The big question: When we open will cases grow exponentially?

Who knows?


We won’t find out, until someone tries it.

Governors are banding together, because none of them have the cajones to do it on their own. If it doesn’t work, they want to be able to pass the buck to the consortium.

As far as I know, anybody that tells us what’s going to happen after opening up is guessing, just like I’m about to do.

This is just a possibility

I made the chart below, using data from the COVID tracking site.

It plots the daily (rather than cumulative) number of tests given in the U.S. (blue line) and the daily number of COVID-19 positives.

The vertical axis is a log scale to make it easier to compare the trends of the two numbers.

One thing that I’ve considered a possibility for awhile is that we mistook the steep growth in the red trend line from March 1 to about March 20 as the spread rate of the virus.

From the chart, it’s apparent that it follows the growth in testing. When testing levels off, so do the number of new cases found each day.

Now, I could be wrong about that. The modelers may had somehow taken into account that the growth in testing was driving that curve, or used an entirely different method to estimate the spread of the virus.

Daily Tests


I also find interesting the trend in the percentage of positive tests. That tells me something about the percentage of people who might have the virus — at least based on the selection criteria for getting to take a test.

Early, testing was restricted to high risk cases with clear symptoms. I would expect a higher positive rate in that period, and that’s what we see.

As testing ramped up, it criteria became less restrictive, but it still is, at best, likely limited only to folks with symptoms. So, it’s still not a good indicator for the percentage of the population with COVID-19 (at the time of the test), but it may be a good indicator of the percentage displaying similar symptoms.

Here’s the trend in that:

Percent Positive

That settled in at around March 20 at 20% and has stayed there. That was just as lockdowns were getting started and testing was getting close to leveling off.

What does this tell me?

I think it’s possible that the virus was spreading slower than we thought before we locked down. In some areas, that have lots of closed, shared spaces with elevators and mass transit, maybe it was spreading faster.

And, maybe that means we won’t see explosive growth when we reopen. And, we might not be up against as much of a tsunami in dense areas if we practice some basic precautions, like wearing masks, washing hands, disinfecting surfaces, etc.

That’s just a guess. I could be wrong.

Another WAG I’ll make from this data, which I’ve stated before, is that the virus might be rolling through the population in 5-10% chunks. That’s based on assuming the 20% positive test rate is double or a little more the rate of COVID in the total population.

If it’s been here since January and took a bit to get to the first 5%, we may be in the 4th to 6th wave of that, which would put us somewhere between 20-40% of the population that has had it.

Something that makes me doubt my guesses is why would we have had the peaks if this thing was spreading steadily?

Though there are a few potential contributors that fit with the guesses above.

First, we brought panicked attention to it. We shined a spotlight on something that could have been already happening just as our capability to identify it (testing) was growing fast, which made it look like it was growing fast.

Two, that panicked attention may have caused more spread as people went to hospitals ill-equipped to keep infected patients from spreading the virus to workers, patients and other visitors.

Third, my ‘shutdowns may have concentrated more cases into a few days‘ theory could also contribute.

The first contributor would not have caused more cases, but as we shone the spotlight on the rapidly increasing cases (red line in first chart), while ignoring the blue line, it made it look like we were heading for a peak.

The second and third contributors could have both caused more cases and more people to be infected in a short window, causing the peaks.

I do not have a high degree of confidence in the above and welcome the poking of holes in it.

Things that make me doubt the idea that this is a steady spreader are the Diamond Princess and the Smithfield meat plant.

Though, both meet the criteria for having a high spread rate (closed and close quarters).

A key piece of info for both is how long it took to get to their infection rates. I haven’t found that data, yet.

What Arnold Kling said

His name is Arnold Kling! I agree. Especially this bit:

I am still cranky, for the same reasons as before. We still do not have results of random-sample testing to have an idea of how many people have the virus in various regions. We still do not know how viral load affects outcomes. We still do not know whether people get it from touching surfaces and then touching their faces, or whether they have to breathe near an infected person. We still are not conducting any scientific experiments. Instead, we are contemplating mass “experiments” with changing regimes for social distancing, arguing over putting at risk either more economic activity or more human lives. Worse yet, we really won’t learn anything from these “experiments.”

We still talk about political leaders “re-opening the economy,” as if the economy is theirs to re-open and individual choices will not be affected by the virus. People still expect that any and every household and business can be saved from the consequences of the virus, because government has the know-how and skill to undertake this. People still conceive of government as an infinite storehouse of riches that is disconnected from any need to obtain the wealth that it purports to distribute.

Any press conference at any level should address what is being done to work toward the info in Kling’s first paragraph and how long it will take.

Why I’m skeptical of experts

I’m skeptical of experts.

Too many folks put blind trust in experts, or people they think are experts, but who really aren’t. Often, they think rich dudes like Bill Gates or Warren Buffet are experts in everything, when the scope of what made them rich was fairly narrow and doesn’t necessarily transfer.

For example, Bill Gates is an expert in writing software and spending the money he made from writing software. It should really go without saying that doesn’t mean he’s an expert in everything.

In this post back in 2011, I criticized his policy-stances on education. I believe I just heard today that he’s running charter schools in Seattle now, instead of wanting to build a single point of failure into the public education system, like he was then. That might be progress for him. Though, given how innumerate we are these days, the damage he has done with common core math may already be baked.

At that time it had taken him 10 years and $5 billion to learn that in public schools he was fighting a vested interest (teacher unions), who also love single points of failure. So, my guess is that he hasn’t suddenly become a believer in innovation and competition, but rather, even his billions couldn’t beat the political power of the union. He may be about to learn the same thing with WHO.

Lots of people try to convince skeptical folks, like myself, that I’m dumb for not trusting the experts, or pseudo-experts, like Gates.

Then the experts blow it. Sometimes big time. This pandemic is a good example of how experts have been blowing it time and time again.

That sometimes puts a little dent in their trust in experts, for a bit.

But, then the next time I express skepticism of some “expert” BS, I hear “you better trust the experts,” again. I ask if they remember the last time the experts blew it? “Well, they’ve learned,” or, “this is different.”

There’s no need to debate how much of an expert someone is. They may be the most expert in whatever topic we are discussing.

I keep one thing in mind: experts have been wrong, plenty.

I especially distrust experts that display classic tells that they use their status to manipulate (and sometimes fool themselves), like speaking with utmost confidence (that tells me they have rationalized away, rather than learned from, their past wrongs), not  willing to explain their work (they think their status puts them above scrutiny) or shutting down disagreement with fallacies (this tells me they may be over dramatizing their expertness to manipulate for achieve another end, like getting us to not wear masks when we should be or to avoid the obvious flaw in their pretty model).

Experts that I am more apt to listen to are ones that are humble because they remember their past failings and they know they could be missing something on this one, too. They often say things like, “I could be wrong.” Or, “If my theory is correct, this is what I would expect” and ask, “What do you think?”

They don’t mind being wrong. They seek it out because they are more interested in getting to a good answer than being right. And they know that it is rare that the first answer is the best answer. They realize that getting to a good answer is an iterative, trial-and-error process, emphasis on the error.

Most of all, they can and love to show their work. They don’t expect you to accept their conclusions just because they are the expert and you are not. They can explain how they got to their conclusions in a way you can understand.

They also welcome feedback, even from folks who aren’t experts. They know that fresh eyes sometimes see glaring flaws missed by experts who are down in the weeds and diverse world experience might bring in a viewpoint or idea from another field that can help.

That is my work on why I am skeptical of experts. Of course, I could be wrong.

Why do American soccer coaches hate juggling?

I see soccer coaches on Twitter and real life snipping at each other all the time about things they don’t think helps players develop.

It’s common to see spats about unopposed training, juggling, fake moves, kicking the ball against the wall, 1v1s, small-sided games, toe taps or whatever.

In my opinion, it’s all good and all has a place. It’s not either one thing or another. What sport is?

The fact that we even have these spats tells me the soccer culture is still in its infancy.

Unfortunately, I see that attitude rub off, which takes away the players’ chances to discover what they can make out of these activities and how much it can help them.

I have never heard a basketball coach complain about kids playing OUT or practicing dribbling on their own.

I also bet soccer coaches in Europe or South American don’t complain about their players wasting time juggling.

So why do we?

I farted around on the basketball court a lot as a kid. Basic games like Around The World or OUT got me started. Over time we built off those basics using our creativity to come up with versions that were even more fun for us.

For example, I had a few different versions of Around The World that I played that kept me shooting when my friends weren’t around (like the Superman, which was flying around the world 7 times as fast as you could go).

I do the same with these basic soccer games. For example, I have a handful of juggling game variations that I cycle through. I started with the basic one that all American kids, except Christian Pulisic, and some coaches seem to hate with passion (how many can I get in a row?), learned a few version from others (e.g. how quickly can I get to 100) and made up a few of my own (e.g. left vs. right, my left wins more often than you might think).

The key isn’t WHAT activity.

The keys are doing an activity, having fun with it and being invested enough to use creativity to mix it up and make it even more fun.

It’s through all of those variations that kids might come up with fun versions that end up going viral and doing more to improve American soccer than the USSF, club or coach could ever do.

Rather than poo-poo’ing these things, we should let it happen.

Are COVID-19 peaks caused by stay-at-home orders? Part 2

A study of outbreaks (where three or more people were infected in one go) in China showed that ~80% were home outbreaks (HT: Marginal Revolution). Transport was second, involved in 34% of the outbreaks (the same outbreak could be in multiple categories).

The same study showed that all outbreaks involving three or more cases were all from indoor transmission, which may be an indication that getting outside might be good.

This might also support the theory that stay-at-home orders caused the initial peaks of cases many places experience.

This theory doesn’t say that stay-at-home orders cause more cases. Rather, it just condenses a week or two worth of new infections into a 1-3 day window after the order goes into effect as those with the virus become more likely to spread it to their families.

This essentially creates a “rush hour” of cases 2-3 weeks later, that was mistook for an exponential growth of the virus.

Consider how viruses typically move through your household. It’s rare for everyone to get it at once. Part of that is driven by incubation period and previous immunity. But, some of that may be other factors like how often you cross paths and for how long.

In a normal day of work, school, soccer practice, meetings, travel, TV and homework, you might have an hour of face time with your family, which might lower the chances of spreading something you are infected with to them.

But, when everyone stayed at home at the same time, you are with them longer. And, if you have the virus, the chance of you spreading it to your family in the days after the order goes up considerably.

If true, you might expect to see less severe initial peaks in areas that did not panic and lockdown at some point. But, those areas are becoming more scarce.

CDC chart updated through April 4

US Deaths by Week 2

They are starting to add the COVID-19 cases. You can see that % of deaths by COVID-19 has spiked up to 6.9% for the week ended on April 4 and was almost as much as by pneumonia.

While the total deaths appear to be on a declining trend, I suspect that they don’t have all the data for the past few weeks, so the columns won’t be reliable to look at for a few weeks.

I also suspect the 6.9% to possibly change, so I’m only looking at it as an early indication. I suspect that number will go up to 15-20% range for last week and maybe this week.

Could the lockdowns cause the peaks? Thinking out loud

I know that sounds stupid.

But, I saw someone point out that places tend to peak in cases and deaths about 2-3 weeks  after lockdowns take effect, which also happens to be the amount of time window for the illness to progress to the point of seeking help and potentially dying.

They also pointed out that places like S. Korea also didn’t lock down.

That made me go hmm…

Theory A is a lot folks were already infected from exponential spread before the lockdowns, the lockdowns stopped them from infecting as many others therefore the peaks lagged the ceasing of the exponential spread when the lockdowns took place. faster.

That’s probably right.

But at least hear this person’s theory out before going off on them in a blind rage. At first I thought it was dumb, then after thinking about it, there might be something to it.

Theory B is that when the general population is interacting business as usual, the virus spreads steadily, rather than exponentially, like other cold viruses.

But, when you stop normal society and lock families in together all day at the same time, that greatly increases the chances of family members who were infected of spreading to their family at that time.

So, something that was slipping through 10-15% of the population of the time, can suddenly spike to 2 to 4 times that at once because families went from spending a few hours together at night to being together 24/7.

Maybe the infected would have eventually spread it to their family, but like other viruses, it would have been over a few weeks just due to normal busy schedules, instead of so much within a few days.

So, if you happened to be locked down with someone who was infected, suddenly your chances of catching the virus at the time of the lockdown, instead of next week or the week after or never, increased, and that’s what caused the cases to spike a few weeks after lockdowns.

Theory B might be a crackpot, but it certainly wouldn’t be the first time that a knee jerk (or non jerk) government action had subtle and bad unintended consequences.



My basic expectations from press briefings

Advice to the President, Governors and Reporters:

If we’ve shut down the economy to keep hospitals below capacity, consider that there are more than a few us out here who have been holed up for a few weeks (and, yes, we notice that your hair is somehow looking fine), I expect every briefing to open with this info:

  • Hospital ICU utilization
  • % infected recovering at home vs. % that need hospitalization
  • Hot spots — where hospital utilization >100% and what % of surge capacity is being used
  • How long is those hot spots expected to last
  • Will surge capacity be adequate there
  • Where are the next hot spots expected to be and is the surge capacity in place
  • What metrics are you tracking to inform your decision to re-open and what are you looking for them to be?
    • If these metrics are different than hospital capacity, why have they changed?

In addition, it would be helpful to give frequent updates on this information:

  • What we know about risk factors:
    • Fatality rates by both age and comorbidity
    • % of population infected from the start (so we know how much of the population is still at risk)
    • What we are doing to get a better estimate on those fatality rates so we’ll better understand our risks (like random antibody testing)
  • What do we know about how treatments are working and preventative measures we can take.
    • Is that having an impact on death rate?
  • Trend in total deaths since this started, not just those we think are COVID-19.
  • Any other things that we’re learning
  • Current plans for re-opening
    • Are we going to test areas and see how that impacts new cases?
    • Are we looking at other countries that have had more success to learn from them?
    • Are we looking to companies that have operated all the way through this so we might apply what they’ve learned?
  • Where are there shortages on PPE and hand sanitizer

Advice to Reporters:

How about we get people back to work before we play the “would’ve/could’ve/should’ve” game?

Quite frankly, you don’t have any legs to stand on in that game.

And, instead of that, when that time comes, how about you ask better questions?

  • What changes are being made to improve how everybody responds to this in the future to make broad shut downs the last resort, instead of the only resort?

I have some thought-starters: Broad, early testing. No centralized CDC testing. Contact tracing. Targeted isolation. Wear masks. Robust stockpiling and production of PPE, masks and hand sanitizer. Antibody testing. Quick setup of mobile testing stations to keep infected patients out of hospitals. Get the private sector involved early so they can provide solutions. Easy setup of surge capacity, separate from existing hospitalized populations.

A numbers team that can put a virus spread into some helpful context — like how does this compare with previous big ones, like H1N1.

Model projects nearly 3 million deaths in the U.S. over the next 12 months!

It occurred to me after writing this post about the chart that will give us a truer sense of the impact COVID-19 has on deaths in the U.S., that I missed the chance to write a headline in the same alarming, sensationalist fashion as the media.

So, here it is.

And, the model this projection is based on is historical averages. The U.S. averages 2.6 to 2.9 million deaths, annually. I predict that will continue.

I hereby nominate myself for the Pulitzer for this groundbreaking reporting.

Time value of money: Wimbledon edition

You’ve all probably heard that Wimbledon paid $2 million per year for the past 34 years for pandemic insurance and now will receive $141 million.

Sounds smart. Maybe it is.

But these are the wrong numbers to determine that.

The right numbers are the rate of return and opportunity cost.

The rate of return on a $2 million annual investment for 34 years with a $141 million payout is about 3.9% per year.

Had they invested that $2 million a year in an S&P 500 mutual fund, they would have had $246 million, so their opportunity cost of that insurance has been about $100 million in today’s dollars. The opportunity cost may even be higher if the markets happen to rebound once things get going again.

Or, that means the folks who gave them the insurance are about $100 million ahead. Not bad for them.

Granted, Wimbledon also wouldn’t have had the coverage had the pandemic happened sooner, but they would have had some nest egg to help them through.