The spy’s dilemma and the lockdown dichotomy

By Christopher Monckton of Brenchley

[Update: Good news: Boris Johnson is now out of intensive care. His prospects for recovery are, therefore, very greatly improved.]

Consider how fair-minded is our kind host. There are two very different policy positions on the handling of the Chinese-virus question: the passivist (let the population acquire “herd immunity” and hope that the virus is not much worse than the annual flu) and the activist (salus populi suprema lex: take whatever steps are needed, even if the economic cost is heavy, to ensure that healthcare systems are not overrun).

My good friend Willis Eschenbach is a protagonist of the passivist position, on the ground that the virus is not much more infectious and not much more fatal than the flu. I am a protagonist of the activist position whenever a new and fatal pathogen emerges, on the ground that until one knows more about the true case rate one must be guided by the growth rate in new cases, which, in the early stages of any uncontrolled pandemic with a population that has no immunity, is necessarily exponential.

It stands greatly to the credit of our kind host that both of these points of view are fairly reflected here, and the quality of the data and arguments being offered on both sides, not only in head postings (for instance Rud Istvan’s excellent medical postings) but also in the discussion between commenters is high. It is not unjustifiable to say that more, and more profound, information about the Chinese virus is being posted here, in a more fair-minded way, than anywhere else. This is how free speech ought to work.

Today’s post will be about how to resolve the dichotomy between the activist and passivist positions. First, the data. Precisely because the early stages of a pandemic necessarily show exponential growth, policymakers in responsible governments are guided, first and foremost, by the mean daily rate of growth in confirmed cases – i.e., cases the great majority of which are identified and reported because they are serious.

During the three weeks up to March 14, the date on which Mr Trump declared a national emergency, the global daily compound growth rate in total confirmed cases was almost 20%. Exponential growth that high, if it had been allowed to continue, could potentially have killed millions to hundreds of millions worldwide. That, above all, was the reason why governments decided, albeit with extreme reluctance and (in the UK and the US for instance, much later than they should have done) to interfere with transmission.

The problem with being late is that the lockdowns had to be much more severe than they would have been if the interventions had been more timely.

However, as our daily graphs here are demonstrating, the lockdowns are working. Of course, some countries – notably Sweden – have not introduced strict lockdowns, and yet the daily case growth rate is falling there too. That is one of the chief reasons why the passivists argue that if we too had not introduced lockdowns the numbers would have fallen just as fast and just as far.

Another reason, well reflected in a recent posting by Willis Eschenbach, is that official sources originally predicted ten times the deaths they are now predicting and that, if they had gotten the predictions right in the first place, no lockdowns would have been needed.

The excellent Dr Fauci, for instance, had predicted 200,000 deaths in the U.S., but Mr Eschenbach, on the basis of a model, considers the number may prove to be only 20,000.

Naturally, any model worth its salt will necessary look at the case growth rate at the time when the run begins. The daily case growth rate of 20% that had obtained before March 24, applied also to deaths (a lagging indicator), would have turned the cumulative 20,000 deaths up until then into 310,000 deaths by April 8 and 17 million deaths by the end of April.

As it is, by 8 April worldwide deaths were less than 90,000. And why? Because the mean daily case growth rate has been falling. Over the 15 days from March 24 to April 8, the daily mean growth rate in deaths was just 10.5%. If this lower growth rate were to be continue till the end of April, there would be 800,000 deaths by then and not 17 million. So of course current runs will be showing far lower estimates of the eventual death toll than earlier runs.

It is elementary calculations like these, based not on predictive models (which are useless in the early stages of a new pandemic) but on the observed exponential growth rates, that led governments to decide that the passivists, for the time being, would not be heeded.

The case growth rate continues to fall. Because lockdowns work, some of that decline is attributable to them. Here is today’s updated graph, showing that, for the world excluding China, whose data are unreliable, the daily mean case growth rate has fallen to less than 8%:

Fig. 1. Mean compound daily growth rates in confirmed cases of COVID-19 infection for the world excluding China (red) and for several individual nations averaged over the successive seven-day periods ending on all dates from March 14 to April 8, 2020.

But what about deaths? Taking the world as a whole, excluding China, the daily compound growth rate in deaths has fallen to about 10.5%, while in the U.S. and U.K. it is about 16-17%. Note that the graph begins on March 23, not on March 14, and that, as with the case graph, the rates shown are weekly-smoothed rates, to iron out the often large daily fluctuations in counts.

Fig. 2. Mean compound daily growth rates in reported COVID-19 deaths for the world excluding China (red) and for several individual nations averaged over the successive seven-day periods ending on all dates from March 23 to April 8, 2020.

Now that we have the data before us, how can the spy’s dilemma assist us in resolving the conflict between the activists and the passivists? That is an important question at present, because the passivists are justifiably impatient to end the lockdowns, for the economic damage they cause is considerable, while the activists, with no less justification, would rather be sure the case growth rate will not return to 20% if the lockdowns are lifted.

This dichotomy between two legitimate and strongly-argued positions is the spy’s dilemma. Imagine an agent in the field. He will nearly always be investigating a subject in which he has no specialist knowledge, and he will also have incomplete and potentially inadequate or even inaccurate data. How, then, can he advise his superiors sensibly?

I once gave a lecture on the spy’s dilemma to 200 trainee James Bonds at the Intelligence School of the Army of Colombia in Bogota. Using global warming as an example, I said that, as a non-specialist in climatology I had had to try to decide between two competing scientific points of view: the passivists, who thought there was nothing much to worry about, and the activists, who thought the planet itself might be at risk of destruction unless capitalism were closed down.

How was I, as a layman, to decide between the graph of the past 1000 years’ temperatures produced by the formidable atmospheric physicist Hubert Lamb, and reproduced in IPCC’s First Assessment Report, and the hokey-stick graph produced in the frankly Communist academic environment of today by Mann, Bradley and Hughes?

I began by saying that data generated by totalitarian are generally more suspect than data produced by those with no Party Line to defend. Therefore, I said, one would instinctively prefer Hubert Lamb’s graph to the hokey-stick graph. However, though evaluation of the likely reliability of source data is always desirable, it is not on its own always definitive.

I explained how Socrates, Plato and Aristotle would have resolved the two competing positions by the use of elenchus, still the most powerful technique for reaching the objective truth ever devised.

Fig. 3. Hokey-cokey: Hubert Lamb’s reconstruction of the past 1000 years’ temperature (top panel, from IPCC, 1990), which shows the medieval warm period as warmer than the present and the little ice age as colder, was replaced in IPCC (2001) by Michael Mann’s infamous hokey-stick graph (bottom panel), followed by many me-too graphs that purport to abolish the medieval warm period and the little ice age.

The climate activists say that global warming goes chiefly into the oceans, causing sea level to rise. Therefore, alongside their statement that the hokey-stick graph is true we can place the new statement, with which we expect them to agree, that sea-level rise or fall is an indication of temperature rise or fall. We can then draw conclusions from that additional statement.

The simplest way to decide which of the two competing 1000-year temperature graphs is correct is to compare them both with an independent graph of the past 1000 years’ sea-level change. Only one of the two competing temperature graphs closely follows the sea-level graph. The other, very conspicuously, does not. Therefore, I said, even a non-specialist with an open and enquiring mind could reach a rational – and correct – discernment of the objective even when faced with two directly-competing expert positions, and even in a field in which he holds no qualifications: for that is what an intelligence agent in the field must do every day. For that lecture, I was awarded the Intelligence Medal of the Army of Colombia:

How, then, should be apply the Spy’s Dilemma to the lockdown question? See how similar it is to the climate question. There are two competing scientific positions, both of them having some sound arguments in their favor. The data are manifestly incomplete, inadequate and often downright inaccurate.

For instance, the British government, comprising an unduly high fraction of innumerates, has not yet understood the importance of keeping a very careful track of how many of its confirmed cases have recovered. The reason why this matters is that, during the early stages of a pandemic, the least inaccurate way of deriving the true case fatality rate is to study the closed cases – those who have had the infection and have either recovered or died. Globally outside China, the confirmed-case fatality rate thus derived is currently about 25%. That seems very much too high, leading to the suspicion that Britain is by no means the only country whose experts have not understood the importance of keeping an accurate count of those who have recovered.

The official figures have stated for several days that only 135 of the 60,000 confirmed cases have recovered. If that were truly the case, it would be an indictment of the National Health Service. So you can expect the head of the Joint Intelligence Committee, who attends all Cabinet Office Briefings on emergencies such as this, to demand that ministers get their act together and require the health service to provide a proper daily count of those who have recovered. It is known, for instance, that of the first 2249 intensive-care cases almost 400 have recovered. Inferentially, a far larger faction of the less serious confirmed cases will have recovered by now.

In tomorrow’s daily update, I shall describe some further methods of intelligence analysis that would assist governments in deciding when and how and to what extent to bring lockdowns to an end. For now, I shall point out that the pandemic will not have reached its peak until the daily compound confirmed-case growth rate becomes negative. At present, it remains strongly positive, though trending in the right direction.

Therefore, it would not be appropriate to assume that half of all cases – let alone half of all deaths – have yet occurred. We all want the lockdowns to end, but at present it is better to wait a little longer. So keep safe.