Coronavirus

[jhm]

Looking at the google mobility reports I was reminded that the real world has ramps and asymptotes to behavior changes. Naively I would have expected step changes in response to public interventions like stay-at-home orders. But people’s behavior is very complicated and a moving target. Not an answer to your question but just some rambling thoughts that remind me how hard the task is.

You are trying to fit a curve over time through both an effective transmission rate that is changing and an active/resolved ratio that is changing. I think an asymptote on the effective transmission rate is screwing up your model so that recent points are now moving downward slowly at the timescale of disease resolution (up to a month to go from new case to death/recovered) - but that’s just my own guess and I wouldn’t weight it too heavily. Also your curve fit implicitly makes a fixed public policy assumption which is likely to reduce its accuracy over longer time horizons because it seems unlikely to stay that way.

I am using a curve that is flexible enough that if policy or cultural changes dictated an increase in death growth rates, it would be able to accommodate that. It just so happens that data does not currently suggest that, at least not yet.

 

[renim]

The coronavirus crisis is creating unusual allegiances as Australia fights to make it to the other side - ABC News

point to make, Japan's population is about 22x the size of Singapore's.
yet the chart treats them as similar.

 

[eevee-fan]

Anyone see the idea of giving people nicotine patch to help prevent COVID-19? Sounds fishy. What are your thoughts?

 

[RobStark]

IHME model updated just now

74k deaths in US. 2k in CA.

Projected containment phase in CA starts on May 20.

Monday June 1 should be absolute latest Fremont opens.

 

[jhm]

The total number of deaths, 52,193 souls, grew more again yesterday than my dumb-ass projections. Indeed this whole week has been more fierce than I expected. So my dumb-ass ultimate projection risen to 75,725, and the last 14-day projection has risen to 119,313.

It does appear that US deaths outside of NY are growing faster than deaths inside NY. Indeed the growth rate in NY is only about half the rate of the nation as a whole. To be precise, the US added 1957 new deaths to prior total of 50,236, a rate of 3.90%. But New York, added 430 new deaths to a prior count of 20,861, a rate of 2.06%. So clearly there are other states growing at rates several times higher than New York, and there are enough of them to boost the national growth rate to 3.9%.

You could say that the epicenter has moved out of New York. I believe this transition is what the Last 14-Day Projection is picking up on and what Dumb-Ass projection did not see coming. The fact that new cases have exploded to 37k yesterday should warn us that the worst may still be ahead of us.

We are going into the weekend. So expect numbers to be lightly reported for a couple of days. Stay safe, y'all.

View attachment 535744


View attachment 535745


View attachment 535746

As of today, 56,797 souls have passed. Weekend reporting lags seem to continue to suppress counts Sunday and Monday. So we are actually at the low point in the weekly cycle.

A curious thing has happened, ultimate projected deaths are nearly the same comparing the dumb-ass model (75,804) and the 14-day trend (79,139). You can also see this in the following chart.

Some folk may have a hard time appreciation the log scale used in this chart. It can give the impression that the model expects rates to decline very quickly. Perhaps using the nominal scale would help this perception.

So of you have questioned whether the growth rate can converge to zero within a finite number of deaths. You might also wonder if an exponential model might fit a little better. Well, I include an exponential fit curve in the above chart. The advantage is that the death rate goes to zero with infinite number of deaths, not finite. But alas the fit is not nearly as good 3 and 4 weeks ago. Let me know if you think this could be worth exploring further.

Here are the other charts.

 

[RobStark]

A curious thing has happened, ultimate projected deaths are nearly the same comparing the dumb-ass model (75,804)

So, what you are saying is that IHME are a bunch of dumbasses?

 

[AlanSubie4Life]

I clearly said no one expected the *percentage* of asymptomatic to be so high

I know. My point was that if they were publishing articles about the importance of accounting for it, it was not a negligible %. It was an appreciable % enough to be a potential issue for containing an outbreak. I am sure if you asked scientists studying it whether 25% was reasonable, at that time, they would have said yes. After all, that’s what the Imperial College study used, approximately.

as a point of information herd immunity at 9% penetration for R 1.1 assumes that the entire population mixes homogeneously.

Yes. Since I was speaking to someone who was having a hard time understanding the basic benefits of test and trace, I was trying to keep it simple.

yet the chart treats them as similar.

Generally since outbreak size is independent of country size, per capita plots don’t make a ton of sense. This was heavily litigated a few weeks ago by the Twitterati.

You know what an essential business can do to help? Open the business

Last I checked they weren’t essential. Seems like producing spare parts is essential, etc., but no reason to build new cars. Few people have a reason for them right now anyway.

he’s been getting increasingly annoyed at the media/popular thought lately (went off on CNN which is a right wing thing these days) and been seeking out more and more fringe outlets like McAfee etc.

Honestly, Elon has only himself to blame. If he had just executed on the “Alien Dreadnought” factory concept a couple years ago, rather than wasting time on a mini sub and some meme-lording and witty rejoinders on Twitter, that Fremont factory would be pumping out Model Y’s with no people at all. Epic fail.

Regarding stock price, it crossed my mind today that maybe the market is pricing in a landslide Democratic victory in the House, Senate, and Presidency in November? Presumably that would be very good for Tesla over the next few years, as it is now abundantly clear to all that we should be heeding scientists and the data now, rather than waiting for disaster to unfold...

I can’t really wrap my head around such an awesome price otherwise, though I can’t say I am unhappy! Who knew Tesla would be the hedge against -20% quarter-over-quarter economic growth?

 

[Norbert]

As of today, 56,797 souls have passed. Weekend reporting lags seem to continue to suppress counts Sunday and Monday. So we are actually at the low point in the weekly cycle.

A curious thing has happened, ultimate projected deaths are nearly the same comparing the dumb-ass model (75,804) and the 14-day trend (79,139).

Surely the 14-day trend is especially sensitive to a Sunday+Monday low count, especially if it is even lower than usual. It'll take more than 1 day to balance it out (if it will).

 

[Norbert]

IHME model updated just now

74k deaths in US. 2k in CA.

Projected containment phase in CA starts on May 20.

Monday June 1 should be absolute latest Fremont opens.

View attachment 536654

Still showing close to zero (57) deaths-per-day for US, June 1.

 

[MC3OZ]

Has this been posted?

https://www.smh.com.au/world/europe...ford-group-makes-strides-20200428-p54nrb.html

 

[eevee-fan]

Has this been posted?

https://www.smh.com.au/world/europe...ford-group-makes-strides-20200428-p54nrb.html

The Las Vegas Mayor would like to help, we can designate the Trump hotel as the control group and another hotel as the test group.

 

[Daniel in SD]

Anyone see the idea of giving people nicotine patch to help prevent COVID-19? Sounds fishy. What are your thoughts?

I want it to be true so badly. haha
I've never smoked a cigarette in my life but I don't like anti-smoking zealots (except @AlanSubie4Life, he's ok).
84% of patients in this recent study of 5700 hospitalized patients in NYC were "never smokers". I don't know how that compares to "never smoker" rates in the demographic around these hospitals but that seems high to me.
https://jamanetwork.com/journals/jama/fullarticle/2765184
Could just be that smokers are less obese though...

 

[dhanson865]

One of these things is not like the others...

 

[dhanson865]

 

[renim]

...

Generally since outbreak size is independent of country size, per capita plots don’t make a ton of sense. This was heavily litigated a few weeks ago by the Twitterati.
?

We are long past outbreak, that idea is indeed correct for early modeling, but that is long past.

Australian federation nearly excluded Western Australia and included New Zealand. That didnt happen, it is really arbitrary and quite unrepresentative to compare excluding per capita. If only because the effect is felt on a per capita basis.

Its an extra step if effort to make the comparisons per capita, but it is relevant.

 

[AlanSubie4Life]

IHME model updated just now

74k deaths in US. 2k in CA.

Projected containment phase in CA starts on May 20.

Monday June 1 should be absolute latest Fremont opens.

View attachment 536654

They are eternal optimists, I guess! In their defense, it was never intended to predict the number of deaths.

As far as factory opening, June 1st seems doable, but they had better start coming up with a workable plan! It isn’t going to just happen! It’s not going to be back to business as usual!

 

[dhanson865]

 

[renim]

View attachment 536684

Again per capita

Even if china is underreporting by a factor of 4, their turmoil must be insignificant compared to italy....

 

[Daniel in SD]

Has this been posted?

https://www.smh.com.au/world/europe...ford-group-makes-strides-20200428-p54nrb.html

The Las Vegas Mayor would like to help, we can designate the Trump hotel as the control group and another hotel as the test group.

Worked for 6 out 6 rhesus macaque monkeys. They're going to test it on 6000 people next month. Seems like they should give it to essential workers in NYC.
There are some naysayers saying that humans have never made a vaccine for a coronavirus but that's not true. Humans have made coronavirus vaccines for farm animals. There's never been a reason to make one for humans.

 

[AlanSubie4Life]

Its an extra step if effort to make the comparisons per capita, but it is relevant.

It’s a complicated topic that can be fully delved into by looking into tweets from about two weeks ago.