Why scientists are worried about the coronavirus variant from Brazil

> for millenia the most-effective method of weather forecasting was constancy: "Tomorrow will be just like today"

I disagree that this like predicting the weather. The weather has consistent seasonal patterns that give you ~90% (brown number) of the story (ironically, this virus is also probably strongly seasonal, but I digress).

Here, we're engaged in speculation that some "new strain" might be worse than the existing strain. It's the opposite of using long-term trends to predict the future. It's a bit like claiming that every new typhoon is going to be the worst one ever, because it's different than the last typhoon in some way.

In other words: if we're seeing a resurgence in cases now, the parsimonious answer is that the blood-bank survey estimate was wrong, not that the virus is escaping all immunity. My prior probability of this explanation is dramatically increased when I see that the seroprevalence data was adjusted upward by a factor of 2-3 to account for "seroreversion" (which is essentially guessing).

This is a CNN article, not a source for numbers on cases in the region. It's extremely hard to draw any conclusions from this.

> The number of burials carried out by victims of Covid-19 in Manaus, in January this year, exceeded the data for 2020 for the city. Last year, 1,285 burials from the disease were counted, against 1,333 in the first 21 days of 2021 alone.

The "first wave" happened starting in ~May 2020 (and testing increased across the year) so this kind of un-normalized comparison across time periods isn't useful.

> The total made in January for all statements (all types of death, including coronavirus), was 2,964 - the highest number of burials carried out in a single month since the start of the pandemic. In 2020, the month with the highest burial record in Manaus was April, with 2,433.

This is a great example of bad data reporting: it's almost impossible to interpret this without comparison to historical mortality trends for the region. It sounds like mortality is slightly higher overall, but it's really difficult to say.

> Last year, May was the month that registered the most burials due to the disease, with 348. In comparison with this year's numbers, January 2021 brings 283% more burial records.

Again, same issue as the first quote: making a non-normalized comparison to May 2020 ignores the substantial increases in testing that has happened since that time.