Here’s another way to think about the probability issue. Suppose you think Neil Ferguson is right, and 250 000 people will die from Covid-19 if there is mitigation but not lockdown. (By mitigation Ferguson means case isolation, including household quarantine, for anyone who has it, and social distancing for the elderly.) And suppose you think that this is too many. In fact, suppose that your limit is 200 000. Any more than that and you will support a long lockdown, because you think that that level of deaths outweighs any amount of economic and social damage done by a long lockdown, not to mention the deaths that will result from the damage done by the lockdown.
But here’s where you have to take into account probability. You can’t just say ‘More than 200 000 could die without a lockdown, and 200 000 is the level at which a lockdown is worth it, so we should have a lockdown’. You have to look at the probability of the claim that 200 000 will die. After all, you did say ‘could’. No-one seems to have come out and just said Ferguson is definitely right, after all. All the talk is of ‘Better safe than sorry’, which is what you say of low probability events, not high probability ones.
You might then say, ‘Well, Ferguson could be either true or false, so it’s probably 50/50’. I don’t think this probability assignment is credible, but let’s say it is. Saying there’s a 50% probability that 200 000 people will die is, in effect (if we are thinking in an expected utility sort of way), the same as saying that 100 000 will definitely die.
(I don’t mean that this sort of way of understanding probability can be taken literally in all real-life probabilistic contexts. For example, if you have a 50% chance of winning nothing or a million — say there are only two people who could win — you don’t literally have a 100% probability of winning half a million. But this way of working with probabilities makes sense in decision theory, although it also caused problems there as well.)
100 000 deaths, though, is not enough to trigger the lockdown even on the terms that you have set. Because you said that 200 000 deaths is the number that justifies the lockdown, not 100 000. So once you take probability into account, things change dramatically. The lockdown that you thought was justified now definitely isn’t. (Things change even more dramatically if you assign more realistic probabilities than 50/50.)
You might then try a different tack. You might say, actually it’s 100 000 deaths that I’m going to have as my trigger for the lockdown, and as I’m 100% certain that they will occur, we have to have a lockdown.
Is it really credible, though, to assign 100% certainty to the claim that 100 000 would die with mitigation but no lockdown? (If we were being sticklers we’d say slightly less than 100%, because only tautologies and logical truths have 100% certainty, but let’s not worry about that here, take 100% to mean almost 100%.) It would be crazy to be that sure about such a bold claim, especially as we’ve seen little sign of a mass killer so far, and it’s only computer models from known doom-mongers that provide any reason to believe such large numbers. It would be crazy enough to apply it in the situation where no measures are introduced, but even crazier to apply it when mitigation will be in place.
You might then go back to your earlier claim, and say that at the very least you’re 50% sure that there will be 200 000 deaths, and that gets us the lockdown.
In this case I would just have to disagree about the probabilities given. The idea that we can be that confident about 200 000 deaths on the basis of some black-box outlier computer models from people with no track record for reliability, and Italy’s dubious stats (which aren’t that bad anyway), and Sweden’s example where there’s no lockdown, strikes me as mad. Personally I would say there’s at most a 5% chance of there being 200 000 deaths, and that’s being pretty generous. You shouldn’t believe extreme claims without very strong evidence, and very strong evidence is most definitely what we don’t have. (Nor do we have a 25% chance of there being 400 000 deaths in this scenario — not even Ferguson has discussed that — so you can’t try to approach it from that angle).
In general we need to be taking probability into account when we discuss Covid-19 predictions. I don’t mean that we can give exact, scientifically-based probabilities for every prediction, but we should be talking in general probabilistic terms. Even if we starting talking more about low, high or medium probability, even when this is based on nothing but gut instinct, the debate can be transformed, as people start to realise that what we are betting the farm on is, after all, pretty shaky.
That doesn’t apply, though, for the other side. The economic damage is not hypothetical. It is not based on a dubious computer model. We know it’s going to happen. It’s certain (at least as certain as an empirical claim can be). We cannot be totally sure how much damage is going to be done, but our choices are between bad, really bad or apocalyptic, and the longer the lockdown the worse it gets. We are deciding between bad health damage which has low probability against some bad economic damage which has very high probability. Really (unless you’re one of those people who say we can’t put a price on anyone’s life, which is an insanely unworkable idea), there’s no contest.
Update: I note that when Toby Young argued that the lockdown doesn’t make sense even if we supposed that Ferguson was completely right, he assumed for the sake of argument that there was 100% certainty in Ferguson’s figures. Despite that, he said, the lockdown still didn’t make sense when looked at it terms of the government’s and NICE’s own concept of QUALYs (quality-adjusted life years).
Sam Bowman took issue with that, although Young defended his claim in a further article. Now, you may be unsure whether Young or Bowman is right here. But bear in mind that Young is assuming, for the sake of argument, that we have 100% certainty in Ferguson’s figures. But once we stop doing that, and get realistic, and downgrade the probability of Ferguson being right to a more realistic figure, then there’s just no question that Young is right. In that case there’s just no contest.
Update 2: I’ve got quite a backlog of graphs and diagrams that people have sent me which I will try to get up on Monday.
41 thoughts on “Taking probability seriously on Covid-19”
I have lost too many hours of my life fretting over frequentist probability versus epistemic probability. And this is not the time to lose ourselves in that maze.
Purely in terms of messaging to those who may be prepared to listen, would it be more effective simply to stick to qualitative uncertain benefits versus certain costs?
Ferguson and his ilk talk only about the uncertain benefits — deaths presumed to be deferred, adding however many months to the lives of generally old sick people. They do not talk about the certain costs of their lockdown — people laid off, businesses made bankrupt, savings destroyed. Those certain costs are not even considered in Fergie’s models.
Perhaps we should demand that any public figure who promotes the benefits of the lockdown should also have to detail the costs that will necessarily be entailed?
Totally agree Gavin.
News – from The Telegraph
Will the UK coronavirus death toll really be as high as we are told?
Apocalyptic predictions that Britain’s coronavirus death toll will be the largest in Europe have abounded over the past week.
The “scaremongering” began after a report from the Institute for Health Metrics and Evaluation (IHME), based at the University of Washington in Seattle, suggested the UK could reach 66,000 deaths by August, peaking at nearly 3,000 a day, and accounting for more than 40 per cent of total deaths across the Continent.
The figures were gleefully seized upon by the Left-wing press and emblazoned over front pages as evidence that the Government’s strategy has failed.
Yet, within hours, British experts had branded the modelling as “absurd”, and by this weekend the IHME had revised down its estimate to 37,494 – and admitted it could be as low as 26,000 which is not hugely dissimilar to Imperial College’s figure of around 20,000.
For anyone following the trajectory of deaths it was clear that something extraordinary would have to happen for our daily death rate to shoot up to 3,000. All other countries have exhibited a smooth upward trend followed by gradual leveling off, so the UK would have needed to experience a trend-defying upward kick to get anywhere close to the IHME figures.
Keith Neal, emeritus professor in the epidemiology of infectious diseases at the University of Nottingham, said: “Redoing their prediction in under a week strongly suggests major flaws in their models. This is not the first model to be shown to have got their projections seriously wrong. Although this is a pandemic, the epidemiology in each country is different and different within countries.”
Explaining the updated figures, the IHME said the new data had taken into account the effect of social distancing and included four more days of data.
But epidemiologists at Imperial also pointed out that the model showed Britain had already exceeded its intensive care capacity by three times, even though the NHS currently has plenty of spare critical care beds.
Prof James Naismith, director of the Rosalind Franklin Institute and Professor of structural biology at Oxford University, said: “I note the IHME updated their forecasts and they have substantially lowered the worst-case and central scenarios for deaths.
“It is to be greatly regretted that too much online and media coverage of the earlier IHME predictions focused on worst-case scenarios without making absolutely clear the very large ranges that the IHME clearly stated for their UK predictions. When these ranges are deliberately omitted (or obscured) by others, who then choose to focus on worst-case scenarios, this is little more than reckless scaremongering.”
It is also unfair to compare countries that have vastly different population densities, social mixing, demographics and family structures. Take Ireland as an example. As of lunchtime yesterday, the country had recorded 8,928 cases and 320 deaths. Which is 65 deaths per million people. In contrast, Britain had 78,991 cases and 9,875 deaths, 145 deaths per million. However, the population density of Ireland is lower than in Britain, approximately 186 people per square mile compared to 727 people in the UK. And while 83 per cent of Britons live in urban areas, just 63 per cent of Irish people do.
We know that the virus spreads through close contact so it is unsurprising that a country with more people, squeezed far more tightly together, is going to have not only higher overall numbers, but also a higher death rate.
The German population density is also lower than the UK at 623 people per square mile and France, which has consistently been lower than Britain, is just 309 people per square mile.
Just 76 per cent of Germans live in urban areas and 81.5 per cent of French people, again lower than the UK. London, which has seen the most cases, is also the largest city in Europe, with nearly nine million people. In contrast, Berlin has just 3.7 million and Paris 2.1 million. The median age of citizens is also likely to play a role in the final death figures. Italy’s median age is 47.3 years while Ireland’s is 38.2.
We know the virus disproportionately affects older people so those with older populations will undoubtedly fare worse. The deadliest day reported in Italy was on March 28 with 971 deaths, and in Spain on April 3 with 950 fatalities, according to the European Centre for Disease Prevention and Control.
Compared to Britain’s 66 million inhabitants, Italy has around 60 million and Spain 46.94 million, so it might be expected for overall figures to be lower. Currently the number of deaths in Italy is 322 per million and 355 in Spain compared to Britain’s 145, so it will be interesting to see where the overall death rate finishes for each country.
Likewise, it may turn out that many of the recorded deaths did not come “from” coronavirus but “with” the virus, which could slash the death rates. The latest report from the Intensive Care, National Audit & Research Centre, ICNARC, shows that around 89 per cent of people dying with coronavirus never made it to intensive care. As of April 9, when the most recent figures were compiled, just 871 patients had died in intensive care out of 7,978 deaths. It suggests that the vast majority of people are dying on other wards, perhaps from other conditions.
When it comes to the number of cases and deaths, it is also unlikely that we can directly compare against other countries because of differences in recording. Although many have criticised the Government for not supplying non-hospital death figures, virtually no other country is managing to release a complete picture of total deaths. In Spain, soldiers have reported finding abandoned bodies in care homes after they were called in to disinfect the rooms.
In France, almost a third of all coronavirus deaths have been of residents in care homes, but in Britain, the Office for National Statistics has said around 92 per cent of deaths are happening in hospitals.
Hundreds of care home deaths have also been reported in Italy and Germany suggesting that their overall figures are likely to be higher than reported. Until this crisis is over it is impossible to know where Britain will fall compared to other countries in terms of mortality but, when the final numbers are known, it will be crucial to factor in demographics or else we will never see a true picture or understand if the Government’s interventions were correct.
Discover more from The Telegraph
Download our free news app
Read more on our website
Access exclusive insight and opinion
Listen to our podcasts
I notice the headlines in the Telegraph have taken a noticeable turn to ‘what happens to the economy?’
More sobering up?
“Yet, within hours, British experts had branded the modelling as “absurd”,”
Didn’t the ‘experts’ predict 500,000?
Most people are totally incapable of cost-benefit analysis. The way most people look at cost-benefit is to consider the benefits of X versus the costs of not-X. The costs of X and benefits of not-X never come into it. And which X they wish to see implemented is determined in advance of the cost-benefit analysis, emotionally, not rationally.
This cost-benefit analysis is further complicated by exchange rate variability (how many lives to the pound?) Again, for most people, in a world where we seem to no longer accept death or have to be confronted with it particularly frequently, that exchange rate tends, in most peoples’ minds, to infinity pounds per life. At which point you have to invoke opportunity cost, and good luck getting that message across!
Most people CAN understand “I/you are RUINED” though.
In terms of economic damage we are already at “bad” and on the way quickly to “very bad”. We need to end the LD not fart about with dozy app tracking or any other face-saving nonsense. That is my fear –that because we have a BluLabour govt rotten with folks who think printing funny money will keep us going–they think they have time to save their face . By not dropping a red hot brick (which they assured us was ice cold) until both their –and our–hands are burned off.
Yes! “Out of 7978 deaths only 871 died in intensive care!” Presumably then they died in other wards from other illnesses?
Not mentioned anywhere I’ve seen is the cost in lives attributable to transferring medical resource away from treating heart complaints, cancer and presumably other things. Unless covid is actually a cure for cancer and heart problems there must be lives lost from this cause.
Yes Mind-bogglingly bloke. We live in a post-modern world now. The vast majority of people who comment here have an empirical view of working things out – BUT that is so last century. Did you not get the memo. Feelings are what its about now!!
We laugh at Orange Man Bad and his “fake News” but really there is now only fake news. It is partly designed that way and partly due to the lack of journalists who never check much and just pass it on. In recent days the radio station I listen to was just giving out daily death figs for C-19 without pointing out that a chunk each day were actually for the previous few days – just catching up. So the total was fine, but the headline was an exaggeration. In some cases like this it is laziness, in others it is by intent to sensationalise.
…maybe not, Pat.
Further up, the article says the NHS is not at capacity in ICU rooms….
I’ve found this website to be a beacon of common sense in this ‘crisis’.
Reading the Times this morning l can’t believe they are talking about extending the LD for another 3 weeks!…is there no one in government that is privy to the stats and, it seems to me, the expert opinions on here??
It seems clear to me that the government is reacting to media hype rather than sound advice.
For the record l’m a so called’key worker’ and I spent two days delivering PPE to care homes in Wakefield district in Yorkshire and none seemed “abandoned” as also stated in the Times
https://drmalcolmkendrick.org/2020/04/06/covid-with-of-or-because-of/
others mentioning it here and there now, and as you can see that includes NHS England
In ICU.
Presumably the treatment formerly being applied in wards and in operating theatres was actually saving some lives.
To listen to the news you would think that every problem apart from Covid has gone away.
And that there exists, or at least could exist, some means of avoiding the ill effects of Covid.
Clearly Covid cures nothing.
And the effects can only be mitigated, not avoided.
I’m hoping that the people who pushed for lockdown (because of lack of critical thinking, or their own political or control freakery reasons) are now beginning to feel the personal distress from their great achievement. When the weather was cold and working from home was a novelty it was all great fun for a couple of weeks. But this weekend really would have been a good one for getting together with friends and sitting outside a pub or whatever. Maybe they’re now thinking of the cancelled holidays, festivals and so on and not feeling quite so ‘fascistic’ in their outlook. A few of them, surely, must be single red-blooded males and looking at the prospect of not having s*x for the next eighteen months (you’d have to have a heart of stone not to laugh, etc.).
Early on in this crisis, I was looking for reasons to believe that it wasn’t as bad as was being made out – and I think those reasons are, in fact, real. For some reason, 86% of the population didn’t think the same way. Either they are incapable of looking past the image created by the media or they see some professional or ideological benefit in milking the crisis. I am hoping that enough of them are now in so much ‘distress’ that the tide may turn quite quickly. If it doesn’t, I am still going to enjoy the thought that some of them are in more distress than me.
Let me make it clear: my points about Covid-19 have nothing whatsoever to do with philosophical debates about how to understand probability talk. Frequentist v epistemic v logicist debates are completely irrelevant here. It’s the probability of x we need to be thinking about, not the issue of what ‘the probability of x’ really means.
I don’t know why you’ve brought this up. If someone tells you we have to be worried about a low probability event, you look into how probable that event really is by looking at evidence that will help you get a grip on that probability. You don’t read up on the philosophy of probability.
Also, ‘qualitative uncertain benefits’ doesn’t strike me as a particularly helpful phrase anyway.
Health Service Jouranl article “Coronavirus response could create very serious unintended consequences”:
https://www.hsj.co.uk/policy-and-regulation/coronavirus-response-could-create-very-serious-unintended-consequences/7027321.article
500k with NO MITIGATION
The experts (Imperial college) presented the outcomes for a range of strategy options. 20k relates to our current lockdown strategy.
“(Nor do we have a 25% chance of there being 400 000 deaths in this scenario — not even Ferguson has discussed that — so you can’t try to approach it from that angle).”
I think you’re wrong about this. You seem to be arguing that Ferguson hasn’t mentioned any potential death counts above 250k – therefore 250k must be the maximum his model allows – therefore any uncertainty about the model must be in the downward direction.
But this is like saying “My friend has invited 1000 people to his birthday party, and he says he thinks about 250 people will come – but I only have 50% confidence in his estimate, so we can stock the bar as if 125 people will come.” Obviously that wouldn’t make sense! If you only have 50% confidence in the estimate, that means your forecast should expand out in both directions, both higher and lower.
And indeed, it seems likely to me that Ferguson’s model gives a range of possibilities from (say) 100k to 400k, but assigns the highest probability to something roughly in the 250k area, so that’s what he talks about in public.
You may say “Well, the reason we know it can’t be higher than 250k but it could be lower is because of scientific reasons X, Y and Z.” But that is simply its own form of modelling! You have attempted to constrain Ferguson’s model for likely deaths with your own DIY model for maximum deaths.
In which case, either you acknowledge that YOUR model is also only 50% certain, meaning we are back where we started – with a range of probabilities which allows for numbers both higher and lower than 250k (and for all we know Ferguson incorporated X, Y and Z into the premises of his own model, so your intervention may be duplicative and redundant). Or you insist that unlike Ferguson’s model, your model is 80, 90, 100% certain, in which case you have to explain to the skeptical reader why your reasoning is so much more airtight.
Has anyone here followed Monckton of Brenchley’s (yes, that Lord Monckton) pieces at Watts Up With That? He’s a real lockdown fan and is now running a series with graphs purporting to show the lockdowns working. A far as I can tell his graphs show the UK lockdown working before it was introduced.
https://wattsupwiththat.com/2020/04/11/coronavirus-good-news-the-downtrend-in-chinese-virus-case-growth-rates-continues/
It’s probably not worth commenting there as it’s massively attended, and he’s a nasty piece of work anyway, using every underhand debate technique to dismiss critics.
Behave, BiG, I did apologise!
Early on,the next-door NHS Trust to where I live announced that, in addition to stopping all planned surgery, it was also suspending cancer treatment. According to the BBC site that day just over one hundred positives had been recorded in the relevant areas – that is one hundred positive tests, not one hundred people needing hospital treatment, much less occupying ICU beds. I haven’t been following the situation as closely as I should – the apoplexy-inducing fear porn abounds and is unbearable – but it’s inconceivable that some people haven’t been very adversely affected to say the least. If I lived in the area and had a loved one suffering from cancer and the worst happened I’d want ‘Neglect due to pre-emptive hysteria’ on the death certificate.
Thanks Adele, don’t know if you’ve seen this but interesting update on the ‘international situation’. Updated every few days
https://swprs.org/a-swiss-doctor-on-covid-19/
*The idea that we can be that confident about 200 000 deaths on the basis of some black-box outlier computer models from people with no track record for reliability,*
According to the paper on mathematical modelling re 2001 Foot and Mouth posted on here recently computer models have no predictive value whatever in the field of public health.
There’s no reason to suppose that the spread of a virus is any more or less predictable than say a football game or horse race; that an experienced virologist / tipster couldn’t be a more reliable guide than any mathematical model. Based on the evidence we should place no more faith in the nostrums of the modelling priesthood than in professional betting tipsters.
It seems that the modellers status is based on their political rather than their truth value. How otherwise could someone with a track record like Ferguson’s still be given credence at all never mind being a guide to public policy?
The so-called consensus over long term weather predictions is often cited as a reason for supposing them to be true. But why shouldn’t official unanimity rather invite suspicion when those who are otherwise engaged in political rivalry find themselves in agreement, especially on something as inherently tenuous or “stochastic” as a weather forecast? – suspicion of officaldom closing ranks against the public, that is.
It’s no different with the virus, only that in the currency of fear the virus has higher value than ‘climate’. Thus political advantage to be had in ramping up the fear, in accusing one’s rivals of being fear-deficient, given that if there’s anything whose spread is “exponential”, it’s fear itself. In absence of strong leadership in defence of “things that are established” you can’t lose politically.
The basic dynamic has never been put better than this which I’ve just copied and pasted from pdf of The Meaning of Conservatism by Roger Scruton to which it is the epigraph:
“He that goeth about to persuade a multitude, that they are not so well governed as they ought to be, shall never want attentive and favourable hearers; because they know the manifold defects whereunto every kind of regiment is subject, but the secret lets and difficulties, which in public proceedings are innumerable and inevitable, they have not ordinarily the judgment to consider. And because such as openly reprove supposed disorders of state are taken for principal friends to the common benefit of all, and for men that carry singular freedom of mind; under this fair and plausible colour whatsoever they utter passeth for good and current. That which wanteth in the weight of their speech, is supplied by the aptness of men’s minds to accept and believe it. Whereas on the other side, if we maintain things that are established, we have not only to strive with a number of heavy prejudices deeply rooted in the hearts of men, who think that herein we serve the time… but also to bear such exceptions as minds so averted beforehand usually take against that which they are loth should be poured into them.
– Hooker, Of the Laws of Ecclesiastical Polity, Book I, Chapter i”
He’s quite the authoritarian manqué and has form. During the AIDS crisis he argued for monthly testing of the entire population and ‘isolation’ of anyone HIV-positive. Not someone I’d care to turn my back on.
Don’t overlook all the deaths averted by patients avoiding medical treatment.
https://www.hopkinsmedicine.org/news/media/releases/study_suggests_medical_errors_now_third_leading_cause_of_death_in_the_us
Sweden wobbling?
https://www.dailymail.co.uk/news/article-8212365/Swedish-PM-tighten-coronavirus-restrictions-Austria-Denmark-begin-easing-lockdown.html
Hopefully just window dressing to fend off the mob.
So, I have a telephone appointment tomorrow with rheumatoid dept at UHCW. I suspect I have RA. Nice time to have an auto immune disease! Do I stay off any medication they may prescribe and take my chances? Or is there an RA Medicine out there which won’t put me at more risk from Covid 19? Decisions decisions…
This Telegraph article seems to me to be a brilliant example of all that’s bad in science and journalism. Great piles of numbers being waved around by people who know nothing about the difference between correlation and causation and have more axes to grind than a lumberjack convention. A collection of national statistics, quickly extracted from whichever reference source is handy, glued together with observations which are either pointless, irrelevant or just plain weird. And then, and then, the blindingly obvious conclusion, “Until this crisis is over it is impossible to know where Britain will fall compared to other countries in terms of mortality …….”.
What am I missing? I thought we had whole university departments spending substantial amounts of time and money understanding these problems not planning more exciting ways of scoring points over their rivals. In my obvious naivety, I imagined we had a solid emergency planning system that spent my money on wisely preparing plans for dealing with all the nasty things that might unknit our society. Did you know that the Civil Contingencies Act requires the NHS to make comprehensive plans to ensure “mandatory disclosures on climate change adaptation and mitigation”. I feel safer already.
Putting aside my unease about possibly being very ill or dying earlier than I’d planned, I feel cheated by those that are supposed to look after me. I know it’s my fault really, I should have kept a better eye on what’s going on in the world. I pay for the fire brigade but I have fire extinguishers in my house. I pay for the police and also buy locks and barbed wire. But do I now need to keep a couple of spare ventilators under the stairs in case the NHS gets short? Or have a few months stored supply of my prescription drugs ?(actually I have always done that). Doesn’t anyone actually do any emergency planning any more? Or learn from the past? It’s not as if we’ve been attacked by interstellar aliens with completely unknown weapons that choke me with my own moustache; we are in thrall to possibly somewhat more lethal versions of last year’s problems. I expected someone to pull the “dealing with a dangerous infectious disease pandemic” plans out of the vault and what actually appeared is advice on how to make sure that ethnic and gender diversity guidelines are being maintained during emergency treatment.
I just wish we could get value for money from experts for a change.
Thanks for the non-herd comments, Mr Drummond and the ability to change your mind in the face of new evidence.
“To listen to the news you would think that every problem apart from Covid has gone away”
I’m almost looking forward to hearing about a new war breaking out somewhere – it would make a change…
Hector, I was simply trying to be helpful with the difficult topic of messaging. It seems obvious to some of us that the lockdowns are a governmental over-reaction. Why is it not obvious to everyone?
It seems that different people have different concepts of probability. The probability of winning the lottery is so low as to be insignificantly trivial — yet millions of people slap down their money each week.
If my suggested alternative is not helpful, no problem. I will keep scratching my head about whether there may be other more effective ways to get people to stop & think.
“It’s possible that people are therefore staying away from hospital then getting even sicker as a result, but there is not the data available yet to prove this hypothesis.”
Are they “staying away” because they’ve been 1) Constantly told to remain at home, 2) Assuming the NHS is completely inundated, or 3) Worried about attending for routine appointments and/or surgery, only to catch CV and potentially suffer far worse consequences?
Dave G: “In my obvious naivety, I imagined we had a solid emergency planning system that spent my money on wisely preparing plans for dealing with all the nasty things that might unknit our society.”
Yes, we are all learning how short-sighted our Betters are. But perhaps this suggests a way forward for us peons. The answer to a Big Fear is — an even Bigger Fear!
Instead of trying to convince our fellow citizens that LockDowns are a damaging over-reaction to a relatively minor viral outbreak, we should be focusing on alarming them about something very much worse — something so devastating that it demands we all get back to work immediately and start preparing.
My nomination for this Bigger Fear is a Carrington Event — where a mass ejected from the Sun hits the Earth like a giant Electro-Magnetic Pulse weapon and fries every piece of electrical equipment on the planet — from your cell phone to the neighborhood nuclear power plant. Talk about scary!
It has happened before, most recently in 1859 — before we all became so dependent on electricity. Those poor C-19 patients on ventilators will all die if the electricity goes out! Fear! Fear! Fear!
We need to save the human race by preparing now! Shield key equipment, build back-up offline power plants, and so much more. We need to get everyone back to work right now!
Okay, I just want to keep the idea of probability simple, without introducing sophisticated philosophical discussions about what probability is.
Also, how we understand the *concept* of probability is very different to making probability estimates. People who buy lottery tickets don’t necessarily have a different concept of probability than other people, they just don’t estimate the probabilities very well. Or they think that despite the low probabilities, it’s still worth it in terms of expected utility (it’s not, but that’s not quite the same as getting the probabilities wrong.) Or they just think it’s a bit of fun, and a chance to dream. Or they indulge in a bit of magical thinking, and think that something in the Universe will eventually reward them. Which is irrational, sure, but that’s not a failure to understand probability.
Thank you Tony! It’s all gone a bit mad hasn’t it?! 🤷♀️
There’s lots of other bigger fears: gamma ray burst hitting earth, super volcano eruption, large asteroid hitting earth etc!
I will shut up after this, but do want to thank you Prof. Drummond for stimulating some of the neurons in my atrophying brain into firing.
The issue is Messaging — how to get people to recognize lockdowns as an over-reaction to this particular threat?
People willingly play the lottery although the probability of winning is insignificant. Perhaps the internal rationale is that the cost to play is small and the potential payout would be life-changing. Would people be so interested in playing the lottery if a ticket cost $1,000 instead of $1 ?
It may be that some of those who are keen on lockdowns see them as sort of a reverse lottery — where the cost of the lockdown to me is small, but the potential negative outcome of my dying is unacceptable, no matter how small that probability may be.
If that hypothesis is correct, then the most effective form of Messaging would be to emphasize the high cost of playing in this lockdown reverse lottery. Lost income; depleted savings; debts; collateral damage from the difficulty of getting medical attention for normal problems; etc.
The cost of a short lockdown may seem small, but it will hopefully start seeming not so small, in fact, pretty damn big, as it goes on and the economy tanks and they lose their jobs.
Problem is, because of the media they’re still comparing that to what they think will be a health apocalypse.
Many thanks for your data and analysis relating to coronavirus, much appreciated. Having read the articles on the lockdown by Toby Young and Sam Bowman with interest I had a go at a cost–benefit analysis myself.
https://medium.com/@martinvsewell/the-coronavirus-lockdown-in-the-united-kingdom-a-cost-benefit-analysis-acd19f635dae
There is not the slightest evidence that 500k would die without mitigation and there is not any chance in the world anyone could revise that number to 20k after ONE day of a ‘lockdown’ that has no historical reference for performance.
It’s bullshit.
Comments are closed.