In close collaboration with Seoul National University's Structural Complexity Laboratory

 

This is an old revision of the document!


Covid-19 Information

Click here if you want to know why I'm sharing this. But much more important…

The line below is a link to the most detailed analysis of the Covid-19 epidemic I've seen. It's information everyone needs, to understand why we are the cusp of the crisis, and what we need to do.

Please click on this link, read the content, and share it to others

How you can use these three resources to figure out for yourself the situation in particular countries

Australia is in the three-day tier: that's a dreadful place to be

I haven't had time to verify this story, but it makes sense – China has ramped up unbelievably fast to a huge productive capacity for the gear necessary for managing the epidemic. They don't need it all now – but we do

If you think that's worrying, try thinking about rapidly getting to herd immmunity without killing unnecessary millions as a control problem: that will really freak you (see below)

This argument makes sense on its own. When you take the next one into account, endgame c the only game in town: it's endgame C or the Italian experience (which will very soon be the even worse US experience)

There is only one option, which is to follow China and South Korea. Anything else leads to tens of thousands of deaths. And also leads to far worse economic disruption. The only question is how fast (and how fast will just be a matter of luck). In the end, we are going to follow China and S. Korea, any government staring into the abyss will blink. Which means we will bear the economic cost of shutdown anyway, but from a far worse base, and therefore for far longer.

And here's where it get's a bit Australian and a bit Political

Lest what is to come below seem all negative, I want to first share a piece of bright sunshine. Australia currently has arrival bans on all countries. The good news is, we don't need to. We could un-ban arrivals from China, Japan, Singapore, Hong Kong or Taiwan with minimal risk. Why? Becsuse their rates of undiagnosed Covid-19 are all far less than ours (and of course, they are not going to let people they know have Covid-19 on a plane). So the safest person you could meet, say on the streets of Sydney, would be someone just off the plane from China. So why is our government maintaining the ban? I think it's because it would be enormously embarrassing for Australia to un-ban say China, but still have hardly anyone come because they would know that China would not let them come back. Spin wins the day yet again! It shows how badly Australia has stuffed up. But it also shows how much hope there is, if only we could get our act together. We do not need to think pessimistically. China has completely solved its problem, they may have fewer active cases than us in a matter of days. despite around 50 times the population. And South Korea is dramatically reducing its active cases too. They are testing more people than us, and their new infection rate is far below ours, strongly suggesting that their undiagnosed infection rate is also way below ours. And they got there without huge social disruption. But the bottom line is, the whole of East Asia is now safer than Australia, despite going to Hell and back. So why are we still on that road to Hell???

Prime Minister Morrison, on the slim chance you encounter this page, right now you have two choices.

One is to ignore your spin doctors, forget the spin, and act now. So far, you have made a complete hash of things. You've now made a start, but nowhere near enough. I don't know whether you are getting bad advice, or whether you are ignoring or dithering on good advice, but the result is a disaster. Don't delay. Your delay last Friday doubled our rate of infection. Days, even hours, matter. If you act right now, today, you will bear some very real political cost for a few months. But your government has long enough to weather that. We won't escape some pretty bad consequences because it's already too late, but once the disaster strikes fully in Europe and the US, as it will, you will look Churchillian, the man who saved the country.

The other is to continue as now, reacting far too little and far too late. People will look at China, NZ, and probably the rest of East Asia, and see that a catastrophe was avoidable, but you failed to avoid it. You will destroy not only your own career, but the chances of your party for the foreseeable future. From a political perspective, you are already facing disaster. Your government, and the NSW liberal government, by your combined bungling of the Ruby Princess and other cruise liners, together with your government's further bungling of other international travel, especially from Italy, Iran and the US, have contributed a huge proportion of Australia's case load. When the general population see through the spin, and they will, no amount of blaming beachgoers – stupid as they may have been – is going to save you.

If both those decisions are too distasteful, there is another: to distance the government from the difficult decisions that need to be taken, by appointing a supremo and giving them the power to take necessary actions. It needs to be someone with a demonstrated capacity for decisiveness, for listening to the experts, and sufficiently respected and independent that their decisions are followed. It could be a thankless task.

Prime Minister, please note that Australia is in the three-day-doubling tier. That's a dreadful place to be. Countries in that tier (or even the four-day tier) with strong leadership are now in full lockdown. You should have done the same weeks ago, but you have had to be forced to do the right thing by the state premiers. Do the right thing now, or get out of the way so someone who isn't perpetually looking at the spin can take over!

One other concern: the UK seems to be softening people up for a strategy of getting to herd immunity as fast as possible, and with as few deaths as possible. If you're considering any variant of this strategy, including 'flattening the curve', you need to consult control theorists because it is a control theoretic problem: please do not filter their advice through others, you need direct advice from control theorists. Please see this article for the beginnings of an analysis of why. I say 'the beginnings' because on a quick read, I think the model omits the key issues of stochasticity in response and stochasticity in measurement. Stochasticity makes the problem even harder. But it's a good start, the only one I can find on the web that even starts on the problem, and his conclusion seems to be pretty similar to mine: that a prudent 'flatten the curve' strategy would be essentially indisinguishable from an infection minimisation strategy.

I've subsequently read that the UK plans to shut everyone vulnerable out of the way for four months (I guess that's how long they expect it will take to get to herd immunity) and just let 'er rip. That might be a good strategy if we knew pretty much everything there is to know about the virus and if the control problem was easier. But we don't and it isn't. Remember SARS? Remember Amoy Gardens in Hong Kong? That's where 321 people acquired SARS because the sewerage system was mis-designed and an aerosol of SARS-laden sewage was able to spread throughout the complex. It can't have been a very obvious problem (who would accept a modern apartment that stinks of sewage?). The Covid-19 virus is very closely related to that of SARS (that's why it's called SARS-COV-2). It's more likely than not that it could spread in the same way. Given Sydney's recent construction stuff-ups, if we can't trust developers to build structurally sound buildings, do you want to bet on the integrity of their sewer designs? More generally, we just don't know enough about Covid-19. Anything we don't know could put us back in the terrifying prospect of the preceding paragraph. Of course, if potential treatments such as Remdesvir or Chloroquine turn out to be effective, this might change – but if they're effective enough, we may not need to aim at herd immunity (and we certainly shouldn't be letting 'er rip, or flattening the curve, until we do know).

I note today (17/3) that Britain has clearly abandoned the 'herd immunity fast' strategy. That seems to have been because Boris Johnson was shown that he could expect 20,000 cases if he acted today, and 200,000 if he didn't. However I'm leaving the preceding two paragraphs here, because

  1. it's not clear that the Australian government has yet learnt that lesson
  2. it's not clear that either the UK or Australian governments have fully recognised how difficult the control problem in a 'flatten the curve' strategy actually is. 'Kill the curve' really is the only rational strategy.

Finally, please don't give in to pessimism. The linked ABC report seems to imply that 1.5 million cases in NSW are inevitable – I'm hoping it's a misinterpretation of the minister. Either way, it is not inevitable. China and South Korea were both in far worse states, within respectively one and two orders of magnitude more population, than NSW when they reacted decisively. They are not going to see 1.5 million cases, China is unlikely to reach 100,000 cases, and Korea will probably just nudge 10,000. OK, China's reactions were draconian (but given they had no warning, probably necessary), and we don't want that level of control here. But we have had at least two months' warning, or mabye six weeks' if you discount the period of distraction by the bushfire crisis. South Korea is a liberal democracy, and its actions have been measured but effective. We could do worse than follow their example. Yes, one key issue (on-line case tracking) infringes on privacy. But if almost everyone gives this information to google and apple and other large businesses with barely a thought, so that they can send us the right advertising, we can certainly do so to save lives.