3 ms·
>That is an unlikely hypothesis, given that we didn't see an exponential explosion in deaths The unmitigated doubling time of COVID-19 seems to be around 2.5 d
by ginko 6y ago
>That is an unlikely hypothesis, given that we didn't see an exponential explosion in deaths
The unmitigated doubling time of COVID-19 seems to be around 2.5 days. So if about 1% gets hospitalized and people noticed after 10 or so patients with the same symptoms (that's before any of them died, mind you) you'd need about 1000 people infected. To go from 1 infected person to 1000 takes log2(1000)=9.96 doublings. So if the doubling time is 2.5 that would be 25 days or over 3 weeks before anyone would notice.
Some of these numbers are a bit of a guess but I'm trying to show that "exponential explosion" can be very small and slow at the very beginning of an outbreak.
- rightbyte 6y agoThe variance in actual increase rate is probably way nosier with fewer subjects. I.e. in the beginning you can probably hover at a somewhat constant number of infected for a while before it takes off. On average an infected person seems to spread it to 2-3 persons, but we don't know the variance. It could be something like 8/10 spread it to zero persons, 1 to 4 persons and 1 to 16 persons. With low number of infected that would make the increase very noisy. In fact, it has to be so, otherwise Covid19 would explode everywhere and not be more or less containable. It has to catch momentum or whatever.
- Gibbon1 6y agoThere is also the potential issue that some of the hot spots are like that because they got seeded by multiple super spreading events. As in unlucky happenstance means three infected people in a tour group show up and infect two to three dozen people each over a weekend. You go from 3 to 100 cases in two days. R instead of being 2.5 jumped to 30! After a week just with normal spreading you have close to a 1000 people ill.