Dumbledore's Army

Karma: 694

Dumbledore's Army 20 Feb 2021 18:48 UTC
7 points
on: Overconfidence is Deceit
Thank you for an interesting article. It helped clarify some things I’ve been thinking about. The question I’m left with is: how practically can someone encourage a culture to be less rewarding of overconfidence?
I guess I’m feeling this particularly strongly because in the last. year I started a new job in a company much more tolerant of overconfidence than my previous employer. I’ve recalibrated my communications with colleagues to the level that is normal for my new employer, but it makes me uncomfortable (my job is to make investment recommendations, and I feel like I’m not adequately communicating risks to my colleagues, because if I do no-one will take up my recommendations, they’ll buy riskier things which are pitched with greater confidence by other analysts). Other than making sure I’m the least-bad offender consistent with actually being listened to, is there something I can do to shift the culture?
And please, no recommendations on the lines of ‘find another job’, that’s not practical right now.

Dumbledore's Army 21 Feb 2021 10:33 UTC
14 points
in reply to: Ben Pace’s comment on: Overconfidence is Deceit
Does anyone have a clear example to give of a time/space where overconfidence seems to them to be doing a lot of harm? I would say making investments in general (I am a professional investment analyst.) This is an area where lots of people are making decisions under uncertainty, and overconfidence can cost everyone a lot of money.
One example would be bank risk modelling pre-2008: ‘our VAR model says that 99.9% of the time we won’t lose more than X’ therefore this bank is well-capitalised. Everyone was overconfident that the models were correct, they weren’t, chaos ensued. (I remember the risk manager of one bank—Goldman Sachs? - bewailing that they had just experienced a 26-standard deviation event, which is basically impossible. No mate, your models were wrong, and you should have known better because financial systems have crises every decade or two.)
Speaking from personal experience, I’d say a frequent failure-mode is excessive belief in modelling. Sometimes it comes from the model-builder: ‘this model is the best model it can be, I’ve spent lots of time and effort tinkering with it, therefore the model must be right’. Sometimes it’s because the model-builder understands that the model is flawed, but is willing to overstate their confidence in the results, and/or the person receiving the communication doesn’t want to listen to that uncertainty.
While my personal experience is mostly around people (including myself) building financial models, I suggest that people building any model of some dynamic system that is not fully understood are likely to suffer the same failure-mode: at some point down the line someone gets very over-confident and starts thinking that the model is right, or at least everyone forgets to explore the possibility that the model is wrong. When those models are used to make decisions with real-life consequences (think epidemiology models in 2020), there is a risk of getting things very wrong, when people start acting on the basis that the model is the reality.
Which brings me on to my second example, which will be more controversial than the first one, so sorry about that. In March 2020, Imperial College released a model predicting an extraordinary death toll if countries didn’t lock down to control Covid. I can’t speak to Imperial’s internal calibration, but the communication to politicians and the public definitely seems to have suffered from over-confidence. The forecasts of a very high death toll pushed governments around the world, including the UK (where I live) into strict lockdowns. Remember that lockdowns themselves are very damaging: mass deprivation of liberty, mass unemployment, stoking a mental health pandemic, depriving children of education—the harms caused by lockdowns will still be with us for decades to come. You need a really strong reason to impose one.
And yet, the one counterfactual we have, Sweden, suggests that Imperial College’s model was wrong by an order of magnitude. When the model was applied to Sweden (link below), it suggested a death toll of 96,000 by 1 July 2020 with no mitigation, or half that level with more aggressive social distancing. Actual reported Covid deaths in Sweden by 1 July were 5,500 (second link below).
So it’s my contention—and I’m aware it’s a controversial view—that overconfidence in the output of an epidemiological model has resulted in strict lockdowns which are a disaster for human welfare and which in themselves do far more harm than they prevent. (This is not an argument for doing nothing: it’s an argument for carefully calibrating a response to try and save the most lives for the least collateral damage.)
Imperial model applied to Sweden: https://www.medrxiv.org/content/10.1101/2020.04.11.20062133v1.full.pdf
Covid deaths in Sweden by date: https://www.statista.com/statistics/1105753/cumulative-coronavirus-deaths-in-sweden/

Dumbledore's Army 21 Feb 2021 16:56 UTC
3 points
in reply to: Ben Pace’s comment on: Overconfidence is Deceit
Thank you!

Dumbledore's Army 22 Feb 2021 9:55 UTC
10 points
in reply to: Ben Pace’s comment on: Overconfidence is Deceit
I was thinking about this a little more, and I think that the difference in our perspectives is that you approached the topic from the point of view of individual psychology, while I (perhaps wrongly) interpreted Duncan’s original post as being about group decision-making. From an individual point of view, I get where you’re coming from, and I would agree that many people need to be more confident rather than less.
But applied to group decision-making, I think the situation is very different. I’ll admit I don’t have hard data on this, but from life experience and anecdotes of others, I would support the claim that most groups are too swayed by the apparent confidence of the person presenting a recommendation/pitch/whatever, and therefore that most groups make sub-optimal decisions because of it. (I think this is also why Duncan somewhat elides the difference between individuals who are genuinely over-confident about their beliefs, and individuals who are deliberately projecting overconfidence: from the point of view of the group listening to them, it looks the same.)
Since groups make a very large number of decisions (in business contexts, in NGOs, in academic research, in regulatory contexts...) I think this is a widespread problem and it’s useful to ask ourselves how to reduce the bias toward over-confidence in group decision-making.

Dumbledore's Army 3 Mar 2021 9:52 UTC
6 points
in reply to: River’s comment on: We should not raise awareness
I think the issue is that ‘raising awareness’ is used to mean three separate things. (I agree that the extra simulacra levels aren’t a helpful explanation.) Using awareness of breast cancer as a reasonably non-controversial example.
1. Give people some useful knowledge or skill to help reduce the problem. Eg teach women how to examine themselves for lumps and when to seek medical advice.
2. Raise the profile of the issue (or in some cases inform people that the issue exists) such that more resources will be devoted to solving it. Eg publishing opinion pieces informing the public how many people are affected by cancer and calling for more facilities for treatment, or encouraging people to donate money to charities researching cures.
3. Signal that you are a virtuous person who is concerned about socially-approved causes. Eg wear a pink ribbon, or like Facebook pages from breast cancer charities.
I agree that there are a lot of people practicing virtue-signalling, while kidding themselves that they are doing level 2 profile-raising, and I also agree that a lot of the profile-raising is transparently ineffective. But I think that there are useful level-1 activities which also come under the banner of ‘raising awareness’ and I wouldn’t want to stigmatise those.
There are also some situations in which the level-2 activities are useful. I suspect you would disagree, but I think sexual assault is a fairly good example: a lot of people have gone to great efforts to explain to the general public that there is a widespread problem that needs action. The result has been an in-progress and partial change in social norms which may well succeed in reducing the levels of sexual assault.

Dumbledore's Army 7 Apr 2021 13:42 UTC
13 points
on: Monastery and Throne
Hi Jacob,
I think your Seeing the Smoke was interesting and the conclusions about human nature are right—particularly the point that most people will fail to do the obvious thing like leave the smoke-filled room out of fear of looking weird. That said, I really wish Cummings had drawn a different conclusion from your blogpost, because I strongly believe that lockdowns were the wrong response to Covid. Specifically, I would prefer Cummings had read Hans Rosling’s excellent book Factfulness, especially chapter 10 on the urgency instinct:
Rosling was investigating a disease in a remote area of Mozambique. The mayor of the nearest city asked Rosling if he should institute a roadblock to prevent sick people from coming to the city. Rosling wasn’t even sure the disease was contagious, but he thought better safe than sorry and said yes, put up the roadblock. When the village women, some of them carrying infants, found the road to market blocked they asked local fishermen to carry them instead. The overloaded boats sank causing the death of about 20 people including the infants. Rosling later found that the disease was caused by eating improperly-prepared cassava, and wasn’t contagious at all, meaning that the roadblock had killed those people for no reason.
Rosling blamed himself for those deaths for years afterwards. As he puts it: “Back in Nacala in 1981, I spent several days carefully investigating the disease but less than a minute of thinking about the consequences of closing the road. Urgency, fear, and a single-minded focus on the risks of the pandemic shut down my ability to think things through. In the rush to do something, I did something terrible.” [My bold.]
I believe that the decision to lockdown followed the exact same error mode as Rosling’s roadblock. Policy-makers got tunnel vision and focused on one thing—the pandemic—and ignored all other consequences including a mental health epidemic, mass unemployment/furlough, serious harm to children’s development, and many other issues. More generally, I suggest that the typical failure mode is to over-react to any threat which is reported in the media (a combination of the availability heuristic making that threat more salient, and asymmetric justice meaning that politicians are more scared of being accused of inaction than they are of wrong action.) The failure mode can be summarised as: we must do something, this is something, therefore we must do it.
Pretty much everyone who has tried to calculate a cost-benefit analysis of lockdowns finds that they cost more years of life than they save. In developed countries, the typical estimate is 10x. (For example, Prof Philip Thomas of Bristol Uni estimating for the UK, or Dr Ari Joffe estimating for Canada.) For developing countries with poorer, younger populations, the ratio is worse. A multi-disciplinary South African team estimated that their lockdown would cost 29x more years of life than it saved.
Those estimates all assume that lockdowns are highly effective at preventing Covid deaths, which was intuitive as of March 2020. However, reality is not always intuitive and the experience of places like Sweden, and those US states which didn’t have winter lockdowns, suggests that the actual benefits of lockdowns are far smaller than thought. Studies like the Stanford University one find small-to-zero benefits of lockdowns vs less-restrictive measures.
For those of us who are more interested in physical truth than in the socially-constructed narrative, we need to start persuading the wider world that lockdowns do more harm than good. I hope you and the rest of our virtual monastery will join me in this effort.

Dumbledore's Army 9 Apr 2021 7:41 UTC
0 points
in reply to: Rob Bensinger’s comment on: Monastery and Throne
Closure of schools. There’s a mountain of evidence that taking kids out of school is harmful. It’s not just the loss of education—although that doesn’t help—but also the loss of socialisation. Less education is directly correlated with shorter life expectancy—a US study found that just that effect was enough to mean that closing schools would cost more years of life than it saved, with 98% probability. That’s before adding in the burden from significantly higher rates of mental health problems in children who have been deprived of school.
Close of schools is disproportionately harmful to those who already come from deprived backgrounds—think about the difference between a middle-class family where every child has their own iPad and the educated parents will help with homework, compared to a lower-class family which has one phone to share among everyone and the single-parent doesn’t have time to help kids and also work. Then consider that closure of schools means loss of free school meal schemes—this caused chaos and serious hardship even in the UK and will have been worse in less developed countries.
Then there’s the extreme cases: school is an escape for kids who live in homes where there’s domestic violence. Teachers can also look for signs that kids need help or are suffering for abuse—if they’re physically present. Closure of schools means that kids in abusive situations are trapped ²⁴⁄₇ with their abusers—whose own behaviour may become worse due to stress of unemployment or isolation.

[Question] What weird beliefs do you have?

Dumbledore's Army14 Apr 2021 10:29 UTC

16 points

77 comments1 min readLW link

Dumbledore's Army 17 Apr 2021 14:05 UTC
18 points
in reply to: Gerald Monroe’s comment on: Why has nuclear power been a flop?
What is your definition of contaminate? If Devanney is correct that low doses of radiation are acceptable—and I believe he is—then much land which is described as ‘contaminated’ is in fact perfectly liveable. (Also see the people who illegally live in the Chernobyl exclusion zone). For a reasonable definition of ’contaminate’ then, it follows that a nuclear accident contaminates much smaller areas of land and is less expensive.
Your anti-nuclear argument also ignores the status quo of non nuclear energy. In America alone, fossil fuels (read coal) kill tens of thousands every year. So if you replaced all coal power with nuclear and had a Chernobyl every year (unrealistic extreme scenario), it would still save lives on net.
That said, I can see the argument that renewables are safer than both today, but OP is absolutely right to analyse the decades-long failure to replace coal with nuclear in the period before we had renewables.

Dumbledore's Army 18 Apr 2021 21:01 UTC
2 points
in reply to: Razied’s comment on: Anger
To expand on your last sentence, anger can be a driver of positive change in the world. Greta Thunberg is angry that people are carelessly wrecking the only planet we have to live on. Racial justice protesters in the US are angry that black people keep getting killed by the police. Unless you’re a saint, being furious about some injustice is much more motivating than the dispassionate thought that ‘x would be a good deed’.
Having said that, I would agree with OP that most of the time in most interpersonal situations anger is damaging, and for most people becoming less angry is a good thing. (Or at least many people should become much more aware about why they are angry, at whom, instead of letting themselves be generically angry and taking it out on the nearest available target.)

Dumbledore's Army 22 Apr 2021 8:50 UTC
1 point
in reply to: mukashi’s comment on: Monastery and Throne
Lockdowns are more harmful than beneficial with the few exceptions of those countries like New Zealand that successfully kept the virus out. For any country where the virus is already endemic, the damage done by lockdowns was immense, and the benefits relatively limited. Remember that the counterfactual is not ‘do nothing’. It’s ‘enact some more reasonable set of restrictions’.
Prof Douglas Allen of SFU just did a really good takedown of bad arguments in favour of lockdown. In his most unrealistic extreme scenario intended to steelman the pro-lockdown case, he finds that lockdowns cost 3.6x more than their benefits, at the opposite end of the spectrum, they might have cost as much as 282x more than their benefits. (His figures for Canada, but the argument should generalise to any developed country.)
This image is a good example of how distorted pro-lockdown arguments are. It’s taken from Neil Ferguson’s Imperial College model used to argue for lockdown. Pro-lockdown cost-benefit analyses generally compare the blue line below (full lockdown) with the black line (do nothing) for an estimate of 120 lives saved per 100,000 people. It would be far more reasonable to compare that blue line to the brown line below (assumes case isolation and household quarantine but not lockdown measures) which immediately halves your assumed benefit for lockdown measures. Then remember that the Imperial College model is grossly overstated and assumes no public behaviour change in the absence of a lockdown. You get the picture.
I really do encourage you to read the whole study: http://www.sfu.ca/~allen/LockdownReport.pdf

Dumbledore's Army 24 Apr 2021 9:47 UTC
2 points
in reply to: Lukas_Gloor’s comment on: Monastery and Throne
I see a lot of nit-picking of my evidence, but you have provided zero support for your own claim that lockdowns do more good than harm. I challenge you to come up with a published cost-benefit analysis that proves the same.
What would a good cost-benefit analysis include? There are a lot of harms caused by lockdowns. Some of them are difficult to quantify (eg my last point), but I think it’s reasonable to demand a cost-benefit analysis takes into account at least three of the following six harms (which are far from an exhaustive list):
- Increased poverty is directly correlated to lower life expectancy, so we should measure the lost years of life from increased poverty. This is a very long-term effect, which will be doing harm for years to come.
- Unemployment and financial problems are both ‘scarring’ (it takes a long time to dig yourself out) and both cause mental health effects. We should include the long-term mental health cost of increased unemployment and individual financial problems.
- Where lockdowns include school closures, the long-term effect on children’s development and socialisation is extreme. We should include estimates of the lifetime impact on children, which will include shorter life expectancy. (Yes, shorter educations are correlated to future life expectancy.)
- Additional mental health effects directly attributable to the lockdown including elevated rates of depression, stress and anxiety. Proper acknowledgement that some of these effects (eg increased alcoholism) have long-term effects which should be accounted for.
- Lockdown sucks for everyone, even if still employed, not in school, and not suffering a formal mental health condition. We should acknowledge that this is a widespread disutility and deserves to be considered.
- Lockdowns have set a deeply disturbing precedent that governments can remove almost all civil liberties when they declare there’s an emergency. This directly harms democracy and raises the risk of future loss of freedoms. (Yes this one is the hardest to quantify, but it’s important and I would like to see more people acknowledge it.)
An ideal cost-benefit analysis would acknowledge that benefits from lockdowns in terms of lives saved are uncertain and include a range of estimates, but if you find one that properly considers at least 3 of the above 6 points, I’ll accept it even if it has a point estimate for benefits. (Those using the initial Imperial College models should be at the top end of the range because the Imperial College figures are too high for all the reasons I’ve already said.) However, since you’ve said that “It’s a strawman that policymakers compare lockdown to “do nothing.”″ then I do expect your superior cost benefit analysis will compare lockdown to more reasonable restrictions, rather than a do nothing option.
So there it is. I challenge you: bring evidence or go home.
PS: If I seem to be beating the drum of long-term effects too hard, it’s because I’m still angry at the UK government’s (belated, poor-quality) excuse for a cost-benefit analysis which, among its many other failures, looked only at harm done over the next five years.

Dumbledore's Army 24 Apr 2021 9:51 UTC
2 points
in reply to: Lukas_Gloor’s comment on: Monastery and Throne
Your arguments about health care systems collapsing in the absence of lockdowns are highly emotive but not factual. The experience from Sweden, from Serbia, from the US states which didn’t do second lockdowns, is that the health care systems didn’t collapse. You are arguing from the unexamined assumption that no lockdown means failed health care system and your assumption is empirically, provably false. Please update your views accordingly.

Dumbledore's Army 15 May 2021 19:30 UTC
3 points
on: Covid 4/9: Another Vaccine Passport Objection
“What’s scary is when I try to put a probability on ‘this turns into a broader thing that outlives Covid and people have to show the computer has them in good standing to get on a train’ or something like that. I still think it’s a super low probability thing, maybe 2% (it’s a Fermi answer, but a considered one), but that’s 2% of a really, really bad outcome.”
I’m sorry, but the 2% strikes me as pure optimism. If the government gets vaccine passports approved for anything, the probability is very high that we’re stuck with them for the long haul, and that use will gradually expand over time from slippery slope effects.
As a comparison, think about all the civil liberties lost after 9/11 and how you never got most of them back. Eliezer Yudkowski wrote that it was a vast underestimate to think that the overreaction to 9/11 would be ten times worse than the event itself. Every bias that caused the overreaction to terrorism is also present and causing the overreaction to Covid. Vaccine passports are part of that overreaction!
Think about how governments learned to shout ‘terrorism!’ any time they wanted more power and how those powers always ended up being used against more than just terrorists. In the US you guys had the NSA snooping on everything and everyone, terrorists or no. Here in the UK one very sinister example is the 2001 Anti-Terrorist Act which our then-government solemnly promised would only be used against terrorists. Within a few years it was being used to seize assets from Icelandic banks for the crime of being inconveniently bankrupt. You simply cannot trust governments to only use their new powers against the original target. Once you add a tool to their toolbox, that tool gets used.
Now really think about what it would take to be the politician or bureaucrat that tried to scrap vaccine passports. You’re up against fear of blame and the Copenhagen interpretation of ethics: whoever makes that decision is at risk of being blamed for every person who ever gets sick again. They will be demonised by the other side as pro-virus. Then there’s simple status quo bias: once the passports exist scrapping them then becomes a change in the status quo and subject to inertia, akrasia, and the difficulty of coming to consensus.
I hope you now understand why I am fundamentally opposed to vaccine passports. I hope I have persuaded you too. This is a terrible, dangerous idea, and the time to stop them is before they become the status quo.

Dumbledore's Army 7 Jun 2021 20:39 UTC
23 points
0
in reply to: Rafael Harth’s comment on: We need a standard set of community advice for how to financially prepare for AGI
Your question ignores timeframes. I’m happy to argue that P(stock rises in the next 5 years |AGI in 20 years)≈P(stock rises in the next 5 years) for all stocks.

I’m a professional equity investor, and trust me, the market isn’t that forward-looking. Unless you believe in AGI within the next 10 years, I suggest ignoring it when it comes to picking investments. Because for the intermediate timeframe until the market begins to take the concept seriously, the value of your investments will be determined by all the other factors which you’re ignoring in favour of focusing on AGI, so unless you want your investment results to be meh for years-to-decades, then don‘t go for some all-out bet on AI.

Dumbledore's Army 7 Sep 2021 9:08 UTC
10 points
on: I read “White Fragility” so you don’t have to (but maybe you should)
Thank you for the write-up. However, I profoundly disagree with the premises of White Fragility. Why? Because DiAngelo’s starting premise is that race is and always must be really important. And I object to that. In fact, I think that ‘believing race is something real and important’ is a necessary precursor to racism.
(And yes, I believe her book to be racist against white people and yes, anti-white racism is a thing. It’s true that in the US context white racism against non-white people is more frequent and often more harmful, but that doesn’t make anti-white racism OK. It’s like saying that men often sexually assault women, so if a woman sexually assaults a man that’s totally fine because women are an oppressed demographic. Nope, that is not how ethics works.)
So what do I believe? I think that a person’s skin colour should be seen as a minor physical detail of no importance. I believe in Martin Luther King’s dream:
“I have a dream that one day this nation will rise up and live out the true meaning of its creed. We hold these truths to be self-evident that all men are created equal.
I have a dream that one day out in the red hills of Georgia the sons of former slaves and the sons of former slaveowners will be able to sit down together at the table of brotherhood.
I have a dream that one day even the state of Mississippi, a state sweltering with the heat of oppression, will be transformed into an oasis of freedom and justice.
I have a dream that my four little children will one day live in a nation where they will not be judged by the color of their skin but by their character.”
A world where people are judged not by the colour of their skin but by their character. Imagine that. I know that we are far from that world today, but it seems to me that our efforts should be directed at getting closer to MLK’s dream world. And that books like DiAngelo’s, however well-intentioned, which demand that we see skin colour as really important, are in fact taking us further away from the ideal.

Dumbledore's Army 7 Sep 2021 14:50 UTC
6 points
in reply to: TekhneMakre’s comment on: I read “White Fragility” so you don’t have to (but maybe you should)
I think your comment assumes bad faith. You assume that the right hand will beat down the black man. That is wrong and discourteous.
You also ignore the point about the direction of travel: we should move toward a world where race is less important, not toward a world where race is more important. We already know what happens if you go the other way. There are places which have entrenched politics based on each racial group getting an entitlement, like Malaysia and Lebanon (the Lebanese system has religious groups instead). It’s a stable equilibrium and it’s a bad equilibrium. Those are not happy countries. The USA doesn’t want to end up like them.

Dumbledore's Army 7 Sep 2021 22:13 UTC
8 points
in reply to: Viliam’s comment on: I read “White Fragility” so you don’t have to (but maybe you should)
Thank you for a thoughtful response. I agree that the current situation in the US is very far from the desired end state and I don’t want to deny the real problems that exist. I’m just deeply concerned that the proposed ‘cure’ is a new permutation of the same disease.

Edited to add: I would also argue that proposing a direction of travel toward the ideal world is not the same as pretending we’re already there.

Dumbledore's Army 7 Sep 2021 22:20 UTC
8 points
in reply to: Dumbledore's Army’s comment on: I read “White Fragility” so you don’t have to (but maybe you should)
I also think that moving the lens of public attention away from racism would make it easier to try solutions to other problems that disproportionately affect black people. As just one example, the system of funding schools from local property taxes means that affluent areas have nice well-funded schools and impoverished areas generally don’t. Systemic reform of school funding to be more equal per child would improve the education available to disadvantaged children (who are disproportionately black). It’s an example of a race-blind policy change that would improve racial justice, and it won’t get attention while everyone is yelling about racism. And yet, even in a world with zero racism (which is not our world), kids born in a poverty trap will have difficulty getting out of the trap and if those kids start off disproportionately black then you will get a situation of ongoing racial disparity in outcomes.

Tl;dr if we spent less time thinking about racism and more time on effective ways to alleviate disadvantage you would get a better and fairer world and also one with less racial disparity.

Dumbledore's Army 9 Sep 2021 5:49 UTC
2 points
in reply to: Brendan Long’s comment on: I read “White Fragility” so you don’t have to (but maybe you should)
Ok thanks for the correction. I’ll pick a different example next time.

Dumbledore's Army

[Question] What weird be­liefs do you have?

[Question] What weird beliefs do you have?