Rally time!
https://twitter.com/coriduke_kjrh/status/1271529717586432000?s=21
8 Likes
How do I get my family over there? I just read a CNN piece that you can get old farm homes in the country in Japan for pretty cheap (relative to buying a home in the US or a city there). They can have my money just get me out of here.
Tons of math and science follows as I work on building a model to predict hospital overruns. If anyone catches anything you think Iâm doing wrong, please let me know! Wall of text/numbers incoming.
OK, thanks so much! This all seems to be working. Whatâs the significance of 5.6 there?
Also, how did you calculate the R0 in the first place? Backfitting it from the rolling 7 day average of day over day increase in total cases? (Or total actives cases?)
There seems to be a problem with my model here for hospital overrun, because itâs got a few states already tipped over that are not (CA, MI, AZ). These ranges are all for 100% capacity to 150% capacity, because Iâm estimating that most hospitals can scale up that much. My guess here is that this is because Iâm either doing a poor job of estimating the active cases in states that do not have the data available (Iâm subtracting deaths and everything > 1 mo old and then applying that ratio to the estimate of total cases to date) or it could have something to do with the lag time between infection and hospital admission.
My guess on California and Michigan is that it has more to do with the active case ratio, given that they have had significant outbreaks for a longer period of time. With Arizona it could mean that the output Iâm getting (-3.23 to -0.05 days) means thatâs when the tipover became inevitable, and weâre just waiting for cases to be admitted in 7-8 days.
This would make a lot of sense, as Iâm working off of the NYC data and believe they tipped over on 3/27, but the order to not transport non-revived cardiac arrest patients to hospitals went out on 4/8, 12 days later. If we assume that order was given within 24 hours of when they maxed out and there was nothing left that they could do to scale up, and we further assume that their field tents, the USNS Mercy, and the Javits Center allowed them to offload more non-COVID patients to create space, then it stands to reason that other states will hit tipover between 7 and 12 days after inevitability.
As a result, the actual range for Arizona to experience full overrun would become 3.77 to 12.05 days from now, or 6/16 to 6/24.
This gives me the following ranges from states Iâve run, rounding off to whole numbers:
Nevada -25 to 13*
California -8 to 15*
Michigan 1 to 11*
Arizona 4 to 12
South Carolina 15 to 26
Vermont 20 to 28
Kentucky 20 to 33
Oregon 24 to 35
Utah 27 to 37
Florida 27 to 43
North Carolina 28 to 43
Arkansas 37 to 48
Texas 49 to 66
Tennessee 85 to 108
Georgia 142 to 375**
- I believe these are outliers based on doing a poor job estimating total active cases. I think Nevada is useless, because they will have too many positive tests there that return home before being hospitalized. I also think their recovery data is garbage (implies 88% of cases are active, which is way higher than other states).
** I think Georgia is full of shit. The R0 Dan has for them is 1.01, and obviously itâs accurate based on their data but I think their data is bullshit. Somehow theyâre holding between .99 and 1.02 from 5/24 to 6/10, which either means that the Brian Kemp administration is fucking crushing it with calculating the perfect degree of re-opening to maintain an R0 of ~1, or theyâre bullshitting the numbers to make it appear that their R0 is ~1. Iâm going with Door #2.
The important thing to keep in mind is that these predictions will change wildly based on R0 numbers. Arizonaâs is at 2.04, theyâre out of control. If we look at South Carolina, expected to tipover in 15 to 26 days, their R0 is 1.5. If it drops to 1.2, they donât experience tipover until 25-43 days. If it increases to 2, they experience it in 12 to 20 days.
Also they are hugely dependent on the death toll and an estimate of IFR at 1%. I believe the actual IFR is going to end up being lower, but I also believe the death tolls are being underestimated. So if the death tolls are 33% higher, but the IFR is .75%, it still works. I think itâs a reasonable method to get a decent model. HOWEVER, if a state is just totally fudging their death toll by not counting a lot of cases, their model will be way off.
Last but not least, some states are just too big to get any reasonable conclusions from a statewide model. I guess if the entire state is at tipover, the big population centers have to be. But while Arizona has about 70% of its population in Phoenix and Arizona, Texas and California have a bunch of population centers. Case in point, Texas is 49 to 66 days from tipover in my modelâŚ
But Texas is releasing regional hospital capacity data. Based on that and current R0 figures, I would estimate the region with Dallas tipping over in 7 to 18 days, and the region with Houston in 20 to 42 days. This is based of hospitalization rates, which makes me think that I should adjust my entire model to reduce total capacity by 25 to 50 percent to account for non-COVID hospitalizations. This would make Arizona tip over in 2 to 10 days, SC in 9 to 18, TX in 40 to 57, etc⌠This feels more rightâŚ
1 Like
Corn tortillas are just masa & water. Ridiculously easy to make.
1 Like
I have one corn tortilla left - auctioning it off to the highest bidder.
2 Likes
Also I highly doubt theyâre testing corpses that died at home, people who died of pneumonia who hadnât had test results yet, etc⌠And that and the bolded DOES equal fudging the numbers are far as Iâm concerned. They may not have 100 COVID-19 deaths and list 75, but theyâre just flat out doing everything they can to avoid testing/confirming COVID-19 deaths.
There are also likely a couple states flat out fudging it. I mean, shit, weâve seen a lot of data manipulated before it was put out. The unemployment rate was âaccidentallyâ released as lower than it was, a chart from a red county in a red state had the Y-axis inverted to make increased cases appear to be decreasing, and some other chart randomly listed the dates in order to create the appearance of declining cases.
Do we really think these assholes draw the line at flipping the Y-axis and mistaken calculations in unemployment reports? We really donât think there might be a âtypoâ or two in the data?
3 Likes
Obit pages donât lie
Iâm not saying theyâre not counting the deaths as deaths, Iâm saying they wonât show up in COVID-19 stats. If someone dies at home or someone dies of pneumonia who hadnât been tested yet, theyâre not going to test to find out after the fact. As a result, COVID-19 deaths will be under reported. Some states will do a better job than others of counting COVID-19 deaths, some states will do a better job than others of âmissingâ COVID-19 deaths.
Iâm just alluding to that story from Lombardy where they compared obit pages this year in Feb or March during a given week and it was like 8 pages longer than last year. Even if govt officials are downplaying COVID deaths (they are for sure) they canât hide from obit counts.
Also,
https://mobile.twitter.com/michaelcrow/status/1271474685713084416
Lol. The amount of deaths outside of covid are up by quite a bit everywhere in the US.
Yes that means not only are the reported Covid deaths not Flu/Pnu counted deaths it also means after you include those deaths and everything else there is still 25-50% increase in death unaccounted for.
ETA- I just read your post above. Seems like we are oj the same side. My bad.
We are on the same side. Itâs all good. I would have assumed like you did as well.
ETA: I donât think all govt officials are willfully misleading COVID-related deaths. Itâs really hard to count some of these deaths; thatâs why hospital capacity and obit pages are best measurements imo.
2 Likes
I worry about the 8 hours of continual exposure.
So the rallying cry of the mask-hating sovereign citizens is âI Canât Breatheâ? Cool.
3 Likes
My mother in laws mom is in the nursing home. Mom is up visiting us from 5 hours and away and just got the call that her mom only has a day or 2 left. Thereâs been several COVID nursing home outbreaks in the super small town they live in. It seems to be the cancer thatâs doing it, but really weird situation right now. Looks like they are going to let her immediate family in, but the brothers are mostly COVIDHOAXERS. Not a good outcome here.
2 Likes
The essence of the poorly educated
1 Like
5.6 is the mean infection time. So when R0 is 1.1 it means that 100 cases today and 110 cases 5.6 days from now.
So each infection cycle is 5.6 days for a calculation basis. Obviously there is a big spread.
*at least this is how I understand it and the math works where my R numbers line up with the official ones with good correlation.
I did 14 day total cases today over 14 day total cases from 14 days ago to get my daily Rd then calcd R0 from there. I did 14 days to represent the normal range of the period of infectivity. 7 would work as well, just a bit Less noise using 14.
1 Like
This is true. Once you get outside of the big cities, the country is full of abandoned homes as the population ages and younger people migrate to the cities in search of better opportunities. I live in one such area. Many of these homes can be had for about the cost of a new car.
The biggest challenge is getting a visa. If youâre willing to teach English, you can get one pretty easily.
1 Like
yeah. no one is doing anything around here. the local sports leagues have all restarted, drove by the place they play volleyball at which is a bar, 200-300 people playing volleyball and drinking. no masks.
there are maybe 1 or 2 masks when i go to the grocery store now. when i see people they are still trying to give hugs and shake handsâŚ
just complete denail or lack of care.
1 Like
I added some detail from the response awhile ago.
Went to the same place to eat as last week. We got our private table on the side. Apparently the business association has done some advertising as the main restaurant row street was filling up earlier with more people than last Friday.
Will try to get pics next time but itâs just what youâd imagine.