R_eff is [maybe 2 to 2.5 in SA](https://twitter.com/lrossouw/status/1464691396581941261?t=5hqcjcu6w8mxa9phfEFFEw&s=19), whereas it was about 1.0 for delta prior to omicron. R_eff in the US for delta is also close to 1.0, so naively, R_eff for omicron will also be 2 to 2.5. If most of omicron's advantage is due to immune escape, then this might not be that naive. If there are ~10 cases per day of omicron in the US already (who knows, wild guess) then it'll take 67 days at an R_eff of 2 and assuming a generation time of 5 days, to exceed 100k cases per da...

Here is a comparison of timelines for the last three prototypes, in terms of how long it took them to fly after being moved to the launch site:


SN11 is due to move to the launch site Monday (March 8th), 5 days after SN10 flew.

— edited by chrisjbillington

Meta: for an international audience, a date format like 2022-03-07 in the title would be better.

It sounds like Andrew Yang is going to appear on Sam Harris's podcast, which has a large listenership.

At the current rate of immunisation, ~70% herd immunity will be reached in July. Add about eight weeks for cases to drop near to zero after that, add two more weeks for recoveries. Best guess: late September.

Reasons this could be a bad estimate: more vulnerable people will have good vaccination coverage much sooner since they're being vaccinated first. This will take a big chunk out of deaths, but I don't think they'll go actually to zero until cases are near zero.

Vaccine rollout could also speed up.

[Article from Israel on some preliminary analysis of vaccine effectiveness and severe disease risk](https://www.mako.co.il/news-lifestyle/2021_q4/Article-0e660b77fe17d71027.htm). Is in Hebrew, google translation of the main points: > According to these data, the effectiveness of the vaccine (for those who received three doses, ie also the booster), decreases only slightly: 90% protection, compared to 95% protection against the Delta strain. > Another worrying statistic has been revealed regarding the unvaccinated: their risk of becoming seriously il...
Id' be interested in a question on whether omicron is *less* lethal than delta. Not quite the same as just predicting "no" on "will omicron be more lethal", since that would include it being equally as lethal as delta, to within statistical significance (to which I would assign maybe ~0.5 probability mass). A lot of talk of omicron being milder than other variants is going around, based on [very flimsy initial evidence](https://twitter.com/BNODesk/status/1464734073897766919): that people who would be unlikely to get severe COVID symptoms anyway did not...

Booster B1049 has just flown its 8th flight for the Starlink 17 mission and landed successfully.

That puts it equal first with B1051, which has also flown and landed 8 times.

@AngraMainyu I certainly made my predictions assuming local time.

A second static fire has been completed. No official word, but all signs point to it being nominal.

@Symmetry Seems that the community thinks the median number of deaths will be somewhere between 25k and 50k, and that the probability distribution starts to fall off pretty sharply above 50k.

This seems reasonable to me. Given Russia's troops in Ukraine were estimated to number ~190k at the start of the war, and given injury is usually a much more likely way for a soldier to be taken out of action than being killed, >25% of the entire force being killed would really be something.

— edited by chrisjbillington

@MaciekK FWIW the probability of a Poisson distribution with mean 5 yielding zero is 0.7%.

Bumped my estimate up to 32% now that we have more data. I was previously doing some exponential fits to the daily cases, but this was hopelessly confounded by the very strong weekly cycle in case reporting. Now with at least a full two weeks of growth in this wave, we can compare week-to-week directly and avoid this problem.

Cases in South Africa in the last 7 days were a factor of 5.24 higher than the previous 7 days, so that comes to +27% per day. Delta was about -5% per day prior to the current wave, so that's an advantage of +32% per day.

@PeterWildeford Any study worth its salt should control for this, and a result that is more of a measure of the case profile instead of the lethality of the virus itself should not resolve this question.

This estimate of prevalence based on sequencing, on which this question resolved, is clearly wrong. The sequencing must be biased toward omicron in some way. We have not seen total case counts increase enough in the US for it to be plausible that omicron is dominant yet. Nonetheless the CDC sequencing proportion is what resolution was dependent on, so I can't really object. But I don't think people should feel like they were wrong about omicron's growth itself - it's much more likely we failed to take into account the heterogeneity of sequencing that h...

Road closure for today cancelled

Also sounds like the weather is not going to be favourable for a flight for approximately the next week or so.

@RyanBeck Sounds plausible to me. I had been thinking geographic heterogeneity (it's plausible some areas are already omicron-dominant), could be either or both.

It seems like such a figure would be dependent on the level of immunity in a given population, and thus the actual number would vary from population to population. For example, let's say omicron's advantage comes entirely from immune escape, but it has the same R0 as delta. In a population that has no immunity to delta or omicron, omicron would not spread any faster than delta. However in a population near the herd immunity threshold for delta (which might be the case in South Africa), omicron would spread faster than delta. You could even imagine omicr...