Currently Online 
| JRPGfan said: What are the chances, of haveing 2 days, of exactly 1,465 deaths in a row, due to coronavirus? (US) |
Revised later as well, since when I copied it down July 29 was 1485, now it's 1465.
Changes later seem to happen all the time now. I had +17606 cases for Europe July 29, now it shows +17626
North America was +80020 on July 29, now shows 78399
Asia went from 79575 to 79509
South America from 89410 to 95675, and already changed for yesterday as well from 80090 to 86899
Thus changing the world total from yesterday +280337 to +287063, deaths from 6221 to 6426
And for July 29 from +284455 to +289053, deaths from 7034 to 6998 now not breaching the 7,000.
I guess more and more countries are 'time stamping' new cases and deaths putting them in the past when reporting new cases or changing the data afterwards.
I don't know how automated worldometer is but I do know countries do things differently and get handled differently. For example in Canada many provinces don't report in the weekend then you get a spike on Monday when they update (x cases on monday + y cases over the weekend), same for Denmark, silent during the weekend then worldometer puts a spike for Monday with cases for 3 days (I separate them out per day to smooth the curves)
However for France and Spain, who also don't report during the weekend, wordometer shows updates from Saturday and Sunday afterwards, hence the counts changing in the past.
It's getting more work revising past numbers more and more. And does it really matter at this point when everything goes week by week. Sweden is the king of making data look good by pushing stuff into the past, it always looks like the pandemic is over in Sweden, yet the whole curve slowly rises back in time.
There will be a human error margin in all the data as well of course, it's all people counting things and entering them into a system. Ball park and trends, absolute numbers aren't all that reliable.
Essay Pro