bluefrogguy

Coronavirus (COVID-19) Discussion

Recommended Posts

Don't think we are slowing down when our critical patients number keeps going up.  Over 12k now.  Our recoveries compared to deaths puts us worse than Italy right now.

Share this post


Link to post
Share on other sites
1 hour ago, daynlokki said:

Don't think we are slowing down when our critical patients number keeps going up.  Over 12k now.  Our recoveries compared to deaths puts us worse than Italy right now.

You guys as a whole as trending in the right direction. Some states are behind, but the big ones like New York and California are certainly on the right paths.

Here in Canada, we are on the right path as well, what is mind boggling to me is we are following the same sort of guidelines and restrictions as you guys, and yet, only averaging 1300-1500 cases in the whole country a day.. and our boy Mr.Ford just announced another 28 days of restrictions to follow the April 23rd deadline he originally set... insanity about to kick in.

Share this post


Link to post
Share on other sites
1 hour ago, jonninho said:

You guys as a whole as trending in the right direction. Some states are behind, but the big ones like New York and California are certainly on the right paths.

Here in Canada, we are on the right path as well, what is mind boggling to me is we are following the same sort of guidelines and restrictions as you guys, and yet, only averaging 1300-1500 cases in the whole country a day.. and our boy Mr.Ford just announced another 28 days of restrictions to follow the April 23rd deadline he originally set... insanity about to kick in.

That’s because nobody is following the social distancing and stay at home orders. My county grades as an F in that. Still nobody is taking it seriously. Someone I went to high school with decided to throw a bonfire and invite about 30 people this last weekend. Zero f***s given. 

  • Haha 1

Share this post


Link to post
Share on other sites

A bit of good news in our small state of Nevada.  It is believed we have passed our peak and, assuming social distancing continues through May, pace to have 250 total deaths, which seems quite reasonable.

https://news3lv.com/news/local/report-nevada-passes-peak-in-covid-19-cases-254-deaths-projected-by-august

 

I mentioned before that I was optimistic about our prospects as our Gov was one of the first to act decisively and we already have a very distant culture.  Outside of the tourist spots, things are fairly spaced out and we drive everywhere. Most people have moved here and there are a lot of single people or nuclear families. It can be a lonely place, but perhaps that is helpful in this particular circumstance. 

The down side is we rely on international tourism to drive a service economy.  So, 1) Almost nobody is working right now.  There are very few "work at home" type jobs.  2) Even though we did a good job, it's not clear when we can restart our economy because our economy involves flying people in and sending them home with empty pockets.  And some of those people come from places like Georgia and Florida, or attend megachurches with irresponsible pastors.  So we might be at the mercy of the most foolish among us.

 

Share this post


Link to post
Share on other sites
Posted (edited)
On 4/12/2020 at 1:52 PM, tonycpsu said:

Combined with your continued gish gallop pattern of making claims and then moving the goalposts when they're proven false

 

Speaking of moving the goalposts, I think it's worth revisiting how the models have been wrong several times now after a few iterations. To be fair, I haven't checked in on this thread in a few days and haven't had the chance to look at previous posts.

 

Not using this as a "gotcha" but the models --and I refer to all the different ones that have received pub, such as the IHME and Washington one-- after the initial one that predicted 2 million deaths have been incorrect. And yes, social distancing has been factored into these previous models.

 

I've done models myself and acknowledge that they are subject to a wide range of error based on your assumptions as inputs. But for all the reverence that the smart scientists/experts get, they deserve some push back when their models are off by so much. I'm not saying to ignore them, but you can understand why some people might be skeptical about the level of doom and gloom being spouted when the models repeatedly get revised down 

Edited by UberRebel
  • Like 2

Share this post


Link to post
Share on other sites
14 minutes ago, UberRebel said:

 

Speaking of moving the goalposts, I think it's worth revisiting how the models have been wrong several times now after a few iterations. To be fair, I haven't checked in on this thread in a few days and haven't had the chance to look at previous posts.

 

Not using this as a "gotcha" but the models --and I refer to all the different ones that have received pub, such as the IHME and Washington one-- after the initial one that predicted 2 million deaths have been incorrect. And yes, social distancing has been factored into these previous models.

 

I've done models myself and acknowledge that they are subject to a wide range of error based on your assumptions as inputs. But for all the reverence that the smart scientists/experts get, they deserve some push back when their models are off by so much. I'm not saying to ignore them, but you can understand why some people might be skeptical about the level of doom and gloom being spouted when the models repeatedly get revised down 

The original model of 2m deaths was without any social distancing. As each individual state changes its social distancing and stay at home orders the models change. As they also do as more hospital beds and ventilators come available. The fact you say the original model included social distancing shows me you haven’t even attempted to look up the subject before posting. 

  • Like 1

Share this post


Link to post
Share on other sites
On 4/12/2020 at 1:52 PM, tonycpsu said:

Automobiles get people to work, to see their loved ones, etc.  They facilitate trade and grow the economy.  This virus just kills people.  If you aren't prepared to account for the positive value of driving in you analysis, then you're not bringing anything compelling to the discussion.

 

The comparison isn't cars vs coronavirus. You're twisting his analogy in order to dismiss it. It's pretty valid. He's saying there's simultaneously a benefit and risk to driving cars just as there's also an obvious benefit and risk of the relaxing social distancing at this time.

 

The implied benefit he was referring to was all the good things you get from being able to interact with others normally (making a living, enjoying life, etc.)

 

He said "The fact is thousands of people die every year in car accidents. Yet, why aren’t people pushing to ban cars? It’s accepted by our society because we value cars that much." To summarize, using a car allows everyone to save time by traveling farther distances, while also carrying risk of dying in a car accident.

 

Comparatively, going out in public to earn a living, enjoying life, etc. also increases likelihood of people getting Coronavirus. But there's a level of risk we should be willing to accept in order to reap some of those benefits.

 

And I'm not saying every person is being prevented from earning money - but many people are. And it's not an all or nothing. It isn't a choice between no social distancing and complete social distancing (which we don't really have). There are gradations we should be willing to have conversations about while balancing risk vs. reward.

  • Like 3
  • Thanks 1

Share this post


Link to post
Share on other sites
Posted (edited)
16 minutes ago, daynlokki said:

The original model of 2m deaths was without any social distancing. As each individual state changes its social distancing and stay at home orders the models change. As they also do as more hospital beds and ventilators come available. The fact you say the original model included social distancing shows me you haven’t even attempted to look up the subject before posting. 

 

Try again. Read this sentence and see where I said "after". I did not edit that part of my original comment (honest to goodness)

 

Not using this as a "gotcha" but the models --and I refer to all the different ones that have received pub, such as the IHME and Washington one-- after the initial one that predicted 2 million deaths have been incorrect. And yes, social distancing has been factored into these previous models.

 

NOTE: I'm not saying EVERY model after the original 2million deaths prediction has social distancing. There are a few out there and I haven't seen all of them. But I am saying that many of them after the original one started "factoring in" social distancing and have still been off by quite a bit.

Edited by UberRebel

Share this post


Link to post
Share on other sites
5 hours ago, UberRebel said:

 

The comparison isn't cars vs coronavirus. You're twisting his analogy in order to dismiss it. It's pretty valid. He's saying there's simultaneously a benefit and risk to driving cars just as there's also an obvious benefit and risk of the relaxing social distancing at this time.

 

The implied benefit he was referring to was all the good things you get from being able to interact with others normally (making a living, enjoying life, etc.)

 

He said "The fact is thousands of people die every year in car accidents. Yet, why aren’t people pushing to ban cars? It’s accepted by our society because we value cars that much." To summarize, using a car allows everyone to save time by traveling farther distances, while also carrying risk of dying in a car accident.

 

Comparatively, going out in public to earn a living, enjoying life, etc. also increases likelihood of people getting Coronavirus. But there's a level of risk we should be willing to accept in order to reap some of those benefits.

 

And I'm not saying every person is being prevented from earning money - but many people are. And it's not an all or nothing. It isn't a choice between no social distancing and complete social distancing (which we don't really have). There are gradations we should be willing to have conversations about while balancing risk vs. reward.

Ok, so in order to minimize your risk to OTHERS while you drive, you have to legally have liability insurance.  What are you gonna have to limit your liability here if you end up infecting someone and killing them from it?  What liability will the states have if they open up too soon and end up having thousands more die needlessly? 

 

5 hours ago, UberRebel said:

 

Try again. Read this sentence and see where I said "after". I did not edit that part of my original comment (honest to goodness)

 

Not using this as a "gotcha" but the models --and I refer to all the different ones that have received pub, such as the IHME and Washington one-- after the initial one that predicted 2 million deaths have been incorrect. And yes, social distancing has been factored into these previous models.

 

NOTE: I'm not saying EVERY model after the original 2million deaths prediction has social distancing. There are a few out there and I haven't seen all of them. But I am saying that many of them after the original one started "factoring in" social distancing and have still been off by quite a bit.

It's a novel virus.  They have known about it all of 5 months.  Any scientific model based on 5 months of anything is going to be wildly off, especially when those same models are including social distancing and the populace isn't following that.  Every single model uses two things: data and assumptions.  With less data, there are more assumptions.

Share this post


Link to post
Share on other sites
Posted (edited)
2 hours ago, daynlokki said:

It's a novel virus.  They have known about it all of 5 months.  Any scientific model based on 5 months of anything is going to be wildly off, especially when those same models are including social distancing and the populace isn't following that.  Every single model uses two things: data and assumptions.  With less data, there are more assumptions.

 

I acknowledged earlier that the models wouldn't be precise (as models are won't to do) and was resigned to that inevitability. Seems you're acknowledging that the models have been "wildly off" and that's part of my point...

 

As the basis for putting close to 20 million Americans out of work and tanking the global economy, the models should have been a lot more accurate. It's not acceptable to just state they were destined to be "off" and ho hum no big deal that's that without need for further discussion.

 

Social distancing factored in, many of them have been off by 100s and in some specific U.S. locales 1000s of percentage points. I don't know if that's pitiful when compared to standard pandemic modeling, but at the very least, you can say that we shouldn't have weighted these numbers so heavily when taking such drastic action.

 

So, now that they have been off by a ton, anyone that tries to use these garbage models as a cudgel for guilting or scaring anyone into responding in X fashion, deserves to have someone raise an eyebrow and say "wait a minute, are we overreacting here?" or "hang on, are you sure about these assumptions as inputs into your model?" This is a perfectly rational take and I'm not seeing enough of that viewpoint being given a chance right now. In fact, what I'm seeing is more often than not, people questioning the data are being called money-grubbers, callous, conspiracy theorists, who don't care about human lives, etc. Not saying that kind of name calling is necessarily going on in this thread

 

To your point about the models being based off of 5 months of data, I say that's even more of a reason why the government response should have been more measured and ramped up accordingly as more information came out. 

 

To be clear, I'm not against social distancing and I'm not saying Coronavirus isn't deadly. But when the stakes are life, liberty, and economics (all of these matter; oh and not trying to tout the Declaration of Independence) it's healthy to question authority and the reasoning authorities use for us to change our ways of life. Humans can be wrong - that includes Dr. Fauci (a good guy). Here are some a couple of other examples from just the past several weeks during this whole ordeal

 

-WHO says that COVID19 cannot be transmitted from human to human

-CDC says masks aren't needed by the average citizen. Only healthcare workers need them. Now, a few weeks later, everyone should make their own masks at home and we'll fine you without one in public (some places)

Edited by UberRebel
  • Like 4
  • Thanks 1

Share this post


Link to post
Share on other sites
1 hour ago, UberRebel said:

 

I acknowledged earlier that the models wouldn't be precise (as models are won't to do) and was resigned to that inevitability. Seems you're acknowledging that the models have been "wildly off" and that's part of my point...

 

As the basis for putting close to 20 million Americans out of work and tanking the global economy, the models should have been a lot more accurate. It's not acceptable to just state they were destined to be "off" and ho hum no big deal that's that without need for further discussion.

 

Social distancing factored in, many of them have been off by 100s and in some specific U.S. locales 1000s of percentage points. I don't know if that's pitiful when compared to standard pandemic modeling, but at the very least, you can say that we shouldn't have weighted these numbers so heavily when taking such drastic action.

 

So, now that they have been off by a ton, anyone that tries to use these garbage models as a cudgel for guilting or scaring anyone into responding in X fashion, deserves to have someone raise an eyebrow and say "wait a minute, are we overreacting here?" or "hang on, are you sure about these assumptions as inputs into your model?" This is a perfectly rational take and I'm not seeing enough of that viewpoint being given a chance right now. In fact, what I'm seeing is more often than not, people questioning the data are being called money-grubbers, callous, conspiracy theorists, who don't care about human lives, etc. Not saying that kind of name calling is necessarily going on in this thread

 

To your point about the models being based off of 5 months of data, I say that's even more of a reason why the government response should have been more measured and ramped up accordingly as more information came out. 

 

To be clear, I'm not against social distancing and I'm not saying Coronavirus isn't deadly. But when the stakes are life, liberty, and economics (all of these matter; oh and not trying to tout the Declaration of Independence) it's healthy to question authority and the reasoning authorities use for us to change our ways of life. Humans can be wrong - that includes Dr. Fauci (a good guy). Here are some a couple of other examples from just the past several weeks during this whole ordeal

 

-WHO says that COVID19 cannot be transmitted from human to human

-CDC says masks aren't needed by the average citizen. Only healthcare workers need them. Now, a few weeks later, everyone should make their own masks at home and we'll fine you without one in public (some places)

You misquoted the WHO. Their exact words were AT THIS TIME there is no evidence of human to human transmission. Which, at that time the only country with any positive tests was China, who we all know weren’t sharing information. 
 

They learned new information. Originally transmission was thought to be by the normal coughing or sneezing. They now know that just breathing on someone while sick with covid can transmit the disease. 
 

This isn’t something you can just slowly ramp up. It’s either full stop or you’re gonna f--- over the ERs in your state. Have you not seen the mass graves in NY? The carrying of the dead by forklift into refrigerated trucks to use as temporary morgues? This disease is more communicable than the flu. By far. Look at the per capita rates of infection to death between last years flu (14 dead per 100,000 infected) and coronavirus while in a complete lockdown nationwide (4,027 per every 100,000 infected). Yes, a lot of people haven’t been tested. Same is true for the flu every season. If we wouldn’t have shutdown our serious and critical condition cases would have overwhelmed our medical system completely and thoroughly. There is already talk that the deaths are low because of the reporting standards. You must have a positive test and you must die in the hospital to count as a death. You die at home or without a positive test and you don’t count towards the statistics. 

  • Like 2

Share this post


Link to post
Share on other sites
1 hour ago, UberRebel said:

 

I acknowledged earlier that the models wouldn't be precise (as models are won't to do) and was resigned to that inevitability. Seems you're acknowledging that the models have been "wildly off" and that's part of my point...

 

As the basis for putting close to 20 million Americans out of work and tanking the global economy, the models should have been a lot more accurate. It's not acceptable to just state they were destined to be "off" and ho hum no big deal that's that without need for further discussion.

 

Social distancing factored in, many of them have been off by 100s and in some specific U.S. locales 1000s of percentage points. I don't know if that's pitiful when compared to standard pandemic modeling, but at the very least, you can say that we shouldn't have weighted these numbers so heavily when taking such drastic action.

 

So, now that they have been off by a ton, anyone that tries to use these garbage models as a cudgel for guilting or scaring anyone into responding in X fashion, deserves to have someone raise an eyebrow and say "wait a minute, are we overreacting here?" or "hang on, are you sure about these assumptions as inputs into your model?" This is a perfectly rational take and I'm not seeing enough of that viewpoint being given a chance right now. In fact, what I'm seeing is more often than not, people questioning the data are being called money-grubbers, callous, conspiracy theorists, who don't care about human lives, etc. Not saying that kind of name calling is necessarily going on in this thread

 

To your point about the models being based off of 5 months of data, I say that's even more of a reason why the government response should have been more measured and ramped up accordingly as more information came out. 

 

To be clear, I'm not against social distancing and I'm not saying Coronavirus isn't deadly. But when the stakes are life, liberty, and economics (all of these matter; oh and not trying to tout the Declaration of Independence) it's healthy to question authority and the reasoning authorities use for us to change our ways of life. Humans can be wrong - that includes Dr. Fauci (a good guy). Here are some a couple of other examples from just the past several weeks during this whole ordeal

 

-WHO says that COVID19 cannot be transmitted from human to human

-CDC says masks aren't needed by the average citizen. Only healthcare workers need them. Now, a few weeks later, everyone should make their own masks at home and we'll fine you without one in public (some places)

There is a good reason to overreact too ANY bat transmitted virus. They have a extremely effective immune system that forces viruses to adapt to dangerous level

This is still a learning process.In both how we should handle it and how virus works. Usually better to safe then sorry with threat we don't have 100% info on. Yeah, the WHO parroting what China was saying did more harm then good. Let the worlds guard down when it needed to be up.

  • Like 1

Share this post


Link to post
Share on other sites
23 minutes ago, Slatykamora said:

There is a good reason to overreact too ANY bat transmitted virus. They have a extremely effective immune system that forces viruses to adapt to dangerous level

This is still a learning process.In both how we should handle it and how virus works. Usually better to safe then sorry with threat we don't have 100% info on. Yeah, the WHO parroting what China was saying did more harm then good. Let the worlds guard down when it needed to be up.

They didn’t parrot anything. The only cases were in China. China wasn’t sharing all the data. The WHO had limited data to work with and none of what they had showed a definitive human to human transfer of the virus. Scientifically the only thing they can say when they don’t have evidence of something is that they don’t have evidence of it. We all know OJ was guilty. There wasn’t enough evidence for it. 

Share this post


Link to post
Share on other sites
2 hours ago, daynlokki said:

They didn’t parrot anything. The only cases were in China. China wasn’t sharing all the data. The WHO had limited data to work with and none of what they had showed a definitive human to human transfer of the virus. Scientifically the only thing they can say when they don’t have evidence of something is that they don’t have evidence of it. We all know OJ was guilty. There wasn’t enough evidence for it. 

O...K.  I'm not even sure what we dis-agree about here? 

Share this post


Link to post
Share on other sites

Any thoughts on why China's numbers are so low...nobody comes close per capita to them? 

I blame the Chinese government CCP ( not their people ) for misguiding the WHO ( which has proven to be useless for pandemics at this point )

 

 

 

Share this post


Link to post
Share on other sites

Interesting that the top two countries with the highest percentage of population over 65 years old ( Japan and Italy ) are at almost opposite ends of the COVID19 spectrum.

Am I to assume that Japan has better healthcare and less Chinese visitors ( i.e. less COVID19 exposure ) than Italy? Your thoughts?

Share this post


Link to post
Share on other sites
27 minutes ago, Dr. Whom said:

Interesting that the top two countries with the highest percentage of population over 65 years old ( Japan and Italy ) are at almost opposite ends of the COVID19 spectrum.

Am I to assume that Japan has better healthcare and less Chinese visitors ( i.e. less COVID19 exposure ) than Italy? Your thoughts?

 

The latter  (less Chinese visitors) is definitely true.... Italy has numerous large-scale business relationships with China .....

Share this post


Link to post
Share on other sites
Posted (edited)
29 minutes ago, Dr. Whom said:

Interesting that the top two countries with the highest percentage of population over 65 years old ( Japan and Italy ) are at almost opposite ends of the COVID19 spectrum.

Am I to assume that Japan has better healthcare and less Chinese visitors ( i.e. less COVID19 exposure ) than Italy? Your thoughts?

+ high population dentistry.

Really does have a pattern with eastern culture handling this better then the west. Though that is only a rough correlation. There is also a correlation between the spread and China's belt and road initiative. As Italy and Iran were the first 2 outside countries hit the hardest at first. Then spread hard to the countries most interconnected to Italy and so on.

When Italy started getting hammed is when this became a wake up call for the rest of world. 

South Korea's provocativeness stems from a partial motivation due to a MERS outbreak before in their country..

Edited by Slatykamora

Share this post


Link to post
Share on other sites
25 minutes ago, Slatykamora said:

+ high population dentistry.

Really does have a pattern with eastern culture handling this better then the west. Though that is only a rough correlation. There is also a correlation between the spread and China's belt and road initiative. As Italy and Iran were the first 2 outside countries hit the hardest at first. Then spread hard to the countries most interconnected to Italy and so on.

When Italy started getting hammed is when this became a wake up call for the rest of world. 

South Korea's provocativeness stems from a partial motivation due to a MERS outbreak before in their country..

What does dentistry have to do with this?

China's numbers are likely under reported.  East Asian countries like to keep everything hush hush.  It's not like in the US where the media reports everything and makes sure to get to the bottom of it. They have also dealt with these pandemics before, and regularly wear masks on an everyday basis, while Western countries have not.

  • Like 1
  • Haha 3

Share this post


Link to post
Share on other sites
7 minutes ago, MrBrown said:

What does dentistry have to do with this?

China's numbers are likely under reported.  East Asian countries like to keep everything hush hush.  It's not like in the US where the media reports everything and makes sure to get to the bottom of it. They have also dealt with these pandemics before, and regularly wear masks on an everyday basis, while Western countries have not.

He is an Anti-Dentite

  • Like 3
  • Haha 3

Share this post


Link to post
Share on other sites
8 minutes ago, MrBrown said:

What does dentistry have to do with this?


Well if you don't have lots of dentists, then soup (bat flavored cup o noodles perhaps) is much easier to eat for those with oral issues. 

  • Like 1
  • Haha 2

Share this post


Link to post
Share on other sites
16 hours ago, UberRebel said:

Speaking of moving the goalposts, I think it's worth revisiting how the models have been wrong several times now after a few iterations. [...]

And yes, social distancing has been factored into these previous models. [...]

I've done models myself and acknowledge that they are subject to a wide range of error based on your assumptions as inputs. But for all the reverence that the smart scientists/experts get, they deserve some push back when their models are off by so much. [...]

I'm not saying to ignore them, but you can understand why some people might be skeptical about the level of doom and gloom being spouted when the models repeatedly get revised down 

 

I guess my response here depends on (a) which other models you're talking about, (b) what "off by so much" means, and (c) what number you're assuming represents the "true" number of deaths for a disease that's still killing thousands daily and hasn't peaked yet nationwide.  Since you've been unclear about all of these variables in your criticism, I will make some assumptions and you can tell me which assumptions of mine are incorrect, just as we would do if we were modeling the pandemic ourselves.

For (a), the only model you're explicitly citing as performing poorly is the IHME one, so let's look at that for now.  If you have any other influential models you think made egregious errors, feel free to cite those as well.  I'll limit this discussion to the number of deaths predicted by that model, since that's what you're focusing on in your post.

Like most models, the UW IHME model predicted a range of outcomes for the number of COVID-19 deaths in the US.  The range of their initial forecast was 38,000 on the low end and 162,000 on the high end, with 81,000 representing their best guess at that time.  That best guess number has been revised several times as new information has come in, which, as you say, is how things are supposed to work.  Here is how the numbers have progressed since the IHME model was released, with the min, max, and median projection for each:

2020-03-26: 38,000 - 162,000 (81,000)
2020-03-30: 40,000 - 142,000 (82,000)
2020-03-31: 37,000 - 153,000 (84,000)
2020-04-01: 41,000 - 177,000 (94,000)
2020-04-02: 40,000 - 178,000 (94,000)
2020-04-05: 49,000 - 136,000 (82,000)
2020-04-07: 31,000 - 126,000 (60,000)
2020-04-10: 26,000 - 155,000 (61,000)
2020-04-13: 30,000 - 176,000 (69,000)

Looking at the data this way, your statement that models are "repeatedly being revised down" is, to put it charitably, an overstatement.  The estimates went up, then down, then up again.  You can click through the links for each date to read about how the model was updated for each revision, and if you're genuinely interested in this stuff, I would suggest doing so.  But the idea you've put forth that IHME has been retreating from a massive overstatement of the death totals is simply not present in the data.

They also talk specifically about revisions based on changing social distancing practices, which, contrary to your post, were not fully baked into the model at the time it was initially released.  In the first release, IHME assumed social distancing would continue as it was at that time, but considering more than 30 states had not even instituted state-wide stay-at-home orders at that time, including some of the hardest-hit states (e.g. FL, GA, PA, MD, RI, and DC), it's natural that there would be some downward revisions as more and more states started.

Here's what that data looks like over time, with the green line representing the prediction, the orange and yellow showing the min and max in the range of outcomes, the red showing the actual numbers at the time the IHME model was updated, and the blue dashed line showing the number of people subject to stay-at-home orders at that time:

image.thumb.png.8c6eff60950ed624b196b636e4d3fd9c.png

 

Now, I'm just a lowly software engineer and not an infectious disease modeler, but knowing that the IHME model was adapting to new social distancing guidelines, and seeing the rising trend in population subject to stay-at-home orders coincides with the downward revisions of the number of deaths, I think the case is pretty strong that the model's downward revisions are a result of including new information not available in the model when it was originally published, and therefore, not indicative of some egregious flaw in the model.

Let's put this in baseball terms.  Suppose we had a normal baseball season, and PECOTA projections come out with the Dodgers predicted to win 100 games, with a minimum of 85 and a maximum of 115.  In May, Bellinger breaks his wrist sliding into 2nd and will miss two months, and then in June, Buehler's elbow starts barking and he's put on ice for a while.  PECOTA incorporates this information into a newly-revised win projection of 90 wins, with a range of 78 and 102.  Is the model really "off by so much"  when the original range of outcomes included a low of 85?

The reason these models use a range of outcomes is to try to incorporate things like injuries to key players, or, in the IMHE case, changing social distancing practices -- things that are "known unknowns" at the time the forecast is produced.  You can call this ass-covering if you'd like, and to some extent it is, but if you're going to say models are "off by so much", you can't ignore the fact that they baked uncertainty into the model from the beginning.

And how much of a difference are we really talking about here in the revised estimate?  The initial estimate on 3/26 was 81,000 dead in the US, while the latest update says 69,000.  That's roughly a 17% revision, hardly something for model skeptics to be taking a victory lap over considering we haven't hit the peak yet, and considering the model's predicting what the number will be at the end of July.

Finally, if you're going to chastise those who made predictions or made policy decisions using those predictions, you should be even more concerned about those who made far larger errors in the opposite direction.  Predictions that were taken seriously by the administration early on had the number of deaths at 500, then 5,000, and are now saying 50,000.  If his new estimate is right, that makes the error of his initial prediction 9,900% as compared to IHME's 17% adjustment.  Which estimate do you think the administration should use going forward?  The one they started with that understated the problem by a factor of 99, or the one that's on target to be within 20% of their initial prediction?

 

  • Like 6

Share this post


Link to post
Share on other sites

I think some posters can count their posts in this thread as their doctoral thesis. 

  • Haha 3

Share this post


Link to post
Share on other sites
2 hours ago, MrBrown said:

What does dentistry have to do with this?

China's numbers are likely under reported.  East Asian countries like to keep everything hush hush.  It's not like in the US where the media reports everything and makes sure to get to the bottom of it. They have also dealt with these pandemics before, and regularly wear masks on an everyday basis, while Western countries have not.

You don't think South Korea and Japan are Transparent?

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

  • Recently Browsing   0 members

    No registered users viewing this page.