If I flip a coin

Or a thread

Italy just reported their most deaths ever in a single day (993). It's a high day, but not a total fluke. Their 7-day average is now over 90% of what it was for the original peak in March. They are almost 2/3 of the way to total reported deaths from that peak, so are virtually certain to pass it in total deaths as well.

Almost no country in Europe, except maybe Scandinavia, is going to have fewer deaths in this second wave than in the Spring. All of Eastern Europe has already passed their small or non-existent first waves, in both peak and overall magnitude (i.e., total deaths). Switzerland and Portugal have now passed it in both as well. Germany has passed it on peak and just about equaled it in magnitude. That just leaves the rest of Western Europe, all of which are likely to have long downslope tails: France topped out at 65% of peak, but is already 70% of the way in magnitude; the UK seems to be topping out now at 50% of peak, but is already 50% of magnitude as well; Belgium topped out at 70% peak, and is now at 70% magnitude; Netherlands topped out at 60% peak, and is now at 60% magnitude; Spain has a very different shape, but may be topping out at 35%, with 60% magnitude so far. The only success stories in Europe are Ireland, with their extreme measures, and Scandinavia, where the jury is still very much out, as cases and deaths have been rising.

Remember all the way back in early September, when we were supposed to be at herd immunity in Arizona? After dropping to a very low number in October, hospitalizations are now at 80% of their peak just 4.5 months ago, and have been climbing rapidly.

#10703, yep. On September 24, the Rational Ground guy* wrote that Florida had likely reached herd immunity. Since then, they've reported ~300,000 new cases, ~5,000 new deaths, and hospitalizations have nearly doubled (although are still <50% of their peak).

* not the one that Florida actually hired to do data analysis.

At a press briefing at the White House today, Kayleigh McEnany's husband, Sean Gilmartin stood at the back and watched the entire event sans mask. He was asked to put a mask on by a pool reporter, so he left. Apparently, they don't just hire replacement level people at the WH, they also marry them.

brian brianson

please understand that i am not mathly gifted. i looked up what square root is but do not understand why it has anything to do with whether or not you get a yes/no answer to something (like the coins). i would appreciate it is you would pls explain. thx

if you have only 20K people who get a vaccine, then you are only guessing that 1% have even been exposed and if you don't continue to test or do antibodies, you can't possibly know how many were actually infected, only people who have symptoms. so 1% of 20K is only 200 people, and the death rate is even .5% then you wouldn't get more than 1 person dying. that is really not enough in mah not so umble opinyin, to be sure that the infections with this vaccine are any good at preventing death. even only 20 people with any symptoms you might not get anything serious or death. i don't understand why you think this is not an extremely small sample size.

ESPECIALLY if any of those 20K people were masking/social distancing, which we don't know

#10706 It has to do with standard error. Which is an important step in "does this study tell us anything interesting?"

Check out the Wikipedia page on standard error.

if you have only 20K people who get a vaccine, then you are only guessing that 1% have even been exposed and if you don't continue to test or do antibodies, you can't possibly know how many were actually infected, only people who have symptoms.

If this helps, I think that all of the 44,000 people in the trial (including those getting the placebo) are regularly tested throughout, so they have a pretty good idea how many were infected - and the fact that nearly 20 times as many people were infected in the placebo group as in the vaccinated group suggests to me a pretty high degree of confidence that it's not just random chance. Although I also don't have the maths to prove it.

ben

i knew how low on the maths scale i am when i was trying to figure out "adjustment factors" for any baseball stat that is not easily done by simple calculator. when aaron gleeman came up with his gpa, he adjusted by "a factor" that i got no idea where he got it from, to make the numbers look the way he wanted them.

but then again, i am the one who thinks reached on error, dropped third strike reaching 1st base should be added up in OBP

Trust me, I thought I could write saber-ish stuff 15 years ago. Put up a couple of articles on a fansite that got laughed out of town. My skills, whatever they are, do not incline mathwards.

if you have only 20K people who get a vaccine, then you are only guessing that 1% have even been exposed and if you don't continue to test or do antibodies, you can't possibly know how many were actually infected, only people who have symptoms.

So, the very short is if you're testing forty thousand people, you give half the vaccine, and half nothing. Then, the half you give nothing will let you know roughly what happened to the other half you gave the vaccine.

There's a little bit of randomness - in a group of 20 000 Americans, ~20 will be named Brian, but if the first group had 24 and the second had 18, that could just be chance. So, you measure the number of Brians in the first group, you find 20, you can reasonably guess the second group will have 15-25, and be very confident it won't be 0 or 1100, right? Any human trials work the same way. The exact numbers are slightly more complicated, but not much. The total number of people you might be wrong by goes up as the total number of people you measure goes up, but the fraction goes down.

The math isn't that hard, but unless you do it a lot it's not on your finger tips; but it's a random walk, or drunken walk. Drink a dozen beer, and walk away from something. The longer you walk, the farther you get. But walk twice as long, you don't get twice as far, because you wander and stagger around. I know I do (but I live closer to Belgium, so obviously I drink more beer).

Lisa, without getting into the math of it, think about it like you would think about sample sizes in baseball. If you have two guys, and one goes 1/10 and the other goes 3/10, it doesn't really tell you much. If they go 10/100 and 30/100, you're pretty confident the guy who had 30 hits is a lot better. And if one bats .100 over a full season and the other bats .300, you're virtually sure the .300 guy is much better and you don't need to see a full career from both of them to come to that conclusion.

This is a similar concept except the difference in magnitude appears to have been even greater -- 162 people on the placebo contracted COVID whereas only 8 people on the vaccine contracted it, if I'm reading the article correctly.

Already another new record for reported new cases today. Likely to be a new record for reported deaths as well

never mind; stat was fixed for NE.
Yes, new records. America #### Yeah!

Drink a dozen beer, and walk away from something.

Hold my beer.

Pandemic Data Are Stalling Out: Thanksgiving has skewed reporting of COVID-19 cases and deaths, but one metric is still clear: Hospitalizations keep rising.

Article by the Covid Tracking Project at the Atlantic site.

Over the summer we were getting +3, +5, +1 new cases a day in Frederick County.

A couple of weeks ago we were getting +55, +60.

Yesterday it was +126.

Thanks, anti-maskers. I hope you get an imaginary vaccine for this imaginary virus. And I'm sure an imaginary hospital bed will fill your needs.

If I had to guess, I would think we are not even a third of the way through reporting backlog deaths that would have otherwise been reported over Thanksgiving. It's very hard to know though.

Ron J Posted: December 03, 2020 at 02:25 PM (#5992281)

#10706 It has to do with standard error...

- i tried reading that page and i know it is not meant for math geenyusses, but i am stuck at the part about "error" because there IS no error, the numbers are real. i know if you check ML stats, that the average BA is gonna be X and there will be players on either side, but their averages are REAL, there isn't any error in uncorrected hitting stats even though the BA lined up from worst to best is gonna show a curve like that

after the first 3 sentences when they started showing actual math with all thse symbols i don't know what they mean, i had to let it slide

Tony S Posted: December 03, 2020 at 08:59 PM (#5992351)
Over the summer we were getting +3, +5, +1 new cases a day in Frederick County.

A couple of weeks ago we were getting +55, +60.

Yesterday it was +126.

Thanks, anti-maskers. I hope you get an imaginary vaccine for this imaginary virus. And I'm sure an imaginary hospital bed will fill your needs

- stay safe tony

and don't worry about the anti maskers. first off, they are White and White people don't get the covid, which don't exist seeing as how donny poo said so, and so they ain't gettin no vaccine, seeing as how it is for no reason, and if they get sick and in the hospital or on a respirator, well then it is because it is really flu and those EVULLLLL doctors are lying about it so as to get more money, as donnie poo said. he has killed more americans than any other president

White House COVID-19 task force pleads with Texas health officials to warn the public

OK, now I fe3e3l totally through the looking glass. Or is it that someone's too busy fighting Hugo Chavez to stop the task force from admitting there's a virus?

harris county is not doing too bad though. thank GOD we have a dem judge, mayor and sheriff. mask compliance in stores is really good except for the usual few males who just gotta poke their dingdongs over their nose diaper

Best guess now for excess deaths, through the end of this week, is over 430,000. Could be as low as 400,000, but probably not too much lower.

CDC estimate is 335,000 or 368,000, depending on which count of theirs you use, but that includes nothing for the last three (very bad) weeks, and is definitely low for at least the three weeks prior to the last three. Adding those in and there's pretty much no way you get an estimate much less than 400,000 for the CDC, and it could very easily be over 450,000. Unfortunately the range of the estimate keeps widening each week as the data gets shittier and shittier.

With 3.5 weeks after Saturday to go until 2021, and no real signs of reversing just yet, if I had to guess right now my estimate by year end is likely to be over 500,000.

yea indeed, Lisa, Hidalgo has been nothing short of wonderful.

Crackpot Nobel laureate on twitter: "Still early to be certain of 2nd wave anywhere."

That was posted 6 hours ago.

Lisa there is standard error in a lot of stats you see all the time. Why for instance do we not use BA to evaluate players? Because various other metrics have smaller standard error when we attempt to use the stat to predict scoring.

Standard error is the scoring system we use to for instance evaluate offensive metrics in baseball.

Standard error is also something I brought up in discussion of baseball's economics. Is a given signing likely to make economic sense (fair bet the answer is no. Free agents as a group are paid about 30% more than their contributions to team's bottom line is worth). But the models I was able to build had a standard error of about $10 million and I always try to put that out there so people get an idea of the general level of precision available. (I'm confident that modern estimates are better and the best ones are trade secrets)

Or: Well there are scientists here (I'm not -- though I've spent a lot of time working with them. My statistics are self-taught and it shows sometimes). I'm told that one serious red flag for any study is that it's missing a discussion of precision. Standard error might not be applicable, but there pretty much always is something that says, I'm "this" confident within "these" bounds.

10724 -- And drew this reply:

@Marco_Piani 11h
Update: today comes with the largest number of deaths reported ever during the pandemic.

Let me go out a limb and say that Italy may have seen the start of the second wave. We need more data to be certain, though.

- i tried reading that page and i know it is not meant for math geenyusses, but i am stuck at the part about "error" because there IS no error, the numbers are real. i know if you check ML stats, that the average BA is gonna be X and there will be players on either side, but their averages are REAL, there isn't any error in uncorrected hitting stats even though the BA lined up from worst to best is gonna show a curve like that

So, we say "error" because we're typically not trying to measure how many times it actually happened, but how often it "should" happen. So, in the vaccine trials, maybe the vaccine is 93% effective in a sample of a billion people, but 94% effective in our trial of 40000 people, we have a 1% error in the measurement of it's intrinsic effectness.

I like coins as an example because they're straightforward. If I want to test a coin to see if it's "fair", I can flip it ten times. If it's fair it "should" come out as 50% heads, 50% tails - but if I got 4 heads, 6 tails, I wouldn't conclude it's unfair; that might be chance. In a trial of 10 flips, the typical error is 3, because (being good at maths) 3x3=10, so a result of 0, or 10 heads is suspicious, but 3 or 7 isn't.

And so, if we're thinking about rates, 10 trials gives you an error, or uncertainty, of 3, or 30%
for 100, it's 10, or 10%
for 1000, it's 30, or 3%
for 10000, it's 100, or 1%
and so we rarely bother with more than that.

In a trial of 40000, if one person dies, well, our relative uncertainty on it is large, it's 1 plus or minus 1, so 100% uncertainty on how many people would typically die, but the absolute rate is still 0%-0.005%, so not that uncertain. And the same more or less applies to how many people were wearing masks, or going to bars, or licking public toilets to achieve internet fame or whatever.

- stay safe tony

and don't worry about the anti maskers. first off, they are White and White people don't get the covid, which don't exist seeing as how donny poo said so, and so they ain't gettin no vaccine, seeing as how it is for no reason, and if they get sick and in the hospital or on a respirator, well then it is because it is really flu and those EVULLLLL doctors are lying about it so as to get more money, as donnie poo said. he has killed more americans than any other president

Thanks bbc. I'm double-masking now when I go to the store, and doing everything I can to minimize physical interactions with anybody. I've pretty much written off Christmas and made my peace with it. I've been hiking the C&O trail, which follows the Potomac, trying to cover all 184 miles of it piece by piece. It's actually kind of awesome to drive out to the remote western reaches on a weekday morning and having the trail all to myself. But it's getting cold now... :(

I'm trying to not get shot climbing the Berlin Wall when it's coming down anyway in a few months.

White House COVID-19 task force pleads with Texas health officials to warn the public

Meanwhile, back in "Sports City, USA" Frisco, TX, the front page above-the-fold story in the free local newspaper this week was a gushing review of our local Convention Bureau's efforts scavenging for bottom-tier bowl games, youth hockey tournaments, and anything else under the banner of "Sports" to come have their event here instead of those states that are taking a -- shall we say -- more active role managing the pandemic.

Something called the New Mexico Bowl has agreed to move here this year (cue those "Los Angeles Angels of Anaheim" jokes), as has some Minnesota travel hockey weekend tournament, and they're hard at work looking for as many others as they can find. The payoff is that each of those events brings the promise of a few thousand dollars of hotel and restaurant tax revenue to the city coffers (and the convention bureau gets paid from a cut of that money). So now we know what price the City puts on public health. Our mayor is owner of what is probably the largest realtor firm in the city and definitely toes the party line when it comes to "Virus? What virus?" rhetoric.

I thought last year's unveiling of a Dallas Cowboys-branded City fire truck would be the absolute nadir of the City's infatuation with sports. Even I never thought they would actively endanger public health over it, but that seems to be where we are.

OK, now I fe3e3l totally through the looking glass. Or is it that someone's too busy fighting Hugo Chavez to stop the task force from admitting there's a virus?

With Trump on the way out, a number of agencies have started to more or less openly defy him.

The latest Trump campaign fundraising numbers ought to clear up any question about why President Donald Trump continues to push the election fraud lie to the bitter, illogical end: He’s raking in hundreds of millions of dollars from it. The campaign announced ahead of its federal filing that it had raised $207.5 million to be exact since Election Day on the back of 498 fundraising pitches to donors since Nov. 3, setting a monthly record for solicitations from the campaign. The fundraising windfall even after losing the election has been generated by Trump’s frothing-at-the-mouth claims of election fraud, which he has used to solicit hundreds of millions from low-dollar donors for his “official election defense fund.”

10729 -- That's not a fire truck. Every boy knows fire trucks are red.

And speaking of colors, here's a graphic rendition of COVID case spread since October

With 3.5 weeks after Saturday to go until 2021, and no real signs of reversing just yet, if I had to guess right now my estimate by year end is likely to be over 500,000.

So, what you are saying is that we have a chance to beat the 650,000 deaths from the 1918 flu pandemic.

So, what you are saying is that we have a chance to beat the 650,000 deaths from the 1918 flu pandemic.

But that's only people not a percentage of the population, so any good Hoovertarian will tell you it's insignificant.

And speaking of colors, here's a graphic rendition of COVID case spread since October

Geez, what could possibly go wrong encouraging people to travel here from the Upper Midwest for the eternal glory of a weekend travelpuck youth tournament?

Crackpot Nobel laureate on twitter: "Still early to be certain of 2nd wave anywhere."

I definitely have no interest in defending crackpottery, but I'd love to see some careful studies of how genuinely second-wave second waves really are. Are the same places getting hit again, or are we seeing other portions of the same countries, that were spared in the spring, now getting hit for the first time? I see here, for example, that Madrid currently has one of the lowest death rates per capita in Spain, so maybe the current "second wave" in Spain is just the rest of the country finally getting their first wave. (Of course, if you drill down far enough, every apparent second wave can be construed as a local first wave.)

You get a very nice double hump of hospitalization rates up I-35 from you in Tarrant County on their chart (slide 7)

I definitely have no interest in defending crackpottery, but I'd love to see some careful studies of how genuinely second-wave second waves really are. Are the same places getting hit again, or are we seeing other portions of the same countries, that were spared in the spring, now getting hit for the first time? I see here, for example, that Madrid currently has one of the lowest death rates per capita in Spain, so maybe the current "second wave" in Spain is just the rest of the country finally getting their first wave. (Of course, if you drill down far enough, every apparent second wave can be construed as a local first wave.)

connecticut, delaware and rhode island should be sufficiently small enough (in both size and population) for you to see clearly that a 2nd wave exists, and that it both appears to be worse than the first wave, and is still increasing in severity.

No second wave in Madrid? Even if you throw out cases as being too unreliable...

On June 19, there were 8416 recorded deaths in Madrid, and deaths had slowed to a trickle. By August 19, two months later, there were only 100 more recorded. There are now 11,426 in Madrid--almost 3000 more, and more than a third as many as there were in the original wave.

That's not that different from the rest of Spain, where the current wave has a little over half as many deaths as the first wave.

edit: excess deaths for Spain are also way above reported deaths, in fact, were almost double as of September. Pro rata, that would put Madrid at around 20,000 dead now, or 0.3% of the population, with a quarter of that coming after the first wave ended.

Even correcting for broadened testing, estimates show cases in places like Rhode Island now similar to what they were in the Spring. Deaths are still expected to be a bit lower at least for now, mostly (it is thought) due to better treatment methods.

Thanks for the various helpful case studies, everyone. So much for that simplistic theory about the geographic spread. (It's mildly tempting to drop down one more geographic level and look at changing rates within neighborhoods. But that's getting dangerously close to the reductio of looking at rates for each individual, and it also just looks prima facie implausible to me that neighborhoods are geographically hermetic enough to have useful individual impact rates.) Well, on the one hand, that leaves me still wanting better information about why things go up and down where and when they do (one thing the whole COVID experience has brought home to me is that I don't really understand why epidemics end. What was going on with multiple waves of bubonic plague in medieval Europe -- why did it stop after a while; where did it go when stopped; why did it return when it did?). On the other hand (and in partial answer to the question of the first hand), I guess it's some evidence of efficacy of social preventative measures, if a first wave can pass through and still leave a population vulnerable to a second wave.

Also, does anyone know what the base population is from which the Navajo Nation COVID statistics are drawn? Wikipedia says ~173,000, which would make their current fatality total of 663 a rate of about 3800 per million (which if I'm remembering correctly would be about equivalent to the excess deaths estimates for the worst-hit areas of NYC). But I wasn't certain that the two sources were using the same population definitions.

It seems like it should be at least 190,000 or so, based on the regions they display. Total population, if you use cases and cases/10,000, is about 160,000, not including "Bordertown", which appears to refer to Navajo Nation people living in numerous border towns, estimated to be over 30,000. In addition, there are another 100,000+ living scattered throughout the USA. I'm not sure if these count.

I'm not sure if these count.

I don't imagine they do. Navajo Nation is not like Red Sox Nation; it's a specific land area.

This document doesn't seem to totally distinguish when it comes to population.

Navajo Nation claims over 300,000 enrolled tribal members and is the second largest tribe in population, following the Cherokee Nation. According to 2010 U.S. Census, there were a total of 332,129 individuals living in the U.S. who claimed to have Navajo ancestry.
The Profile includes the population on the Navajo Nation, the Navajo population in the bordering towns of the Navajo Nation, and in the metropolitan areas with a high populace of Navajos. Using the 2010 U.S. Census American Indian and Alaska Native (AIAN) Summary Files data, the Profile provides a documentation of the population in three enumeration groups, listed below.

Yes, but on p. 13,

Population on Navajo Nation
Total Population on the Navajo Nation
The 2010 U.S. Census enumerated 173,667 people living on the Navajo Nation including all races.

To include the entire Navajo population, people in LA and Houston would have to identify as Navajo on their tests and be coded and reported as such. That's a big ask. And there would have to be a Navajo Nation identifier without any other Native American identifiers, which seems odd.

Agree it would seem strange. What about "Bordertown"? It's probably easier for Navajos there to be be coded as Navajo Nation in the nearby towns even if technically they are not in the land area.

Nope, that is broken out as a separate population (something like 70,000) on the next page. The report gives total, nation, bordertown, metropolitan pop. breakdowns.

My guess is that the NN number comes from people who use the health service on the Rez. I don't know if folk in the bordertowns can, between BIA and insurance companies, they'd probably die waiting for an answer.

I'm still confused. This dashboard implies the border town people are counted, though doesn't clearly state that. And this document says the border town population is a little over 33,000.

OK, the document you linked carefully distinguishes the two realms, and also distinguishes All Navajo, Navajo Plus and Other Race on the Nation. The dashboard does appear more inclusive.
The dashboard explicitly (note 3) includes Bordertown case in total positives.

My error in Bordertown population is no more than further proof that I've reached an age at which I should not rely on recall; I recalled the Navajo Alone population.

Dark Clouds over the Sun Bowl
Coronavirus causes El Paso's Sun Bowl game to be canceled for the first time in 85 years

Has there been much/any public talk about priorities of vaccine distribution? That is, who gets it first, and who has to wait for manufacturing production to catch up with demand?

I've seen widespread talk that vaccines will go first to "the most vulnerable" and frontline health care workers. It's hard to argue that, if it means everyone over 65 because the death rate is clearly skewing higher with age (and while I'd like politicians and public health officials to stand up and say this bluntly in public, I understand they probably never will: too much of a hot potato) and first to everyone who works in ICU because of greater exposure to corona.

But beyond the obvious, has there been any discussion of the next, and further, levels of priority?
Answers from around the world are invited here.
Just some questions to chew on:
1) Will governments use public health records to "fast-track" some individuals or cohorts?
(a) for example, if they decide after everyone over 65 gets the vaccine that next priority would be those over 50, would they also contact (or start an application process for) people under 50 who have a long or difficult history within the medical system and would move them ahead of those with few health problems?
(b) for example, will it be distribution by population, or has anyone said it should/will go to areas of greatest cases/and/or/death while areas with proportionally lower numbers will be vaccinated later?

2) Will the priority system differentiate between those who have tested positive for covid and recovered, compared to those who have tested negative, or never been tested?

3) Will antibody (blood/serology) tests for past infection and recovery from covid become more important than diagnostic PCR ("currently infected" swab) tests? Will resources be applied to antibody testing to help prioritize vaccine distribution?

Has there been much/any public talk about priorities of vaccine distribution? That is, who gets it first, and who has to wait for manufacturing production to catch up with demand?

So far, they’ve said that they are going to prioritize healthcare providers and older people in long-term care facilities and their caregivers. If they have made a decision for sure after that, I haven’t seen it. Apparently, the thought is to give it to people who are most vulnerable to dying, which would be older people. The alternative, I guess, would be to give it to the people most likely to spread it, but it may be too late to make much of a difference there.

The other high-priority targets would be first responders, essential workers, people in prisons, and hoping by then that there is an adequate supply for everyone else.

The NY Times has a little questionnaire that estimates your priority to get it. No guarantees as to its accuracy, of course.

In the UK, they managed a U-turn on prioritisation about 24 hours after the MHRA announced emergency approval of the Pfizer/BioNTech vaccine - originally health care workers were to be at the front of the queue, but then the decision was switched so that staff in care homes were before them. Which meant that some NHS staff had already booked vaccine appointments before the news came out. These will apparently still be honoured.

The need for cold storage creates some interesting logistical requirements - you don't want to break up packages of this vaccine too much, as then your storage needs multiply, and you're at much higher risk of throwing away part of what is quite a limited supply. The UK also needs to balance local distribution, not least because there are political implications. Some of the north of England, like Manchester, has been voicing dissatisfaction at the amount of central government support received. Scotland's health service is run independently by a political party (the SNP) that strongly opposes the governing party in Westminster. Same is true, though to a lesser extent, for Wales - but their health service is integrated with England's.

It's certainly a complex challenge, but needing to change course within days of approving the first vaccine doesn't bolster confidence that the UK used the quiet summer months to prepare very well.

In Connecticut, they have laid out a plan for vaccine distribution along the lines discussed above.

Phase 1a (to begin within the next couple of weeks) will include healthcare workers, nursing home residents and medical first responders.
Phase 1b (mid-January to late May) will include critical workforce, other congregate settings, adults 65 and up and anyone considered “high-risk” under age 65.
Phase 2 (early June) will include anyone under 18-years-old, remaining residents over 18-years-old.

My friend who is a public school teacher told me that they are included in the "critical workforce".

The U.S. Has Passed the Hospital Breaking Point: A new statistic shows that health-care workers are running out of space to treat COVID-19 patients
Robinson Meyer and Alexis C. Madrigal, The Atlantic

A month ago, in early November, hospitalizations passed 60,000—and kept climbing, quickly. On Wednesday, the country tore past a nauseating virus record. For the first time since the pandemic began, more than 100,000 people were hospitalized with COVID-19 in the United States, nearly double the record highs seen during the spring and summer surges.

The pandemic nightmare scenario—the buckling of hospital and health-care systems nationwide—has arrived. Several lines of evidence are now sending us the same message: Hospitals are becoming overwhelmed, causing them to restrict whom they admit and leading more Americans to die needlessly.

Focused on lives and the function of the health-care system, it takes no position on "wave."

Geez, what could possibly go wrong encouraging people to travel here from the Upper Midwest for the eternal glory of a weekend travelpuck youth tournament?

You mean because that's the only youth sport so far correlated to virus spread?

lets talk about all those supposed people in the hospital. first of all they are making it all up. fake news. there aren't really people in thehospital and they aren't dying. well, the ones that are dying are dying from starvation because of those stupid librills and dokterz insisting on closing down all the bars and Fun Places to go. if they are sick or dying it can't be from covid because there ain't no covid and besides White people are immune

as for me and Husband, well, we'll keep wearing our masks. even after getting the vaccine, we're wearing our masks until the virus is gone because i don't trust no vaccine that doesn't protect from disease

Has there been much/any public talk about priorities of vaccine distribution? That is, who gets it first, and who has to wait for manufacturing production to catch up with demand?

like every other commodity in modern capitalism, priority will go to those who have the connections to jump the line. some performative altruism will take place, and it be hyped up as a "feel good" story by the media, but 6 months from now, 9 months from now, there will still be inner city hospitals and rural hospitals that will not have had access to even a single vaccine dose for their staff.


fun fact:

this will become a cudgel that right wing media will use to attack the biden administration and within his first 6 weeks in office, foxes and friends will return to "don't pull the plug on grandma" republicanism. those of us who aren't immediately decapitated by whiplash will immediately die from aneurysms.

Another 200,000+ cases, 2,000+ fatalities day. I fear we are in for a run where this is the rule, not the exception.

as for me and Husband, well, we'll keep wearing our masks. even after getting the vaccine, we're wearing our masks until the virus is gone because i don't trust no vaccine that doesn't protect from disease

yeah, i would definitely recommend continuing to do that, if only to piss off the people who are going to lie about being vaccinated whenever they refuse to wear masks.

I believe you are correct, Srul; haven't checked the teevee newsfeed, but I assume GA's totals will be spiking in couple weeks from today

Another 200,000+ cases, 2,000+ fatalities day. I fear we are in for a run where this is the rule, not the exception.

i don't know that we'll hit 5,000 deaths per day by january 1st, but it might be close.

this will become a cudgel that right wing media will use to attack the biden administration and within his first 6 weeks in office, foxes and friends will return to "don't pull the plug on grandma" republicanism.

Yeah, I was just thinking about this yesterday. How long will it take Fox News to start blaming Biden for the pandemic deaths (and the size of the budget deficit)? A month?

Florida is back up to averaging nearly 10,000 cases per day -- not quite peak levels but pretty close. I will probably re-run my analysis from the summer sometime this week, since they are almost at the point where the number of cases exceeds the number of rows in an Excel spreadsheet (about 1.05 million), at which point such analysis will become tougher for me. Maybe this will give me a reason to learn Python, which I've been meaning to do for a couple of years.

Also, has anyone noticed a lot of people who seem to think that the COVID IFR is 0.25% again? I thought we put that delusion to rest months ago when the CDC finally updated their best estimate to something like 0.65%, but I keep seeing people--even some otherwise smart people--cite numbers like 0.25% or 0.3% online.

Yeah, I was just thinking about this yesterday. How long will it take Fox News to start blaming Biden for the pandemic deaths (and the size of the budget deficit)? A month?

You're way too generous. January 21st. We had, what, ONE death from Ebola and the GOP was ready to impeach Obama. And, yes, the deficit will suddenly become an issue again.

But this might finally get the anti-maskers to join the real world.

Anyone have a theory on what is going on with Utah's numbers? Seems like a real outlier with overall positivity rate close to 10% and yet CFR is under 0.5%, about half of the next lowest state.

Depending on who you ask, Utah can be as low as 9% total infected or as high as 13.5% as of three weeks ago (as of the time deaths would now be reported). Reported deaths is 939, or close to 0.03% of population. Utah's excess deaths appear to be close to twice their reported deaths, which puts them at the high end for the US, so you can estimate that as 0.055%-0.06% of their population. Dividing the two and you get an IFR of 0.40% to 0.67% or so (measured by excess deaths). Then, Utah has the youngest population in the country, so to put it on roughly the same scale as the US you should multiply by around 1.5. That leaves you with 0.6%-1.0% IFR. That's not out of the normal range for this later phase of the virus, now that treatments have improved.

If you do this same exercise for all states, you generally get similar results.

Utah: also low smoking, low diabetes, low obesity.

Pat Rapper must be truly delighted to know that Frisco, TX, is the home of this year's New Mexico Bowl as well as the Tropical Smoothie Bowl and all the displaced travelpuck squads.

Giuliani has Covid. (per Trump tweet, so evaluate credibility accordingly)

I don't recall seeing him wearing a mask anytime during his traveling clown show the last few weeks.

Yeah it’s just weird. I’m used to seeing the COVID deniers claim it’s just the flu, but even some people who take the pandemic seriously claim the IFR is only 0.25%. (I was having this debate with one of our former resident posters on Twitter recently. He was also claiming the case multiplier was like 8x which still seems impossibly high to me.)

Giuliani has Covid.

Thoughts.

Giuliani got hospitalized within two hours of the announcement (which is probably why the news came out now). So he's obviously had it for awhile.

If he knew about it and still cavorted around maskless, he's a potential murderer.

Responding to President Donald Trump's tweet that Rudy Giuliani has tested positive for coronavirus, Dr. Megan Ranney says that Giuliani could have potentially exposed "hundreds and hundreds" of people to the virus during a recent trip to Atlanta. CNN has reached out to Giuliani for comment.

CNN

Pat Rapper must be truly delighted to know that Frisco, TX, is the home of this year's New Mexico Bowl as well as the Tropical Smoothie Bowl and all the displaced travelpuck squads.

I have no doubt City officials must be hard at work trying to scavenge the Sun Bowl, the Redbox Bowl, the Bahamas Bowl, the Hawaii Bowl, and all the other canceled bottom-tier ESPN bowls too.

This is not the first time Marcus has found herself caught in the middle of a tense, complicated cultural battle involving medical misinformation and an us-against-them vibe taking hold in her community. I first met Marcus in April 2019, as a measles outbreak raged in Orthodox Brooklyn, culminating in the official declaration of a public health emergency in select Brooklyn ZIP codes and the shuttering of local yeshivas. Marcus had just launched the EMES Initiative—EMES, which stands for engaging in medical education with sensitivity, means truth in Hebrew. When we spoke last year, she was in the midst of editing, printing, and delivering 10,000 copies of a 40-page informational booklet about vaccines to the doorsteps of ultra-Orthodox homes across the five boroughs to combat a similar booklet disseminated by anti-vaxxer groups. She was organizing small informational sessions for ultra-Orthodox women afraid to vaccinate their children, trying to stop a neighborhood “measles party,” and meeting with New York City public health officials and Centers for Disease Control and Prevention members around her dining room table to talk about culturally sensitive ways to reach ultra-Orthodox people.

Back then, Marcus was energized. Today, she seems beat. As the coronavirus ravaged her community, the EMES Initiative escalated its efforts. It’s now grown to include 30 or so Orthodox health care workers. But the challenge they are trying to meet feels nearly insurmountable. “I have nothing left,” Marcus told me on the phone a few weeks ago, as health officials tracked a “slow but steady” rise in positive coronavirus cases across New York City.

In March and April, the coronavirus hit the Hasidic Jewish community in New York with devastating force. On March 17, the White House organized a call with 15 leading Orthodox rabbis in New York City, including prominent Hasidic leaders, to urge the community to shutter key institutions and adhere to social-distancing protocols. By March 19, even before New York’s major hospitals were overwhelmed with critically ill COVID-19 patients, more than 500 cases of the coronavirus had been identified by one urgent care center serving ultra-Orthodox neighborhoods in Brooklyn. Hasidic media outlets were reporting on the deaths of prominent rabbis and community leaders daily. “We were waking up each day to find out who died overnight,” said Marcus.

“Many believe that their family members died in the hospital because of neglect,” he told me. “It’s easier to accept that your mother was sick and the hospital neglected her, rather than believe that the grandkids came over, gave Bubbe COVID, and then Bubbe died.”

Within the Orthodox community, there are several ways community leaders can quash dissent. One particularly effective method is to threaten a person’s marriage prospects—or the marriage prospects of that person’s children or grandchildren. The threat of poisoned matchmaking prospects, or “shidduchim,” as it is colloquially called, can quickly muzzle those who might otherwise be compelled to speak out on issues from child sexual abuse to COVID-19 safety protocols

“I can’t go on living in fear,” a young Orthodox woman who lost her father-in-law to COVID in March recently told me, explaining that she tested positive for COVID-19 antibodies and therefore thinks she is inoculated against reinfection. She is not interested in wearing a mask or taking other precautions, and she believes that God, whom Jews refer to as “Hashem,” has a plan. “Not everything is in our control,” she said. “Hashem has something to do with it.”

the chosen people have embraced natural selection.

(which is probably why the news came out now).

Definitely.

Mostly unrelatedly, new denialist theory alert! Then CDC has a lower expected deaths baseline this year than they did for last year (2019), which is counterintuitive because population grows yearly. Therefore, this year's must be artificially low, to the tune of 100,000-150,000 or more. Sounds plausible, but it's wrong. The baseline IS lower than last year, but the baseline they use is a simple average of the last X years adjusted for population growt, potentially weighted more heavily by the more recent years. There's a massive problem with that for 2019--the 2017-2018 flu season, which mostly hit the 2018 year (Jan-Mar), was way above normal. There's also a secondary problem--population has been growing yearly buy it grew much faster in the 80+ cohort in 2016 and 2017 than it did in more recent years. Both of these have the result that using a simple average was going to be way too high for 2019.

So what happened coming into this year? In 2017, which was a more or less normal flu year (mostly 2016-2017 flu year with a small part of 2017-2018), and an average over 80+ population growth for recent years, the expected deaths were close to actual deaths. In 2018, which was a high flu year, but a low increase in over 80+, the two factors cancelled out and again we had close to expectation. In 2019, where flu season was again low (and prior year was high, boosting the average), and 80+ growth was low, the actual deaths were way below the expected based on simple averages. Thus, coming into this year, the CDC reduced their expectation base, as we were one further year away from 2018 which inflated it, and also had a year of low increase in old people factored in to their expectation. This reduced the 2020 baseline quite a bit.

So what actually happened this year? Until covid, deaths were BELOW the baseline coming into the year... most likely due to 80+ increase still being overestimated (growth was low coming into 2020) and slightly lower than average flu season. It wasn't until March, with the arrival of covid, that we went positive on excess deaths. In other words, the idea that the baseline is way too low this year is not backed by the numbers at all.

year     expected       actual (millions)

2017     2.81           2.82
2018     2.86           2.85
2019     2.93           2.86
2020     2.89           

We have probably around 3.10 million deaths so far, best estimate, and at the very least 3.05 million or so. With 26 days left that's another 0.22 million or so baseline (obviously, we will end up with way more than baseline over these last 26 days, but I'm ignoring that for now), so 3.27 so dar at the low end (380,000 excess already) and maybe close to 3.37 so far at the high end (480,000 already), using the CDC's baseline.

My baseline, adjusting for average flu and population growth, is very similar to the CDC for 2020: 2.91 or so.

Giuliani has Covid.


Thoughts.

wish my mouth hadn't been full when I read that.

Any detailed analysis of virtually every country in the world shows excess deaths well above reported deaths, with the US close to what would be the middle of the pack for Western Europe, in terms of underreporting. Thus, no particular reason to think that we are drastically miscounting excess deaths. Nevertheless, twitter denialists claim that the true number will be in the range of "200,000-250,000" (due to this new theory, above), with 50,000-100,000 of that due to lockdowns. So, with some creative accounting, you can turn 500,000 excess deaths into just 100,000, y'know, no worse than a "bad flu". Well done!

Ok figured out what the extent of the theory to reduce deaths. It's a fun one! You take the CDC's estimate for 2019 v 2018, which is 70,000 more. Then you say 70,000 is an average expected increase, and because we had a small increase last year due to an "exceptionally mild" flu season (which in reality was about 15,000 less than recent average), you ignore last year and add 140,000 to 2018. Leave aside that the very high CDC estimate in 2019 was based in large part on the much larger than average 2018 flu season and rapidly aging population for 2018, both trends that would not have been expected to increase under a more sophisticated analysis. That leaves you with a whopping 3.07 million expected deaths this year, which would be an increase of 220,000 over last year, and a massive jump in mortality rate.

In this way you can slice off 180,000 deaths, reducing the CDC's reported 335,000 (which does not take into account the last 3 weeks, and is low for the prior 3+ weeks), reducing excess deaths to just 155,000, half of which are lockdown deaths!

Rinse and repeat.

One sports bubble has burst - the England cricket team are on their way back from South Africa after players on both sides, at least one coach, and some of the staff at their hotel tested positive. The two teams did manage to play three games over the last week or so, but the rest of the series is now cancelled. A significant economic hit for South Africa's cricket board.

US deaths have increased an average of about 45,000 per year over the last 9 years (and not at all before then, for a decade). The CDC's baseline is 35,000 more than last year. The latest theory is that the real number should be around 145,000 more than last year, an increase that might not have been seen since the 1918 flu. The biggest recent increases were around 80,000.

There is some basis for the expectation being a bit more than 45,000--the graph of deaths each year shows that smaller years tend to come after bigger ones, and the last two years were smaller increases. However, 2-year periods over the last 10 years have never been more than about 110,000, and these increases came when the US population was increasing close to 2.5 m per year. In recent years it's been increasing about 2/3 that rate, and on top of that less the oldest cohort was expected to increase at a lower rate than the rest of the population. (edit: other sources are only showing a slight decrease per year, but still a large decrease in the oldest cohort).

Using 120,000 as the max over 2 years (to be generous), and multiplying by 2/3 for the slower increase in the oldest population gives you 80,000. Subtract out last year's 20,000 increase and pretty much the max increase you could expect this year would have been around 60,000. Add 10,000 for a leap year and you get at most 2.93 expected this year.

mortality rate (per 1,000) the last 10 years, based on CDC numbers:

2010      7.99
2011      8.07
2012      8.11
2013      8.22
2014      8.25
2015      8.45
2016      8.49
2017      8.66
2018      8.69
2019      8.71

2.89 million deaths this year (CDC estimate) would have been 8.76 per thousand. 2.91 million (my estimate coming in, since revised down slightly due to lower than average flu season) would have been 8.82 per thousand.

3 million would have been 9.09 per thousand--simply a massive and unprecedented increase--it's a ridiculous baseline, used only to massage the data into a more palatable form.

The administration's final middle finger to America.

The administration's final middle finger to America.

that seems very optimistic.

First they came for the scientists...

State police brandishing firearms raided the home of Rebekah Jones, the former Department of Health employee who built the state's much-praised COVID-19 dashboard before being fired over what she said was refusing to "manipulate data."

"They pointed a gun in my face. They pointed guns at my kids," Jones tweeted shortly before 5 p.m. Monday.

ok I ran a more sophisticated analysis of baseline expectation, using mortality tables and actual numbers of men/women each year per age range, since 2010. The expected increase in deaths is only about 65-70% of what it was in 2015-2017, when the bigger increases in yearly death totals were being recorded. I get 2.92 million expected this year--slightly higher than the CDC estimate and very slightly higher than my estimate, but well less than what you would get if you simply extrapolate out the peak increase years from the middle of the decade out through 2020, ignoring the changing demographics, and assume all deficit from the prior two years would have been made up this year--all bad assumptions.

47 year-old Magistrate Judge in El Paso placed first for election on 3 November, went into the hospital for the second time for COVID on the 9th and died Monday.

10787 do you know whether this was her second infection? Or had she recently been released from the hospital and then returned? The article doesn’t say.

10785 will be interesting to find out what that's about. FWIW, I remember finding her complaints about Florida's data early in the pandemic somewhat overblown. What she described seemed consistent with how other states reported information. And FL actually provides better data to the public than just about any other state, even if it's far from perfect.

I don't either, Dave.

10789: “ In a search warrant, an investigator with the Florida Department of Law Enforcement said a person at Jones' home who was using her email address illegally gained access to a state-run communications platform and sent a group text Nov. 10 telling people that it was "time to speak up before another 17,000 people are dead."”

Link

Edit: Same info is in Tony’s link.

10785 will be interesting to find out what that's about. FWIW, I remember finding her complaints about Florida's data early in the pandemic somewhat overblown. What she described seemed consistent with how other states reported information. And FL actually provides better data to the public than just about any other state, even if it's far from perfect.

This was my impression as well. That said, obviously unnecessary to serve a warrant against a non-violent offender at gunpoint.

Well, vaccines are rolling out. I understand at least one non-trial person has now been vaccinated, so it's only a matter of time until I can vacation in Bruges, or maybe even the Jurassic coast.

Anyways, it's been fun being terrified but also detached with all y'all.

The "fun" ain't quite over, Brian. It looks as if the US has a gppd shot at 2700 deaths for today on Worldometer, and I'm not surprised. W/w numbers had been at least 20 percent higher for infections and deaths the last several days, and a fair number of states' Monday reports are weekend numbers. Shouldn't be hard to top 200000 infections, either.

Deadliest Days In American History

1. Galveston Hurricane 8000
2. Antietam 3600
3. 9/11 2977
4. Last Thursday 2861
5. Last Wednesday 2762
6. Last Tuesday 2461
7. Last Friday 2439
8. Pearl Harbor 2403

Deadliest Days In American History

1. Galveston Hurricane 8000
2. Antietam 3600
3. 9/11 2977
4. Last Thursday 2861
5. Last Wednesday 2762
6. Last Tuesday 2461
7. Last Friday 2439
8. Pearl Harbor 2403

New #4 12/8/2020 2913

I assume we don’t have actual daily death counts for the 1918 influenza pandemic. In October 1918, an average of 6,290 people died of influenza per day.

Measured by excess deaths, a three weak period in early April averaged over 3000 deaths per day. The last three weeks might have been even higher. We'll find out eventually.

(edited since it was really about 3 weeks rather than 2)

Hank -- you're using worldometer, AuntBea is using one of the other sites. Today lagged Friday's 1926 on Worldometer

I've been using Worldometer as well as Covid Tracking. For excess deaths I've been using the CDC.

By Worldometer, the largest day is now December 3, last Thursday, with 2926. Both Worldometer and Covid Tracking add deaths to old dates not that rarely.

New York vs Florida is somewhat interesting right now. We are averaging similar number of cases, but NY is doing like 2x as many tests. NY is only averaging about 25% deaths over the past week, however. Assuming that deaths are the best indicator of the true number of cases out there (with a lag built in, of course), this implies that there are diminishing returns to additional testing. That’s not surprising. It also implies, I think, that the 8-10x case multipliers that some people are asserting just can’t be true. If there were really 7-9 undiscovered cases for every discovered case, then you’d expect the state doing 2x as much testing to have a lot more cases, given the reported death numbers are not that far off.