@(InsanityCheck) Yup. There's this theory going around that there's a giant iceberg of asymptomatic carriers who get missed in the testing, and that existing IFR estimates are too high by a factor 5x or more. Ioannidis has jumped on this bandwagon and even some epidemiologists with decent credentials. And a ton of wannabe epidemiologists with economist backgrounds and naive takes have also jumped on this. Maybe they're seeing something others are missing, but to me this whole thing seems bizarre like some runaway process where people lend each other cred...
@(Linch) according to the graphs shared on this tweet (https://twitter.com/jburnmurdoch/status/1254484489067167745), excess death in Guayas province, Ecuador, are up by 10,000. I think excess deaths are a good indicator of how many people died of Covid-19 (this is worth flagging as an additional premise). According to Worldometer, Ecuador has a median age of 27.9 compared to the UK's median age of 40.5. Guayas has about 3M inhabitants and I imagine the capital Guayaquil with 2M inhabitants to be particularly badly hit. The IFR is already 0.33% with those...
I'd be interested in a question about the CFR for young-ish healthy people who contract the disease. It seems to me that there's very little information about victim demographics so far. Initially, it was reported that the disease mostly kills elderly people or people with pre-existing health conditions, but there were I think 1-2 exceptions among the first couple dozens of deaths (not totally sure about this). And the famous whistleblower doctor who died is said to have been healthy, and 34 years old. If it takes younger and healthy people longer to die...
@(Tamay) I wrote some comments about their page here: https://pandemic.metaculus.com/questions/3755/what-will-be-the-ratio-of-fatalities-to-total-estimated-infections-for-covid-19-by-the-end-of-2020/#comment-25199
I was unimpressed.
It seems like Iceland is one of the main data sets they rely on, so it would seem important to pay close attention to what continues to happen there. I think 4 deaths is still compatible with their hypothesis, but 8 deaths would already start to violate it and it we're not far away from that.
Edit: I read that 12 patients ...
@(Jgalt)
One thing in particular that stood out to me from that interview (the "Gazette" one):
Singapore has hot temperatures too, like Cambodia and Indonesia where experts were concerned about cases going unnoticed. I initially found it plausible that warmer countries have much fewer cases of new transmissions, but this passage by the interviewee provides evidence against this:
>Unfortunately, I think it’s more likely to be that it’s gathering steam. We’ve released a preprint that we’ve been discussing publicly — and trying to get peer reviewed in the...
@(EvanHarper) Wuhan hospitals were already completely overwhelmed, so to some degree hospital overflow is already factored into the 1% total CFR estimates made by the Imperial College (https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-2019-nCoV-severity-10-02-2020.pdf). While it seems defensible to me to expect hospital crowding to get worse, I don't see why the effect would be as big as you suggest. Countries will also have more time to prepare. (I'm currently predicting even more deaths than you are, b...
@(J08nY) I never know how to update based on news like this. On the one hand, they obviously have an outbreak. On the other hand, they obviously seem to want to continue to pretend to pretend that they don't, and other countries seem to be okay with keeping up that pretense. There's a sense in which the longer this goes on, the weirder it would be if suddenly NK admitted "Oh btw, we really do have an outbreak." So maybe this current state will continue except if NK gets hit so badly by the virus that they will be forced to give up their dignity and ask f...
@(Jgalt) I had noticed that their numbers were off by 2x at least. Also Amesh Adelja on the Sam Harris podcast was wrong here. I don't think new info from Italy or Britain was needed to figure this out. I think they misinterpreted a study or otherwise got false assumptions stuck in their head. I think the culprit might be this cruise ship study: https://www.medrxiv.org/content/10.1101/2020.03.05.20031773v2.full.pdf?fbclid=IwAR3KZVM6R7kqQd_X0MzwnwliBVdpzDa5jvlovqmWFcIh4gZy6K5lnTzwWts
Its abstract says this:
"[...] we estimate IFR and CFR in China to be 0...
@steven0461 I think if the question resolves in 2 days already, then users who predicted really high only get like 2 points in total, so it doesn't matter too much? (I might be totally wrong!)
However, if I'm right then there's still the perverse incentive from better aesthetics: When I scroll down the page of resolved questions, it feels nicer the better the forecasts are, even if this mostly just reflects updating things right before a potential closing, with essentially zero point gains. :)
@(Jgalt)
That Mail Online article is not very good.
>Of these cases, there were only seven deaths, indicating a fatality rate of just 0.004 per cent, which is significantly lower than other countries, including the UK.
Their math is off by a factor 100. And the thing about 50% asymptomatic is totally wrong, see this interview with an Icelandic expert: https://www.icelandreview.com/sci-tech/is-icelands-coronavirus-testing-showing-that-50-of-cases-have-no-symptoms/?fbclid=IwAR2aBMKHGpqdE0bdMb007Rt7_9JrKOtZP-clCMJSeuObZZtAkQJD3GNR4TA
>Several large med...
@(sebk) I'm aware of all that. I still think going below 1% IFR for 2020 isn't reasonable:
- South Korea's testing is quite massive. As I understand it, they had somewhat localized outbreaks especially initially, and those areas got tested very heavily. The entire membership list of the Christian sect got tested (notably a relatively young demograph). And yet their death rate is already above 1% (it's still climbing). It's unlikely that South Korea is missing more than 20% of cases IMO.
- The newest point estimate by the Imperial College is an IFR o...
I updated that <1% is less than 25% likely. It's established that the death rate can be as low as 0.5% if the number of infected people doesn't overstrain hospitals (edit: that's not even true! out of diagnosed cases on the cruise ship it's 1% and in South Korea 1%+!), but unfortunately I expect a large fraction of cases thiss year to happen in places where hospitals will be overstrained.
I updated downward on the impact of hospital overstrain. I think it's unlikely that even the most extreme hospital overstrain would increase the infection fatality rate by much more than 2x.
I now think that at least 400 million infected is more likely than 100-200 million infected.
@(Nostradamnus) the way to fix this is by resolving not on the outcome "humanity survives" or "humanity doesn't survive," but on whether survival will from hindsight be judged to be due to luck or not. I would imagine that our civilization's future experts will have neat technology at their disposal to judge historical counterfactuals. I wouldn't want them to run any sentient ancestor simulations, but maybe they can assess the question in a way that doesn't replay the Holocaust & crudely evaluate predictions that way. Then, if you predicted a 1% chance o...
Oh I see. Going to lose a bunch of points on this because I thought "Democratic People's Republic of Korea" can't possibly refer to an authoritarian regime.
@(ignorance.prior)
>Mortality in children seems to be near zero (unlike flu) which will drive down the IFR significantly.
That's a good point. I don't know to what degree other IFR estimates factor in children. They seem to rarely be among confirmed cases (even though I've read that children tend to be mildly symptomatic, as opposed to being asymptomatic or immune).
>Sarah Newy reports Italy’s death rate might be higher because of how fatalities are recorded. In Italy, all those who die in hospitals with Coronavirus will be included in the death ...
@(DavidMathers)
Unless I made a mistake the last time I looked this up, I think on February 10th there were only 29,600 cases confirmed in Hubei. And unlike Wuhan, the UK didn't try to break a world record with 40,000 people gathering for a potluck (https://www.thestar.com.my/news/regional/2020/02/06/wuhan-neighbourhood-sees-infections-after-40000-families-gather-for-potluck) during the disease outbreak, so that should help keep the situation somewhat under control. Because of better preparation and anticipation of the current situation, I think the UK'...