I’ve done a very rough fermi estimate (I even pull some numbers out of my ass), according to which I don’t need to worry. The probability of me dying due to ncov 2019 is
3⋅1068⋅109⋅1100⋅4⋅15⋅2≈6⋅10−6
Here 3⋅106 is my expected number of infected people, which I got by looking at median according to metaculus, which is 311k, and multiplying it by 10 (a number I pulled out of my ass), since I think the probability density function of the probability distribution of the number of infections is mostly convex.8⋅109 is Earth population.
1100 is the probability of a random person dying if given that they were infected—I got this number by seeing on EA forum a comment that said 3% of identified ncov 2019 cases die, and that probably many infected people haven’t been identified but most deaths have, so I decreased it to 1% (another number out of my ass).
4 - I multiply the probability by 4 for myself since I live in a dense city, and my girlfriend works in a large office with a lot of people. (another number out of my ass)
15 - since mostly old people die, with another number pulling I decrease my probability of dying by 5.
2 - my immune system seems weaker than average, since I often get common cold.
In total I get approximately 6 micromort, which is really low—on average I probably get about 25 micromort per day according to wikipedia. (EDIT: 25 micromort per day is a huge overestimate for my age of 25 - actually I have about 1.34 all causes micromort per day)
The death percentage should perhaps be higher than that 3% actually, since it takes time for someone to be killed by the illness, and most of the people infected have not been infected long enough to die of it.
I’ve done a very rough fermi estimate (I even pull some numbers out of my ass), according to which I don’t need to worry. The probability of me dying due to ncov 2019 is
3⋅1068⋅109⋅1100⋅4⋅15⋅2≈6⋅10−6
Here 3⋅106 is my expected number of infected people, which I got by looking at median according to metaculus, which is 311k, and multiplying it by 10 (a number I pulled out of my ass), since I think the probability density function of the probability distribution of the number of infections is mostly convex.8⋅109 is Earth population.
1100 is the probability of a random person dying if given that they were infected—I got this number by seeing on EA forum a comment that said 3% of identified ncov 2019 cases die, and that probably many infected people haven’t been identified but most deaths have, so I decreased it to 1% (another number out of my ass).
4 - I multiply the probability by 4 for myself since I live in a dense city, and my girlfriend works in a large office with a lot of people. (another number out of my ass)
15 - since mostly old people die, with another number pulling I decrease my probability of dying by 5.
2 - my immune system seems weaker than average, since I often get common cold.
In total I get approximately 6 micromort, which is really low—on average I probably get about 25 micromort per day according to wikipedia. (EDIT: 25 micromort per day is a huge overestimate for my age of 25 - actually I have about 1.34 all causes micromort per day)
The death percentage should perhaps be higher than that 3% actually, since it takes time for someone to be killed by the illness, and most of the people infected have not been infected long enough to die of it.
There’s the opposite effect: the early diagnosed cases tend to concentrate on the more obviously serious ones (more likely to die).
https://twitter.com/robertwiblin/status/1222577306515976192
It is probably wrong to take median (size of pandemic), if we speak of the risk of events with heavy tails.
That’s why I multiplied it by 10.