Economist : "Pandemic worse than figures show" - But the worst is over.

Serology testing for antibodies reveals that actual COVID-19 infections outside China peaked between late March and early April in the U.S. and Europe, then in early May in India, Brazil and Africa.

Global pandemic infections have been declining since May, regardless what "confirmed cases" (positive tests) seem to suggest.

Highly misleading "second wave" graphs depicting more U.S. "cases" in July than in April, or more European cases today than in March, only tell us there was very little testing in March and April.

The actual number of worldwide cases, according to The Economist's serological estimates, has fallen from over 5 million a day in May to about a tenth as many by mid September.

If Europe had tested as intensively in March as they did in September, they'd have seen more than ten times as many infections in March as they do now.

Mining the same vein for the USA, the CDC finds the number in the New York City metro area who have been infected and recovered is at least six times larger than the known tally of "confirmed" cases - making it tougher for the virus to jump to new hosts. This also means the true infection death rate (after uncounted mild and asymptomatic cases are included) is much, much lower than once imagined.

Graphs purporting to show the ups and downs of cases over the past six months are almost nonsensical, because we had almost no testing when infections peaked (see the Economist graph above), so the number of "confirmed" cases in March-April missed a lot more than it included.

Testing was trivial in March-April, so the number of cases confirmed by testing amounted to a tiny fraction of reality - particularly in New York.