slowed across Europe to 1.8 million cases, compared with 2 million the week before. Some of the worst-hit countries — including Belgium, France and the Czech Republic — have seen exponential declines, while in Germany and elsewhere the curve is just beginning to bend.

But hospitals remain packed, and deaths across the WHO’s 51-nation European region have been rising.