XPost: alt.global-warming   
      
     Coronavirus may have infected 10 times more Americans than reported, CDC says   
     MSN Money, 25 Jun 2020 17:43Z   
     Washington (Reuters). Government experts believe more than 20 mn   
     Americans could have contracted the coronavirus, 10 times more than ...   
      
   Amusingly Brietbart and Friends are trying to argue the pandemic is   
   much smaller than it appears to be.  Various reasons have been   
   offered. E.g. covid19 deaths are assigned to anyone with the virus   
   whether it's proven it killed the victim or not.  Or covid19 tests are   
   not accurate and people are said to have it when they do not.   
      
   Some of these things are laughable on their face. Or, when pushed, we   
   can see the errors usually mean the official numbers are way way WAY   
   smaller that they should be in most countries.   
      
   E.g. covid19 tests usually err in the "-ve" direction much more than   
   in the +ve dir. Most tests look for RNA that has features of the SARS2   
   virus. In one common test this means if it says it's seen the virus it   
   is 99% likely to be right. If it says it has NOT seen the virus is is   
   approx 30% likely to be wrong. IOW a single run of this test is likely   
   to understate the number of cases 30%.   
      
   Of course doctors generally use multiple tests per patient to make   
   sure the error in either direction is as small as possible.  Hospital   
   beds are expensive, after all, and we better not suck up space and   
   money treating people that actually don't have it.   
      
   Most number crunchers find the official data understates the situation   
   quite a lot.   
      
   E.g. in the US up to today there have supposedly been 2.7 mn cases and   
   almost 130k deaths.   
      
   But the ratio of these is around 4.7%. Of the cases registered about   
   4.7% result in death of the patient.   
      
   But we know this is WAY too high. From studies of randomly sampled   
   cases that include very weak disease and also very strong disease it's   
   known about 1% of people that get it whether they have symptoms or not   
   actually end up dying from it.   
      
   IOW the US data shows the number of ACTUAL cases is likely to be   
   around 4-5 times more than the official numbers to make the number of   
   deaths work out around 1% of the total.   
      
   But on top of that we also suspect the number of deaths does not   
   capture every death from covid19.   
      
   There are different ways to check this. E.g. we can randomly sample   
   the population and find out how many people have SARS2   
   antibodies. There is some question how long antibodies last after   
   someone is exposed to the disease, but studies in the US and elsewhere   
   find from 1 to 4% of populations with big outbreaks appear to have   
   been exposed to SARS2 no matter what the official stats say.   
      
   Of we can look at the number of registered deaths and compare that with   
   the "normal" number in past years. In the US it's been found there may   
   be twice as many deaths from covid19 than are recorded in official   
   numbers.   
      
   Together all these results point at real US deaths from covid19 being   
   around 2x the official numbers, while covid19 cases are around 10x   
   larger than the official numbers say.   
      
      
      
   Meanwhile, looking at the numbers being reported continues to startle   
   various s/w packages being asked to find patterns in the data.  Some of   
   these s/w are reasonably complex but they sometimes just drop their   
   bundles when new numbers come in because they seem to rigged.   
      
   One way data gathering groups try to ensure data is honest is to   
   require running totals of cases, deaths and recoveries.  This is like   
   double entry bookkeeping. Because running totals are supposed to keep   
   climbing. If they go backwards then someone has done something funny.   
      
   We have seen in the case of France and the UK that several times   
   either running #cases or running #deaths have gone backwards.   
   Sometimes by 10s of 1000s. Usually these big hiccoughs remain   
   unexplained.  In a minority of cases the s/w can see someone got one   
   digit wrong.  Hard to imagine how these data can rely on 1 person   
   typing in a number right at a keyboard and maybe hitting a 2   
   someplace that really mean to hit a 1, but it seems to happen. In   
   such cases the number is generally corrected after a few days by   
   subtracting 100,000 (or whatever) off the running total which can   
   sometimes push the running count for today below yesterday.   
      
   But in other cases a random number seems to have been subtracted from   
   a running total and no-one is owning up as to why.   
      
   These funny edits are seemingly now coming more often. I hate to say   
   it but maybe because things are not looking politically good for some   
   countries and the govt's are trying to avoid being wiped out at the   
   next national poll because of it.   
      
   Here are just 2 examples the s/w squawked loud at in the past 24-48 hrs.   
      
   UK daily cases:   
   172 1295   
   173 1223   
   174 958   
   175 921   
   176 655   
   177 1118   
   178 1381   
   179 891   
   180 913   
   181 830   
   182 690   
   183 832   
   184 -29724	<-- 2 Jul   
      
   WHAAAA? The number of cases in the UK on 2 Jul was NEGATIVE 30k cases?   
   Someone decided that cases that had been recorded were somehow not   
   cases anymore? Generally, as above, the number of "true" cases is   
   many times more than those officially recorded and we can guess that   
   from the ratio of deaths to cases. There should be no reason to   
   subtract some odd number nr 30k off the number of cases.   
      
   Brazil daily recoveries:   
   175 26227   
   176 32506   
   177 19055   
   178 22875   
   179 25316   
   180 18303   
   181 11793   
   182 30507   
   183 137874	<-- 1 Jul   
   184 31500   
      
   Whaaaaa? Brazil is seeing 10-20k recoveries per day then suddenly it   
   sees 10 TIMES MORE on one day?   
      
   Interestingly, Brazil has a history of doing this. Every few months   
   some YUGE number of cases are judged on some day to have spontaneously   
   recovered. Either that or the relevant death certificates have become   
   lost so they MUST have recovered, right?   
      
   Some of the plots for Brazil show an amazing sawtooth pattern.  The   
   numbers build up and up and up at a certain rate.  Then on some day   
   they drop back mightily.  Only to resume building again the day after   
   at exactly the same rate they had before the big drop. And repeat 2-3   
   times to emphasise the books are being cooked.   
      
   Meanwhile, numbers in the US remain remarkably honest.  They might be   
   understated for various "measurement error" reasons but no-one appears   
   to be deliberately editing them to make things look good for   
   them. Well, maybe apart from Florida, and not to any lasting effect.   
      
   It would be "disappointing" if a country that has loudly criticised   
   others for fraudulent record keeping to be found "obviously" guilty of   
   the same.   
      
   --   
   Upcoming events:   
   08 Jul 2020	June US Release NOAA SOTC   
   08 Jul 2020	2020Q2 Billion Dollar Disasters NOAA   
      
   A man may fail many times but he isn't a failure until he begins to blame   
   someone else.   
   -- J. Paul Getty, 29 Aug 2017   
      
   [Post ExtremeEvent:]   
   Jakob Zscheischler @ZscheischlerJak 02 Jul 2020 16:49Z   
   "Moving beyond isolated extremes" Editorial in @NatureClimate about   
   the emerging field of compound events @Compound_Event   
   rdcu.be/b5nUe   
      
   [Severe Weather Season!]   
   Weather Underground @wunderground 02 Jul 2020 19:54Z   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)