Hacker News new | past | comments | ask | show | jobs | submit login
MicroCOVID: Estimate the Covid risk from your ordinary daily activities (microcovid.org)
46 points by sundarurfriend on Oct 1, 2020 | hide | past | favorite | 14 comments

For anyone else without an average foot, 6ft is 1.8m and 10ft is 3m.

This is utterly rubbish.

My scenario: Live with my partner, neither of us goes out unless we REALLY have to. All shopping doorstep delivered.

The dropdowns do not cater for this at all. I might check back if/when some grown-up data is added to it, but for now I'll pass thanks.

That’s exactly there.

You can pre-pick the scenario, “Live in partner with no interactions besides you.”

You can also get there through the series of drop-downs.

Your data is wrong. You estimate 10% of all tests are positive in Luxembourg. That's utter BS. We have tested more poeple than our population and have done multiple testing. The quote is around 1-2% and not 10% as you claim.

This[1] is their source description for Luxembourg. Following their "country profile" link from there, their source seems to be saying around 1% positive test rate for Luxembourg[2].

So it's not clear where the 10.8% figure comes from. I'll try contacting them at their Contact Us email address [3] and see where the difference is coming from.

[1] https://ourworldindata.org/coronavirus-testing#luxembourg

[2] https://ourworldindata.org/grapher/covid-19-daily-tests-vs-d...

[3] https://www.microcovid.org/contact

Surprisingly where I live, a one night stand is less risky that eating indoors in a restaurant. I wouldn’t have guessed.

That said, the chances are still pretty low I think, 1000 in a million isn’t very high is it?

One thousand in a million is the same as one in a thousand.

I'm in the office twice a week. Sadly no good way to set that up and if I try it says something way too high I think.

Does this take into account that people with symptoms should stay home?

I love the effort to quantify things! It's a great approach.

I am sure there are a bunch of ways to nitpick but I appreciate the approach.

I personally think you need a drop-down for age and your risk tolerances are roughly 10x off what they should be.

COVID is almost better thought of as 2 diseases. A dangerous disease for people over the age of 60, and a historically mild disease (less dangerous than many diseases all children got in the 1969for those under 60. If older, taking precautions makes sense.

Very rough ways to think of the death rates are: -Kids have roughly 1 in 10,000 chance of death from COVID. (1/20 as deadly as measles, which almost all children got until the 1963 vaccine) -Parents are roughly 10x more risk than kids. (1 in 1,000) -Grandparents have roughly 1,000x more risk than kids (1 in 10).

Based on: (a) https://reason.com/2020/09/29/the-latest-cdc-estimates-of-co... (which shows alltime numbers) (B) https://www.cebm.net/covid-19/declining-covid-19-case-fatali... (which shows how treatments have improved outcomes well over 60% over time). (c) https://en.wikipedia.org/wiki/Measles

However at a society level, its only because we have become so good and safe that this disease is really relevant.

The age-adjusted death rate, which accounts for the aging population, is about 732 deaths per 100,000 people in the U.S, it has trended down over time. https://budgetmodel.wharton.upenn.edu/issues/2016/1/25/morta...

In 2017 there were over 2.8M deaths. Estimating a 2020 COVID death count at 280k, you are looking at a 10% increase in deaths due to COVID. Since most COVID deaths are in older populations, the age adjusted death rate is smaller, under 5%.

So if you adjust the death rate due to COVID we are looking at approximately 769 per 100,000 dead. A bit safer than in, slightly under 2008's 775.

So the world with COVID is basically like living in the year 2008 (with respect to deadliness).

You can also ask "what happens if we just let COVID run lose?" If you assume it is 2x as deadly, then you would have a 10% age adjusted mortality increase. This would equal an age adjusted deaths would be 805 per 100,000, slightly under 2005's number of 815.

Your risk tolerances (the bottom drop-down) are also way too low. Stating 1% per year recommended risk for many people is basically saying you should take 1 in 100,000 chance of death in a year. Car crash fatalities are 10x that annually in the US (38k deaths / 350M). Most people willingly assume car crash risk as a given.

Anyway, this is too long of a response but you need age as an input as it changes the outcomes by many orders of magnitude and you your risk tolerances are an order of magnitude off IMO (I like providing options though).

Are you suggesting that letting covid run loose is OK because society's overall death rate will be the same as it was 15 years ago?

That argument has severe moral and ethical problems.

I also think you're underestimating the impact of letting the virus run loose. Countries with very poor pandemic management have seen large increases in deaths. For example, Peru's death rate this year is 156% of what it was last year. And that is only half of the year, and summer. See https://www.ft.com/content/a2901ce8-5eb7-4633-b89c-cbdf5b386...

Absolute trash. According to this I should be dead by now because my GF works in a hospital.

“Dead by now?”

When I (NYC) put that scenario, live with a person who is an essential or front-line worker, it says:

“This is a roughly 6,000-in-a-million (0.6%) per week chance of getting COVID from this activity with these people.“

So maybe 15%-20% chance so far of experiencing your own case (and of course a very small case fatality rate if you did), hardly seams “should be dead by now”.

If I try to simulate that, it says 50% risk of getting COVID.

But most people won't die from it.

As a first approximation: 206,852 ÷ 7,262,629 cases = 2.8%?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact