| Author | Message | ||
     Sifo |
Blake, sorry I haven't kept this conversation rolling. Very rough day yesterday. The house sale failed to close yesterday, and I found some pretty bad errors from our financial institution while trying to get taxes done. Spent hours on the phone not getting answers. Good news, house is supposed to close today. Also, despite a lack of answers for the missing data from our financial institution, last night my wife figured out what they had done in their reporting, and we will be able to straighten it out. It looks to be about a $6,000 difference in my tax liability this year! Don't schedule a house closing or do your taxes on April Fools day! I think you're making conclusion about Italy and Spain that are not supported by facts. How are you concluding that Italy or Spain's pandemic trajectory was mitigated and by what measures? Remember that correlation is not causation. By what measures? Limiting travel and limiting social interactions. Similar to what is going on in many of out states right now. We know that these things will limit the spread of viruses. It's more that theory at this point. Once they identify how the virus is spread (aerosol, droplets, surfaces, etc.) they know measures that will help stop it. When they realized how rapidly this virus was spreading, they took NPI (non-pharmaceutical intervention) measures. Scientific (not fooling ourselves) analysis requires falsifiable hard data and accounting for all factors that may significantly affect the outcome of what is being observed. This pandemic data so far defies scientific analyses beyond mere observation of actual cold hard scientific data. You do what you can with the data you can get. Data collection is almost never perfect. It does affect the error confidence of your projections though. All epidemics that are not extinction events tend to follow a bell curve type trajectory with early slow onset, then rapid acceleration, followed by a relatively linear growth period, then rapid deceleration relatively linear decline, and asymptotic tail off. That can be the end of it, or like the mutating seasonal flue, it never totally goes away, and follows a cyclic trend. Of course, there are reasons for that to happen. Mostly, we have an understanding of how viruses are spread, and we react, which changes the rate of spread. See my above response about NPI measures in Italy and Spain. Seasonal viruses will to that as they have a hard time surviving and spreading seasonally. We aren't sure about that with the WuFlu yet. Or a virus can run it's course through a population until they develop a herd immunity. All will create that classic bell curve. The main point is that there is no solid scientific conclusion to be had from the pandemic data, other than to observe and note what its trend (trajectory) has been so far relative to that in other places. The world has been fooled into thinking that science is capable of helping us know a lot more than it truly can. Two days in a row now, Dr. Fauci has tried to set the record straight about the utility of modelling, that it is only as good as the assumptions used to create it. BRAVO DOCTOR FAUCI!!!. If only the climate crisis alarmists would be so forthright and honest! What may be missing? What say you? Fun stuff. Way better and more fun to discuss than politics! Hahah. I largely agree with all of that. I would say however, projections can be made. Obviously, the one I laid out was very simple. I also only projected out 10 days. More robust projections would require adjustments for factors that will, in what we are discussing here, start to bend the curve down. Let's agree to critique my projection after day 38.  I'm guessing I won't miss the mark by more than 2 days. It would be nice to be wrong on this one. Tom you ought to enjoy this: Just another way of looking at the data. This plot of daily mortality will better help identify the peak of the epidemic. It appears that both Spain and Italy are heading for the downside. Our USA experts (Fauci, Birx and friends) are estimating our peak in two weeks or so, at 26 days in the chart below. That does look encouraging. I find it interesting that in a day you have gone from "This pandemic data so far defies scientific analyses beyond mere observation of actual cold hard scientific data." to "Our USA experts (Fauci, Birx and friends) are estimating our peak in two weeks or so, at 26 days in the chart below.". I'll be fair to you though, things are changing fast with this virus. On that note, it is sad that not long ago, it was being asked if we knew anyone who had been infected, with no responses. Suddenly our little corner of the world is having personal stories come forward. Some of them, not so good. I've got taxes to finish. House closing should happen without me being there today. Lots of other chores. Hope you have a good day, and prayers for those being affected by this thing. | ||
| Ebutch |
 | ||
| Ebutch |
Poor Fredo  | ||
| Aesquire |
It's unseemly to cheer about people exposed to a disease, even if they are evil lying threatening scum who cheer about people exposed to a disease. Not being accusatory, just reminding myself to be a better person. Unlike Fredo. | ||
| Ratbuell |
Clarence Darrow said: I have never killed any one, but I have read some obituary notices with great satisfaction. You aren't alone. | ||
| Tpehak |
I would love to see Seattle police is wearing coronavirus shaped helmets and hats. | ||
| Hughlysses |
More backing for hydroxychloroquine: https://www.washingtontimes.com/news/2020/apr/2/hy droxychloroquine-rated-most-effective-therapy-do/ | ||
| Hughlysses |
Remember how they’ve been saying “masks don’t work”, “they may do more bad than good”, and “you don’t need one”? They’re fixing to tell us we should be wearing masks: https://apnews.com/a7857eb0250093ab796daf6122c7104 7 | ||
| Sifo |
Got the financial reporting fixed up for my taxes. What a PITA!. It did reduce my tax liability by over $5,800 though. Still waiting to hear that the house is closed. Closing by mail is a new thing for me. They’re fixing to tell us we should be wearing masks: I said that quite a while ago. I was walking the pup at a nature area this afternoon. Ran into a neighbor from the hold hood. He's a dentist. He's taking advantage of this time freshening up the dental office. He said he's in the process of applying for the business loans for his business. Also applying for whatever has to be done to delay mortgage payments on his house. I don't know what's available for that kind of stuff, but if it will help you out, I wouldn't delay in applying. Finally feeling like spring around here! | ||
| Tpehak |
They also stated toilet paper does not work, it looks like they are about stating TP is essential product during coronavirus epidemic. | ||
| Ourdee |
https://gisanddata.maps.arcgis.com/apps/opsdashboa rd/index.html#/bda7594740fd40299423467b48e9ecf6 1,011,490 confirmed 52,771 dead 210,186 recovered | ||
| Blake |
Tom: Congrats on kicking the tax man's arse. I get it the other way round. It's not scientific to offer an estimate. That's just human rationalization. Science is observation and testing. If it hasn't been observed yet, then it isn't science. Sorry for being hard headed stance on that; it's a huge issue for me. The relentless popular culture science, science, sciency, scientists say, religion of scientism has become a cult these days that is a mockery of actual science. For example, you're asserting that the decelerating trajectories in Spain and Italy are due to measures taken. But Tom, you don't know that. You're guessing at it. You assume it the case. But it may well be the case that the disease is simply running its course through the population and on its natural decline. The normal bell curve previously noted. The data to conclude one or the other doesn't yet exist. It may never exist. One may reason based upon assumptions, like the one that the epidemic here will peak in two weeks. But that's not science. It's human reasoning based on incomplete data. How might it be off? If weather and temperature are a factor, and if suddenly an effective cure/treatment hits the market, and if infection rates are much higher than guessed, and if Putin is helping us... <smirk> I don't remember your prediction. I'm sticking with 30,000 total mortality. I'm an optimist.  | ||
| Blake |
Tom: This prediction? 100*330 = 33,000 dead in America by April 10th? That's another 28,000 plus in 8 days, which would require continuous exponential increase from the current daily mortality. But that hasn't happened anywhere else during this stage of a nation's epidemic. Look at what the curves for Italy and Spain once they hit the 10 per million mark. Their exponential growth phase ended and went linear then decelerated, though that's tough to impossible to see on a logarithmic scale. Going from 1 to 10 deaths per million took Italy 7 days plus change. Spain 6 days. America 9 days. Going from 10 to 100 deaths per million took Italy 13 days Spain 11 days America 9 days is your guess? Based on the relative milder trajectory of America's mortality data, we should take significantly longer than even Italy to reach that grim milestone. I'd estimate that you're going to be off by a week or more. The logarithmic scale is deceiving you.  | ||
| Blake |
Here's a plot with a decent correlating trend line. If we're going to guess, then...  | ||
| Sifo |
Going back to March 27, I made this reply to Ratbuell... Nobody is talking about the comment made by the President's staff yesterday in the presser, that "models showed that we'd have tens of thousands of bodies by now and we simply don't - the models were wrong". I'll talk about it. I don't know what models predicted that, but right now we are at about 1,300 dead. Double that 3 times and we are there. This virus is progressing with a doubling in the US about every 4-5 days. So the models may have been off by about 2 weeks. That's not that bad when you are dealing with incomplete data in the early stages of a pandemic. Depending on when the models were done, the entire difference may have been made by limiting travel from infected countries. It's a dynamic situation where things are changing daily, including what government is doing to slow it down. It's clear to me that you still don't grasp the exponential nature of what we're dealing with. Very small changes early on have drastic impact on the outcome later on. It's just how exponential math works. So it's been a week now, 1 more week to go. I see in that week we have doubled the deaths twice already, and then a bit more. It's starting to look like we will hit 10,000 deaths after just about 10 to 11 days. That would be April 5 or 6. We'll see how that projection holds up. As I recall, part of science, is taking observations, developing a theory and testing that theory. That requires predicting the results of that test. Is that not what's going on with making a projection during a pandemic? I don't remember your prediction. I'm sticking with 30,000 total mortality. I'm an optimist. I would say you are. I hope you're right. I did predict/project 100 deaths per million people in the US by day 38 on your chart. I'm not sure exactly what day that would be. I think April 8 or 9. Let's hope I'm wrong. | ||
| Sifo |
Looks like we were both putting posts together at the same time. | ||
| Ebutch |
 | ||
| Sifo |
| ||
| Tpehak |
There is a point in amount of infected people when no matter what you do the spread will keep gaining speed. And it looks like this point has been passed. | ||
| Tpehak |
Fun fact, when some time ago I said that they can use closed business facilities as hospitals for infected no one here took it seriously, and now they do exactly this thing turning those facilities to hospitals and quarantine areas. | ||
| 86129squids |
I follow this thread for just a small modicum of logic, let's take a look at those who are running out of it. And then, ignore them. I really really appreciate you guys out there who are statisticians keypad on KeepSafe keep LLC I really don't think Mom asking is that bad I just think my phone is that dumb | ||
| 86129squids |
I don't really think my accent is that bad | ||
| 86129squids |
Hells Bells I guess I better try better at annunciating annunciating annunciating this ••••••• thing doesn't know how to spell e annunciating Are you guys be safe get ready to cook dinner and shelter love one another from a long distance by take a ten-foot pole | ||
| 86129squids |
Let me try again. All of you guys, all my friends be safe, observe what we've been told- although there are some of you who seem to doubt that, I'm worried about my sister... worry about your own. DCUO peace to you all Hells Bells I hate this ••••••• fun phone phone phone | ||
| Aesquire |
https://www.powerlineblog.com/archives/2020/04/the -china-syndrome-3.php Autocorrect is ducking wonderful, eh? | ||
| Aesquire |
https://www.americanthinker.com/blog/2020/04/dont_ count_on_the_model_prediction_for_coronavirus_deat hs.html The assumption, and it's a good one, is that at disease is going to make the same shaped curve on a graph. The numbers and steepness of the slopes is as yet unknown, but the general shape is agreed with. Trying to predict WHERE you are on the curve can be frustrating. Maybe impossible on a day by day basis, since the noise level isn't known precisely. In other words, deciding when we've nosed over on the curve cannot be based on day to day swings. Tuesday May show less new cases than Monday, and Wednesday May show many more. Taking the slope from Tuesday to Wednesday gives you a Bad prediction. It's a common mistake, and a common way to lie with statistics, to cherry pick the start and end dates of a graph. Anyway, we can guess with fair certainly that "it gets worse before it gets better" but which exact day it starts getting better is a full on WAG that is going to change up to and past the predicted date. | ||
| Gregtonn |
An infamous quote comes to mind: "What difference at this point does it make?"_Hillary Clinton G | ||
| Blake |
Tom, What do you say in response to my comments re the history of Italy's and Spain's pandemic trajectories going from 1 to 10 deaths per million (DPM) and from 10 to 100 DPM in comparison to the USA data? Going from 1 to 10 deaths per million took Spain 6 days. Italy 7 days plus change. America 9 days. Going from 10 to 100 deaths per million took Spain 11 days Italy 13 days America 9 days is your guess? What do you say to the semi-log chart for data since exceeding 1 DPM? Note the USA trend line with R2=0.998, excellent correlation. If you use that chart to project an intersection with 100 DPM for the USA, what do you get? How many days from 10 DPM? At what death proportion, at what DPM did Italy's and Spain's trajectories transition from exponential growth (straight line on a logarithmic scale) towards linear growth (downward curving line on logarithmic scale)? The logarithmic scale can be very deceptive. The acceleration of the USA trajectory is decreasing. It was previously doubling in 2-3 days. You note that now it's near 4-5 days. Extend that trend forward. In a week's time it will be...? 7 days? Maybe not, but we'll see. | ||
| Blake |
Tom: As for curve trends being science: Projecting some state of affairs into the future by extending apparent trends on a plot is not science, at all. It's just guessing. Now, if you were to propose that virtually all such viral epidemics follow a certain overall general shape in their trajectory (the bell curve for daily mortality for example), and you presented lots of hard data to support that, and then you applied that "known" shape and best-fit it to existing data then that might be scientific. If you further assumed a duration for the epidemic, say three or four months, and you further constrained your data curve fitting to that, and you reported the statistical correlation and 95% confidence range, then you may be doing scientific work. But it's not completed scientific effort until it's falsifiable, so more data will be required. Just visual trend guessing though, no, that's not science at all. That's how we fool ourselves, and we are each the easiest one to be fooled by ourselves. Dr, Feinman said something like that. | ||
| Sifo |
Blake, first thing I would say would be a question. What are you trying to answer? I'm not sure that using deaths is a good way to track an epidemic in the first place. It is an important metric regarding the impact on society however. Tracking an epidemic with numbers of infected, or critical cases is difficult when comparing country A to country B just because they will almost certainly have different social responses that affect mitigation of the virus. Looking at deaths to compare country A to country B will have all the upfront variables, plus all the variables involved in differences in treatment of the patients along the way. On top of all of that, looking at the US specifically, we have a huge factor working for us. We were late on having the virus getting a foothold. That gives us data collected from many countries around the globe that can help us respond better than the early victims did. We also have some of the best medical care in the world, as long as it doesn't become overwhelmed. So I do agree that projections on something like this are full of problems. That's what leads to the projections I mentioned form Dr. Birx of 100,000 to 240,000 deaths from this. It's still far better than saying we have no ability to see forward on this. So what do I say as far as my prediction vs. yours? I just drew a straight line on a graph. Crude at best. I didn't mean it to be the best that "science" could do, just an example. Yours may be much better. I honestly hope it is. My point though, was that projections do have predictive value. I trust the experts in epidemiology, such as Dr. Birx are using better tools than drawing a straight line projection on a rough graph. I'm curious about something. You seem to argue that this is just following a bell curve based on a "normal" example of a virus running it's course. I'm saying the curve we see right now is in large part to our response to the virus. If I'm correct in my understanding of your argument, what factors do you see causing the slowing of the infection at this time? |
