| Author | Message | ||
     Sifo |
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. I think I touch on some of that in the above answer. Again, we seem to be composing posts at roughly the same time. I didn't mean to imply that what I did was hard science. The epidemiologists, such as Dr. Birx as an example, I think are doing work far closer to what you are describing. Projections for any virus will only be falsifiable once the virus has run it's course. The work done during the epidemic still has much value though, IMO. | ||
| Tpehak |
Second week of the quarantine is ending. It's time to go back to work. | ||
| Hughlysses |
I figured Tom would have already posted this one: Time to get a mask! | ||
| Blake |
Tom: How does infections or death rate slow over time? Because the population is finite. Worst case is everybody dies. Still a bell curve. As to the what is "science" issue: I'm not critiquing your estimate specifically, just sharing a philosophy of science observation, that no projections from models or trends are science, that models are not science. Models can be dead wrong and often are. I can share some stories. Remind me to tell you sometime. <smirk> Though the REALLY interesting ones are still highly classified. I've been doing modeling of complex systems for a long time. Models are not science, because models can be and often are horribly wrong, only as good as the assumptions and methods used to create them. Model reliability (accuracy) must be proved or disproved by actual testing/observation. At that point, we can only say whether or not and to what extent the model may have been useful. Let's hope our govt's pandemic models prove useful. Models are not science. They are hopefully based upon scientific principles and laws, but they necessarily include lots of boundary conditions and assumptions and even modelling techniques that if poorly chosen can render them horribly wrong. The unknowns, and especially the dreaded "unknown unknowns" can really bite you. The big question according to Fauci is the infection rate and the proportion of population that is asymptomatic. We currently have no idea of those vital statistics. We're guessing at them. And like you say, this particular issue involves human factors. Gah! (Message edited by blake on April 03, 2020) | ||
| Ebutch |
A message From hillary:  | ||
| Blake |
Tom: That video is so bad. Mr. "I've got a degree and am a surgeon so I'm credible" is full crap. He's actually imagining that he knows the death rate. We don't know it. We know how many people die (the numerator), but have no handle on how many people are infected (the denominator). We may know the death rate relative to people tested positive, and the death rate relative to those who are hospitalized. If he wants to compare this pandemic to the flu, then he needs to do so apples to apples based on death rate per hospitalizations. It's so frustrating to see that kind of garbage being spread and scaring people even more than otherwise. Then he dwells on number of cases, and even relative to the liars in communist China no less. Complete garbage. Who has more cases, USA or EU? Gee, why doesn't the news media refer to EU cases?  You've still not commented on the semi-log plot from 1 DPM with the trendline, and what you'd estimate from it. Or what you'd estimate by looking at the different number of days for each nation to progress from 1-10 DPM and 10-100 DPM. Consider what your projections would be for Spain and Italy if all you had was data through to the 10 DPM point, similar to the current situation for USA.. | ||
| Blake |
Science vs hard science? Not sure what the difference would be. Maybe one is based on proxy data versus actual direct observations (hard data). It's gotten so bad in our modern culture that we have been taught that evolution is science. Oh brother. I need Semi to come help explain this philosophical issue. | ||
| Gregtonn |
The "modelers" are starting to run for cover. Many are admitting there were too many variables to come up with an accurate models. Fauci is even admitting that the projections of deaths have been based on inaccurate or unknown data points. I'm not going to quote sources. I'll let you use your own time to find them. They are out there. G | ||
| Sifo |
Tom: How does infections or death rate slow over time? Because the population is finite. But we are nowhere near the point where it runs out of people to spread it to. How do I know that you may ask? For example, MI is now showing data from testing. You can find it here if you care to look. At the moment, they show 23,545 negative tests, 8,005 positive tests, and 31,177 specimines tested. Roughly 3/4 of the tests are negative. BTW, MI is one of the hot spot states at the moment. The tests prove you wrong. Despite what has been claimed by some here, testing has not been wide spread, but targeted at those who are likely sick or exposed to known infection. With the estimated rate of infection, we would need 80% of the population to be infected before acheiving herd immunity. With a much higher infection rate that you imply, you would need even an even higher percentage of the population to have been infected before acheiving herd immunity. To think that we, or any other country have had that happen is just not supported by anything at all. I'm quite startled by your view of modeling not being science. Modeling is math. Was Einstein not doing science? Do engineers not use science when designing a bridge? I share your frustration with certain modeling. Specifically for me, climate modeling. I don't see the problem as being one of modeling though. That one is just a complete perversion of science. That video is so bad. Mr. "I've got a degree and am a surgeon so I'm credible" is full crap. He's actually imagining that he knows the death rate. We don't know it. We know how many people die (the numerator), but have no handle on how many people are infected (the denominator). I believe he was discussing CFR (Case Fatality Rate). We do have the denominator on that. We just don't have the denominator that you demand we must have. You've still not commented on the semi-log plot from 1 DPM with the trendline, and what you'd estimate from it. Or what you'd estimate by looking at the different number of days for each nation to progress from 1-10 DPM and 10-100 DPM. I've discussed it. I wouldn't try to eyeball a curve into a log chart. It's just not a natural visual thing to do. Obvoiusly, there are valid reasons to view the data differently depending on what you are trying to see. Like I said, I hope you are right with your estimate. It certainly differs from the experts who are advising our President. Time will tell. | ||
| Gregtonn |
I believe he was discussing CFR (Case Fatality Rate). We do have the denominator on that. That is an inaccurate statement. Without knowing how many were infected and recovered it is impossible to know the denominator (Cases of the virus). CFR by definition means you must know how many cases there have been. This is why Blake is correct in saying this form of modeling is not science. G | ||
| Ourdee |
Just another dark art. | ||
| Crusty |
So tell me; is calling this Pandemic the Chinese Virus really a racist term? And why? | ||
| Whisperstealth |
My friend, who is a retired nurse, just got a phone call. They offered her $10,000 a week to come and work in an ICU in New York City. | ||
| Tpehak |
Yup, they pay more than $1000 per day. | ||
| Sifo |
That is an inaccurate statement. Without knowing how many were infected and recovered it is impossible to know the denominator (Cases of the virus). CFR by definition means you must know how many cases there have been. This is why Blake is correct in saying this form of modeling is not science. G NO. Stay ignorant. | ||
| Blake |
Tom, The current testing checks for the virus. People who've been infected and recovered don't test positive for the virus, but they were infected. The only way to detect the portion of the population who've been infected is to test for anti-bodies. That test is just now hitting the market. We do not know mortality rate of infected cases, because we don't know how many people were infected. We may be close to knowing how many were infected and showed serious symptoms, but that may be a small or large part of the total infected population. We don't know. Modeling and science: We need to clarify terms. There are models and there are models. Scientific models like Newtonian physics, the laws of thermodynamics, Ohm's law, and Einstein's theories of relativity are simple and concise mathematical descriptions of the fundamental behavior of the physical realm. That is science. The models we're talking about are computer simulations contrived by a modeler who aims to accurately simulate reality; they're not descriptions of the laws of science. Your darn right that modeling as such isn't science, whether it's an engineer of bridges, aircraft, or automobiles, or anything. There's a good reason why no aircraft gets into service absent a full and comprehensive complement of airworthiness testing. Same for vehicles and roadworthiness testing. I don't know if they test scale models of bridges or not. If it's a novel new design approach, they surely do. Boiler plate designs, not so much. We had tens of thousands of unique load cases for the B2 that we used to engineer it. We didn't science it.  Yet we also conducted an absolutelymassive amount of testing, of materials, of components, of subsystems, of full scale aircraft in a huge laboratory hooked up to hydraulic actuators, and of multiple flight test vehicles. The modeling is not science. It's based on scientific principles. It can and will be off the mark often. I wish I could share. Part of my job was to define, witness, and evaluate testing versus what the modeling predicted. My modeling and that of others. The fun we had. I recently had a really sweet model. I was so proud of it. They built the thing. My model was off by double. I'd modeled a major bearing as preloaded. It wasn't preloaded. It was floppy and loose. After the test, I fixed that discrepancy and the model proved useful. And they switched to preloaded bearings like I'd recommended from the start. The testing is the science. Modeling is educated analysis and reasoning. This all goes back to the issue of people's sense of what science really is. Engineering is applying scientific principles and knowledge, but it isn't science. We're constantly revising our code books and modeling techniques. Modeling via mathematics and physical principles or what have you can be wrong. Modeling is based on scientific principles, but modeling and models are not science. You cannot prove a damn thing using a model. You can only hope to accurately/reliably simulate some portion of the material world. Science is testing and observation. Models can help understand how a system responds to certain perturbations and boundary conditions. Did the modeler get them all and model them correctly? Hah/ | ||
| Sifo |
OK, would you go along with models are used in the practice of science? | ||
| Zacks |
Just remember, models are used to predict weather too. | ||
| Aesquire |
The map is not the territory. The model is not the real thing. Models are only as good as the testing you already did to fill in the equations the model uses. Useful, sure, sometimes. IF you are dealing with knowns. It's the unknowns that bite you in the butt. If you want to model a bridge made of reinforced concrete, then decades, centuries even, of testing give you the right numbers and equations to plug into your model. You should have a pretty good idea of what to expect. But a bridge made of Spectra fiber composite with thermoplastic resin? You will need to build real, physical samples of beams and trusses, and more important, connections, and then test them to destruction, before your model is worth the time to even consider. Because few people have built bridges, or anything else with those materials. In aviation, the B-2 is a great example of a flying machine that needed extensive testing to be built with any confidence it would fly and not fall apart. The flying part, that was able to lean on the decades of research that Northrop had put into the aerodynamics of flying wings. But the construction? In large part the darn thing is made from tape. A huge part of the structure was literally on rolls fed into a robotic tape dispenser that laid layer upon layer of tape at carefully calculated angles and thicknesses. The model for that had to be created nearly from scratch, and would have been fiction without actually building big pieces and then carefully, accurately, breaking them. OK, would you go along with models are used in the practice of science? Along with HOW models are used? Or the use of models at all? Not always, or sure, with caution. You have to understand the limitations. Just remember, models are used to predict weather too. Yes, and they are good for a week, 2 weeks is getting iffy, and 200 years is a joke. There just are so many variables in a planetary weather system. I used to pretty much have a weather cube radio with me most of my waking hours not in a classroom or working. ( Glider pilot ) I've spent months, years even, of my life running around looking for the weather conditions that would give me the desired result. ( to soar, sometimes for hours, and many miles, without an engine onboard ) I can assure you, prediction is best vague. That's why Nostradomus is popular, and Fred the Specific Warning Prophet is unknown. ( and, besides, they burned him at the stake for being right when they were wrong ) As to Pooh's Breath? So tell me; is calling this Pandemic the Chinese Virus really a racist term? And why? Nope. And the people telling you that it is are shilling for the Chinese Communist Party that is trying to deflect the blame for lying, continuously, about the disease. Stupid and vile people may increase their existing racist idiocy at the name Chinese Coronavirus, but they also get more racist watching football. Or Breathing. However, I prefer to use a MORE INSULTING name, hence Pooh's Breath. It helps to know the Dictator of China, Xi, was mocked for being a fat guy who supposedly resembled Winnie the Pooh. So he banned the character form his Kingdom, and will have you tortured and murdered for insulting him. So If you want to assign blame for the virus on other than the pre-medieval agriculture system and the lack of sanitation in street markets, rather than pick on the GEOGRAPHICAL ORIGIN, which is NOT racist, I prefer to blame the human monster in charge of the country. And insult him, deliberately, on top of that. Unfortunately I have near zero Social Media presence, so the odds on MY made up name getting popular is near zero as well. Which might be a good thing, since if the TV news parrots started using the phrase, Xi might just nuke them. And as much as I dislike Big City Leftist Hells, most of the good museums and other cultural spots are there. I can get good BBQ in the sticks, but Ballet?  | ||
| Gregtonn |
NO. Tom, Without explanation that is nothing but the sort of childish retort you might hear from an angry two-year-old. Where am I wrong? Like Blake I have years of experience as a design engineer. Projects include modeling and testing for parts of the Space Station and many other satellites. Modeling and testing results are sometimes separated by several orders of magnitude. The thing about most satellite design is that you only get one chance to get it right. G | ||
| Blake |
Tom: >>> OK, would you go along with models are used in the practice of science? Maybe, but only once all the data is in that is required to either confirm or falsify the model's integrity. I'd be hesitant to do so otherwise. Not the kind of models/simulations we're talking about, no. Why? Because science is observation, theory, testing, confirmation and/or falsification. I'd just call it human reasoning, a study of the system of interest that may or may not prove valuable. A model may not even accurately represent a theory as intended. Think of a model as a math equation with millions of degrees of freedom, thousands of variables, and a bunch of unknowns, with a collection of idealized boundary conditions and all kinds of simplifying assumptions. If "scientists" doing models were only paid based on the success of their modeling, there would be a lot fewer models and modelers. hahahah One of the issues I notice a LOT is the complete lack of statistical rigor in what the modelers report. It's a sin against science in my view. Fauci payed it all out, 100,000 to 240,000 is his current best guesstimate. That's a HUGE range. Not as bad as the 41 year old global warming estimate that remains to this day at 1.5-4.5 deg. Celcius, but still pretty massive. What I would say is that model are tools that can be used to try to understand what the science may or may not lead. | ||
| Blake |
Greg, Well, he didn't say "you poopie-head", so I think you're being too harsh. I appreciated the wiki explanation: "In epidemiology, a case fatality rate (CFR) — sometimes called case fatality risk — is the proportion of deaths from a certain disease compared to the total number of people diagnosed with the disease for a certain period of time." What is the case fatality rate for flu in a bad year? | ||
| Sifo |
Tom, Without explanation that is nothing but the sort of childish retort you might hear from an angry two-year-old. Where am I wrong? That there was no explanation. What is the case fatality rate for flu in a bad year? https://journals.lww.com/imd/Fulltext/2019/09000/A _Severe_Seasonal_Influenza_Epidemic_During.4.aspx
Of course, I'm not sure this qualifies as science.  I'm not sure how much of medicine you would accept as science. | ||
| Gregtonn |
Blake, Maybe I was a little harsh but I did come up in an age when spanking was considered a useful form of discipline. For the record: The word denominator comes from the Latin de-"completely" and nominare-"name". Therefore without complete information you do not have a denominator. One other observation: Doctors are said to be "Medical Professionals" but what they do is called "Practicing Medicine". Does anyone see the irony there? G | ||
| Sifo |
Deaths overlaid on top of cases. Cases does look to have a real change in the trend. Cases will follow about 14 or so days after.  Perhaps Greg didn't get spanked enough. Blake, for such a stickler for science, you seem extremely willing to accept a pretty crazy idea that most of the population has already had this virus, recovered, and has antibodies, providing a herd immunity that is reducing the infection rate. I've seen zero evidence for this assertion. The math... Let's just call it very improbable. | ||
| H0gwash |
is calling this Pandemic the Chinese Virus really a racist term? Technically yes, in the same way 'German measles' is technically racist. IMHO the bigger problem is it's an irresponsible term for a leader to use, especially if they work with Asians or Germans. Now that the Chinese are a tiny percentage of all COVID cases, Trump has wisely quit using the term. | ||
| Aesquire |
I doubt most of the population already had Pooh's Breath, but the number is probably quite high. It's now becoming probable, if not yet confirmed that Covid-19 was actually released into the wild/was first contracted by a human ( the difference is political and not relevant to the numbers here ) back in fall 2019, maybe as early as October. If so, some of the flu fatalities and a lot of minor illness for a few months were actually Covid-19. This throws off the numbers a bit, eh? Sometimes you just cannot get the numbers you want. For example it would be useful to get an antibody test on a large sample, but the tests are just now becoming available and then you would want to pick a representative group. Sampling everyone in a U.S. carrier group that's been in the Atlantic training for 4 months would give you a different result than, say, the University of Rochester. And you are never going to get a definitive answer on how many Americans caught Pooh's Breath, ever. Best you can do is make informed guesses. Those flu numbers are informed guesses. I'm not on anyone's side here as far as the numbers go. I'm not even sure how to define the differences you are having. If you take one as pessimistic and one as optimistic? Or is it levels of cynicism? In any event play nice, and remember, this will eventually be over, even if the fatality rate is 100%. ( Cheerful, aren't I? ) | ||
| Hootowl |
“is calling this Pandemic the Chinese Virus really a racist term” No. It is simply descriptive of its origin. Many pathogens/diseases are known by their geographic origin. None of those nomenclatures are racist. The only way this is racist is if “get Trump orange man bad” is all you think about. | ||
| Sifo |
I do find one thing humorous. I've noticed at least 2 trends among those who don't want to believe epidemiologists. One group looks at the cruise ship where all were tested and claims the infection rate is obviously much lower than what we are being told (They ignore the fact that they were under quarantine conditions on the ship). The other group claims that the infection rate is much higher than what we are being told, but most just don't know they've been infected. Both of these groups seem to band together in a united denial of the middle ground. I'm pretty sure that some have gone back and forth on what they claim, just so they don't agree with the best evidence. I figured Tom would have already posted this one: I hadn't seen that one yet. Very hectic lately. It's a short one too. I suggested that we would go there some time ago. But remember, there's no need. You've probably already had it and didn't know, or you're less likely to get it than the seasonal flu. Just ignore what those who have made this their life's work have to say.  | ||
| Gregtonn |
Perhaps Greg didn't get spanked enough. Perhaps. I grew up with a father who rarely spanked any of his eight kids but regularly beat them with whatever was handy. On the farm that might be a hose, broken tree branch, section of hay rope or any number of things . More notably he didn't stop when he thought we had learned a lesson, he stopped when he wasn't mad any more. Once again the term CFR is bastardized by using "reported" cases. Not only is that unscientific it is a patently stupid attempt determine how "deadly" this or any other virus is. How many cases go unreported when someone simply "calls in sick"? G |
