tag:blogger.com,1999:blog-7612793.post6683417424365614174..comments2024-03-27T23:59:49.801+00:00Comments on Stoat: Into the distance disappear the mounds of human headsWilliam M. Connolleyhttp://www.blogger.com/profile/05836299130680534926noreply@blogger.comBlogger27125tag:blogger.com,1999:blog-7612793.post-73468141799633117672020-12-10T11:16:56.287+00:002020-12-10T11:16:56.287+00:00"William M. Connolley said...
Incidentally, a..."William M. Connolley said...<br />Incidentally, anyone wanna bet? Daily deaths in the US won't go above 2k.<br /><br />23/07/2020, 11:39"<br /><br />Someone should've taken these odds!!<br />Next bet, when will the weekly daily average drop below 2000? January?Nathanhttps://www.blogger.com/profile/12139055978545659341noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-84513685795668790622020-08-16T02:41:45.177+00:002020-08-16T02:41:45.177+00:00I'm sticking to my original Small Hands parano...I'm sticking to my original Small Hands paranoid conspiracy theory. Also CIP is wrong, there are no daily tallies above 2K ... <br />https://github.com/CSSEGISandData/COVID-19/tree/master/csse_covid_19_data<br />You have to download csv files for the JHU time series, then correct Spain, Peru, Iceland, UK, India, ... based on actual country websites (JHU is now rather notorious for skipping seemingly random dailies for just about all countries, it is now rather annoying to correct the ones that I have accidentally tripped over to date).<br /><br />Then there are the under counts based on mortality data ... <br />https://www.nytimes.com/interactive/2020/05/05/us/coronavirus-death-toll-us.html<br />https://www.nytimes.com/interactive/2020/08/12/us/covid-deaths-us.html<br /><br />I'd expect the under counts to be less troublesome for the EU (excepting the UK for the moment) and very much more troublesome for places like Iran, Iraq, Russia, Peru, Mexico, India, South Africa, Columbia ... do they report a death or just go bury them in their neighbors backyard or just go with a funeral pyre?<br /><br />For the UK it is now a real mess (but nothing like in the US where Small hands is literally pulling POB's out of the ground, tangential but illustrative nonetheless) ... <br />https://coronavirus.data.gov.uk/cases<br />https://www.gov.uk/government/publications/phe-data-series-on-deaths-in-people-with-covid-19-technical-summary<br />Where you now get to pick from six columns (in two tables), all several thousand deaths less then what was being reported just say three days ago.Everett F Sargenthttps://www.blogger.com/profile/00201577558036010680noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-17571495300601207062020-08-11T12:33:16.853+00:002020-08-11T12:33:16.853+00:00Thanks. That gives me a funky map, which is quite ...Thanks. That gives me a funky map, which is quite nice, but I totally failed to find a times eries of deaths in the US from it.William M. Connolleyhttps://www.blogger.com/profile/05836299130680534926noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-64430918947715536442020-08-11T10:15:21.650+00:002020-08-11T10:15:21.650+00:00John Hopkins University figures are at
https://gis...John Hopkins University figures are at<br />https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6<br /><br />I think Barrons just used one update being earlier than another to get 2 days of deaths "reported in 24 hours" to make a headline.crandleshttps://www.blogger.com/profile/15181530527401007161noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-42762826426160311152020-08-10T14:22:47.324+00:002020-08-10T14:22:47.324+00:00The 2k "prediction" was deliberately ris...The 2k "prediction" was deliberately risky, and mostly intended to crystallise opinion, in which it succeeded: no-one felt confident in taking the bet. The stats I'm seeing, per A, don't look like 2k was exceeded (I'm using the FT day-by-day for convenience: https://ig.ft.com/coronavirus-chart/?areas=usa&areas=gbr&areasRegional=usny&areasRegional=usca&areasRegional=usfl&areasRegional=ustx&cumulative=0&logScale=1&perMillion=0&values=deaths). the Barron's story unfortunately doesn't link to actual data and I couldn't easily find it at JHU. But the FT 7-day avg is currently at ~1k, slightly below the peak; and new cases appear to be slowly declining too.William M. Connolleyhttps://www.blogger.com/profile/05836299130680534926noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-60722927757367695082020-08-07T13:23:05.947+00:002020-08-07T13:23:05.947+00:00I was tempted, albeit briefly, by William’s bet. H...I was tempted, albeit briefly, by William’s bet. Had I done so, I wouldn’t currently think of accepting any of his money, even if he offered it - my assumption would have been that the 2k figure was a seven day average, daily figures being noisily variable. The worldometer site I use has the daily death rate at about 1,100 (and falling) so I think I was wise not to accept the wager..Anterosnoreply@blogger.comtag:blogger.com,1999:blog-7612793.post-77179261603601752362020-08-07T02:59:54.884+00:002020-08-07T02:59:54.884+00:00You would have lost William, quite unfortunately: ...You would have lost William, quite unfortunately: https://www.barrons.com/news/us-tops-2-000-deaths-in-24-hours-for-first-time-in-three-months-johns-hopkins-01596761106?tesla=yCapitalistImperialistPighttps://www.blogger.com/profile/17919587059726337357noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-508806953829245292020-08-01T08:15:03.950+00:002020-08-01T08:15:03.950+00:00@-WC
"Incidentally, anyone wanna bet? Daily d...@-WC<br />"Incidentally, anyone wanna bet? Daily deaths in the US won't go above 2k."<br /><br />You may be right, although as of 1st August the daily deaths have reached ~1.4k.<br />But the total deaths from a brief peak above 2k are likely to be less than a month or more at 1K.<br />It is how fast the infection rate can be locked down in a region rather than the size of the initial peak that matters. The amount of lockdown being re-imposed in those US states that 'liberated' (Florida, Georgia, Texas) is apparently a matter of politics rather than a measured response to the rate of spread. That does not bode well for the reversal, or even flattening of the curve.izenhttps://www.blogger.com/profile/13782090976414192570noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-17076993831621383162020-07-27T10:59:02.384+00:002020-07-27T10:59:02.384+00:00UK US France
Apr 15 Cases (7da) 4884 30269 3451
Ma... UK US France<br />Apr 15 Cases (7da) 4884 30269 3451<br />May 1 Deaths(7da) 674 1982 335<br />Jul 8 Cases (7da) 545 54585 536<br />Jul 24 Deaths(7da) 63 918 9<br /> <br />Apr deaths per case 0.13800 0.06548 0.09707<br />Jul deaths per case 0.11560 0.01682 0.01679<br /> <br />Improve factor 1.19382 3.89346 5.78126<br /><br />Formatting will probably be terrible sorry. Seems to question why US was used for puzzle. Is the puzzle more of a why hasn't the UK improved at finding milder cases as much as other countries?<br />crandleshttps://www.blogger.com/profile/15181530527401007161noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-38499757739072091072020-07-24T15:15:06.699+00:002020-07-24T15:15:06.699+00:00UK is somewhat ahead of US in both deaths per 1m p...UK is somewhat ahead of US in both deaths per 1m population and further past their peak.<br /><br />Here are UK new cases on 7 day average basis<br />July 8 545<br />July 9 554<br />July10 552<br />July11 581<br />July12 600<br />July13 625<br />July14 598<br />July15 586<br />July16 584<br />July17 610<br />July18 610<br />July19 621<br />July20 628<br />July21 635<br />July22 638<br />July23 656<br /><br />If R is greater than 1 that is causing the increase we should expect growth to be exponential or perhaps worse if further easing of lockdown restrictions occur.<br /><br />A different explanation might be that we could be getting better at finding milder cases.<br /><br />It could of course be a bit of both and/or different reasons.<br /><br />So a possible question to bet on is whether in 2 weeks time this growth will look above or below the mid point between linear and exponential growth.<br /><br />Not sure I particularly want to bet but if forced to choose a side, I think I would prefer to go for the optimistic lower side.crandleshttps://www.blogger.com/profile/15181530527401007161noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-75311805090525537452020-07-23T21:28:13.309+00:002020-07-23T21:28:13.309+00:00Talk about flattening the curve, Small Hands would...Talk about flattening the curve, Small Hands would take a 365-day rolling average of Arctic sea ice extent. Then take a Sharpie to that graph to make it look like we were at the Last Glacial Maximum.<br /><br />You just don't give numbers to a person who has gone bankrupt at least six times, ever ... <br />https://www.washingtonpost.com/politics/2016/live-updates/general-election/real-time-fact-checking-and-analysis-of-the-first-presidential-debate/fact-check-has-trump-declared-bankruptcy-four-or-six-times/Everett F Sargenthttps://www.blogger.com/profile/00201577558036010680noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-58836976853719855232020-07-23T21:22:39.243+00:002020-07-23T21:22:39.243+00:00On sea ice, I'd argue that the chance of a new...On sea ice, I'd argue that the chance of a new minimum is still under 50%. 2012 was a really exceptional year. If we look at the 5 years with the lowest extent on today's date other than 2012 (e.g., 2019, 2017, 2016, 2011, and 2007), look at the difference from 2020 to that year, and then apply that difference to the minimum for that year, only 2016 and 2007 would beat 2012. And given reversion to the mean, I'd argue it is likelier for the difference between 2020 and previous years to decrease than to increase. So that converts to a less than 3/6 chance for 2020 to be a record year. <br /><br />Similarly, if we apply the difference from 2020 to the 2011-19 average (6.58 to 7.47) to the minimum of that average (4.49), we get 3.6. 3.6 is still larger than the 2012 minimum of 3.39. <br /><br />So, yeah, both calculations point to less than 50% chance of a record. <br /><br />-MMMAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-7612793.post-71640441745034773362020-07-23T20:59:20.750+00:002020-07-23T20:59:20.750+00:00"Incidentally, anyone wanna bet? Daily deaths..."Incidentally, anyone wanna bet? Daily deaths in the US won't go above 2k."<br /><br />No effin' way dude ... <br />https://www.cdc.gov/nhsn/covid19/report-overview.html<br /><br />The 7-day centered rolling average for deaths took an ominous bend (lower slope) on July 15th, the same day Small Hands took those midget fingers to the data.<br /><br />Effin' paramilitary police state.<br /><br />What was it you all said four years ago about Small Hands? Worst prognostication ever. That dipshit would sell his own kids and grand-kids into slavery and/or prostitution if he thought that would win him the election come November.Everett F Sargenthttps://www.blogger.com/profile/00201577558036010680noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-18675160945015586872020-07-23T17:10:07.903+00:002020-07-23T17:10:07.903+00:00Sea ice: interesting, isn't it? Just to be orn...Sea ice: interesting, isn't it? Just to be ornery, I'd be prepared to bet on no new min this summer, if anyone wants to take the other side.William M. Connolleyhttps://www.blogger.com/profile/05836299130680534926noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-24858699558648802602020-07-23T17:01:01.730+00:002020-07-23T17:01:01.730+00:00The USA is doing a poor job now of counting cases....The USA is doing a poor job now of counting cases. Not nearly enough testing for the number of cases. True rate is several times the measured rate. Poor is better than abysmal.<br /><br />Infection => illness => hospitalization => ICU => death<br /><br />Each with a lag time and a percentage going a different direction.<br /><br />https://nyti.ms/2WOwXnj<br />Philhttps://www.blogger.com/profile/07567197089095711546noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-64116789719735383082020-07-23T16:21:39.568+00:002020-07-23T16:21:39.568+00:00Ah, perhaps I failed to point out the disparity be...<i>Ah, perhaps I failed to point out the disparity between the initial curve, when the lag was small, and now. Although as Phil points out, that may be due to changing circumstances.</i><br /><br />Part of that is USA did an abysmal job of counting cases at the outset but is doing reasonably well now. Initially they were not capturing pre-symptomatic cases.Layzejhttps://www.blogger.com/profile/11346550512734519728noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-43494105146322409992020-07-23T15:18:34.457+00:002020-07-23T15:18:34.457+00:00Oh, speaking of bets, what you think about the Arc...Oh, speaking of bets, what you think about the Arctic sea ice, and the odds of a new minimum record this summer?<br />Philhttps://www.blogger.com/profile/07567197089095711546noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-20434216216827309872020-07-23T15:17:45.463+00:002020-07-23T15:17:45.463+00:00Sometimes you make a bet hoping you lose. Like buy...Sometimes you make a bet hoping you lose. Like buying life insurance. This would be one of those cases. No reason for me to bet on this.<br /><br />Yet there is still a data quality problem. Deaths in most countries, including the USA, are likely undercounted. I'd suggest anyone that wants to bet frame the payoff in terms of excess deaths.Philhttps://www.blogger.com/profile/07567197089095711546noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-4054408347300389362020-07-23T14:55:59.100+00:002020-07-23T14:55:59.100+00:00I certainly hope you are right on the daily deaths...I certainly hope you are right on the daily deaths but I fear that you are not. I certainly would not be party to such a macabre bet.CapitalistImperialistPighttps://www.blogger.com/profile/17523405806602731435noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-77459609777251640922020-07-23T14:46:22.830+00:002020-07-23T14:46:22.830+00:00Alas, I don't bet. If I did, I'd think exc...Alas, I don't bet. If I did, I'd think exceeding 2k for a single day count would be close to coin flip odds. A couple data points to inform my guess: <br /><br />1) the low point of cases/day was around June 10, at about 21,300. <br />2) the low point of deaths/day was July 5th, at 517<br />3) cases hit a plateau around July 17 at about 68,000/day<br /><br />extrapolating: with a 25 day lag, we'd expect to hit 68/21*517 = 1700 deaths/day around August 11th.<br />alternatively: in the 18 days since July 5th, we've seen an increase from 517 to 862. That's about half the time to August 11th, so then we'd expect to peak at around 1200 deaths/day. <br /><br /><br />some caveats:<br />1) testing has been increasing, such that we might want to adjust older case data upwards to compensate. This would lead to a decrease in my mortality projection<br />2) the mortality low was the July 4th weekend, which might have artificially depressed deaths. So increase the projection<br />3) deaths are presumably still being undercounted, and my guess is that the undercount will be worse in places that are run by republicans - e.g., Florida, Texas, Arizona, etc. This would probably depress future deaths relative to July 5th deaths, as I'm guessing that a higher percentage of future deaths will be in those republican states. So lower the projection. <br />4) all my calculations above were with 7 day averages. Single day counts can be a couple hundred above the average. So if the 7 day average is in the 1600-1700 range, there will probably be a single day which would exceed 2000 at some point. <br /><br />(and now the latest single day mortality count was 1205, after yesterday's 1165. These mortality rates are way higher than should be acceptable...)<br /><br />-MMM<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7612793.post-80743429431310167262020-07-23T11:39:06.063+00:002020-07-23T11:39:06.063+00:00Incidentally, anyone wanna bet? Daily deaths in th...Incidentally, anyone wanna bet? Daily deaths in the US won't go above 2k.William M. Connolleyhttps://www.blogger.com/profile/05836299130680534926noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-75920249686188929792020-07-23T11:38:21.559+00:002020-07-23T11:38:21.559+00:00> deaths often lag diagnosis by several weeks
...> deaths often lag diagnosis by several weeks<br /><br />Ah, perhaps I failed to point out the disparity between the initial curve, when the lag was small, and now. Although as Phil points out, that may be due to changing circumstances.William M. Connolleyhttps://www.blogger.com/profile/05836299130680534926noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-91183660283066516722020-07-22T22:02:01.804+00:002020-07-22T22:02:01.804+00:00Oh, and this would be a good way to monitor over t...Oh, and this would be a good way to monitor over time.<br /><br />Sewage has information.<br /><br />https://www.washingtonpost.com/climate-environment/2020/05/01/coronavirus-sewage-wastewater/?%20environment_2<br />Philhttps://www.blogger.com/profile/07567197089095711546noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-55775353642575594742020-07-22T21:16:32.089+00:002020-07-22T21:16:32.089+00:00Confirmed cases don't have the same meaning ov...<br />Confirmed cases don't have the same meaning over time. Apples to the left, oranges to the right. Apricots stuck in the middle. Remember that the early days in NYC people were mostly tested as they were being moved to the ICU, later on hospital admission, still later on symptoms, now at least in WA State you can generally get tested by just asking for a test. Test positive rates were over 50%. Now under 5%. Drive up testing and generally more testing means the positive tests are earlier in the infection, and more are asymptomatic. Just to make things more fun, the CDC (mostly non-political) has been replaced with HHS, (mostly political) and predictable chaos and missing/untrustworthy data abounds.<br /><br />And yes, deaths are rising. Yesterday 1,119, highest since June 7th.<br />Philhttps://www.blogger.com/profile/07567197089095711546noreply@blogger.comtag:blogger.com,1999:blog-7612793.post-71510781450793436442020-07-22T21:06:07.122+00:002020-07-22T21:06:07.122+00:00Aside from the effects mentioned (age of infected,...Aside from the effects mentioned (age of infected, proficiency in treatment) there is the fact that deaths often lag diagnosis by several weeks. In fact, the last week has seen sharp increases in the number of daily deaths.CapitalistImperialistPighttps://www.blogger.com/profile/17919587059726337357noreply@blogger.com