3 ms·

The Economist's model is complete nonsense when it comes to China. The Economist trained a black box machine learning algorithm on data from a bunch of countri

The Economist's model is complete nonsense when it comes to China. The Economist trained a black box machine learning algorithm on data from a bunch of countries that let the virus rip, and then blindly applied the resulting model to China. Garbage in, garbage out. From April 2020 through the end of 2021, people in China were out and about, living their lives almost as normal. Nobody knew anyone who was getting sick, much less dying. The hospitals weren't overflowing. Almost nowhere was under lockdown. During this time, the Economist's model spits out a death toll of over a million. As I said, total nonsense. Anyone who works with ML models and statistics can tell you that when you train a black box and it produces silly output, your model is junk. You don't use your model's output to make crazy claims like, "1.7 million Chinese people died, but nobody (including nobody in China) noticed." One final thing: When it comes to China, a lot of people in the West are living in absolute fantasy land. Disbelief/denial that cases in China have been essentially zero for most of the last two years or claims that a million Chinese people have died are examples of this. As someone who actually knows what's happening in the country, it's quite frustrating.

sweezyjeezy 4y ago
I haven't read the article - I thought parent comment was suggesting this is based on excess mortality?
- DiogenesKynikos 4y ago
  Not based on any actual measurement of excess mortality. They're citing a machine learning model that tries to correlate various other measures (like GDP/capita, unemployment, etc.) with mortality in some countries, and then extrapolates to every country. That model spits out ridiculous numbers for China. A normal statistician would look at that and say, "This model doesn't generalize well," not, "1.7 million people died without leaving a trace." The best actual measurement of excess mortality in China was published in the BMJ (formerly known as the British Medical Journal).[0] It found that in the initial outbreak in early 2020, there were just under 5,000 excess pneumonia deaths. Almost all of them were in Wuhan itself. In most of China, there were actually negative excess deaths, because the lockdowns reduced flu and car crash deaths. 0. https://www.bmj.com/content/372/bmj.n415 https://www.bmj.com/content/372/bmj.n415
  - sweezyjeezy 4y ago
    Got it, thanks for this
NeutralCrane 4y ago
> From April 2020 through the end of 2021, people in China were out and about, living their lives almost as normal. Nobody knew anyone who was getting sick, much less dying. The hospitals weren't overflowing. Almost nowhere was under lockdown. During this time, the Economist's model spits out a death toll of over a million. China has 4x the number of people living there as the US. Back in November 2020, approximately 0.08% of all Americans had already died of Covid. That's about 0.01% of America's population dying a month for the first 8 months. And yet if you asked most people in the US if they actually knew of someone who got sick and went to the hospital with Covid, let alone die from it, the overwhelming response would have been no. Most of people's understanding of the toll of the virus was through the media and second hand reports, which is why Covid denialism was able to flourish. And look at case rates today. Thanks to vaccines and immunity, death rates are low, but recent CDC reports have official Covid cases for this week being at the same level as they were in November 2020, with the true number being almost certainly far higher due to the availability of home testing now. And yet American society has largely returned to business as usual, and most people don't know anyone with an active infection right now. Now look at China. A total of 1.7 million Chinese people dying through then end of 2021 would only amount to 0.12% of Chinese people dying, but over the course of 24 months, which is about 0.05% of the population dying per month over the course of the pandemic. If most Americans were still mostly only encountering Covid through the media as late as November 2020, then I imagine it would be pretty easy to feel like things are under control in China when the death rate is half that, combined with strict government control of media, hospitals, and the internet.
- CorrectHorseBat 4y ago
  I don't live in the USA, but since the beginning of 2020 until now basically everyone I know has had covid. No direct contacts I know went to the hospital or died, but I also don't know that many people who are older. My friends in China on the other hand know no one who has had covid, let alone someone who went to the hospital or died. That could maybe be media control, but I don't buy it. They tried that in Wuhan and it miserable failed, why would it suddenly work now? And how do you explain they didn't have mass lockdowns for 2 years and now they are back? These lockdowns are super unpopular and wreck the economy so I don't see why they would do it unless really necessary.
belter 4y ago
"China hiding real Covid deaths by masking cause of death: Report" "Amid a massive surge in COVID-19 cases in China, Beijing is trying to hide the real death toll by masking the cause of death. Citing Financial Times, Taiwan News reported that if someone dies after contracting COVID-19 but had, for example, cancer, heart disease, or diabetes at the time, Chinese hospitals would not classify the death as resulting from COVID-19, but as a chronic illness instead. The faulty methodology was confirmed by Jin Dong-yan, a virologist at Hong Kong University." "The numbers are not accurate, but Shanghai hospitals are not necessarily doing this on purpose. From the start, China had this method of recording deaths," he added." https://www.business-standard.com/article/current-affairs/china-s-hiding-real-covid-deaths-by-masking-cause-of-death-report-122041700536_1.html https://www.business-standard.com/article/current-affairs/ch...
- DiogenesKynikos 4y ago
  First, taiwannews.com is not even remotely reliable for anything related to mainland China. This is the same outlet that claimed the Three Gorges Dam was about to break last year. Taiwannews.com loves disaster stories about China, and isn't afraid to invent them. Second, all these stories about China supposedly concealing deaths in the Shanghai outbreak came out just a few weeks after cases began to rise. Back then, there were only a handful of reported deaths. A few weeks later, there have been over 500 deaths (roughly in line with what one would expect, given the number of infections), and these claims of concealing deaths look pretty silly. Deaths always lag cases, and there can also be reporting delays.
  - belter 4y ago
    "The pandemic’s true death toll" https://www.economist.com/graphic-detail/coronavirus-excess-deaths-estimates https://www.economist.com/graphic-detail/coronavirus-excess-...
    - DiogenesKynikos 4y ago
      I discussed The Economist's model a few comments up in this chain. It's a case of garbage-in, garbage-out. When a black box ML model spits out a highly surprising answer that contradicts everything that is known, your response should be, "Why should I believe this model?"

sweezyjeezy 4y ago

I haven't read the article - I thought parent comment was suggesting this is based on excess mortality?

DiogenesKynikos 4y ago
Not based on any actual measurement of excess mortality. They're citing a machine learning model that tries to correlate various other measures (like GDP/capita, unemployment, etc.) with mortality in some countries, and then extrapolates to every country. That model spits out ridiculous numbers for China. A normal statistician would look at that and say, "This model doesn't generalize well," not, "1.7 million people died without leaving a trace." The best actual measurement of excess mortality in China was published in the BMJ (formerly known as the British Medical Journal).[0] It found that in the initial outbreak in early 2020, there were just under 5,000 excess pneumonia deaths. Almost all of them were in Wuhan itself. In most of China, there were actually negative excess deaths, because the lockdowns reduced flu and car crash deaths. 0. https://www.bmj.com/content/372/bmj.n415 https://www.bmj.com/content/372/bmj.n415
- sweezyjeezy 4y ago
  Got it, thanks for this

NeutralCrane 4y ago

> From April 2020 through the end of 2021, people in China were out and about, living their lives almost as normal. Nobody knew anyone who was getting sick, much less dying. The hospitals weren't overflowing. Almost nowhere was under lockdown. During this time, the Economist's model spits out a death toll of over a million. China has 4x the number of people living there as the US. Back in November 2020, approximately 0.08% of all Americans had already died of Covid. That's about 0.01% of America's population dying a month for the first 8 months. And yet if you asked most people in the US if they actually knew of someone who got sick and went to the hospital with Covid, let alone die from it, the overwhelming response would have been no. Most of people's understanding of the toll of the virus was through the media and second hand reports, which is why Covid denialism was able to flourish. And look at case rates today. Thanks to vaccines and immunity, death rates are low, but recent CDC reports have official Covid cases for this week being at the same level as they were in November 2020, with the true number being almost certainly far higher due to the availability of home testing now. And yet American society has largely returned to business as usual, and most people don't know anyone with an active infection right now. Now look at China. A total of 1.7 million Chinese people dying through then end of 2021 would only amount to 0.12% of Chinese people dying, but over the course of 24 months, which is about 0.05% of the population dying per month over the course of the pandemic. If most Americans were still mostly only encountering Covid through the media as late as November 2020, then I imagine it would be pretty easy to feel like things are under control in China when the death rate is half that, combined with strict government control of media, hospitals, and the internet.

CorrectHorseBat 4y ago
I don't live in the USA, but since the beginning of 2020 until now basically everyone I know has had covid. No direct contacts I know went to the hospital or died, but I also don't know that many people who are older. My friends in China on the other hand know no one who has had covid, let alone someone who went to the hospital or died. That could maybe be media control, but I don't buy it. They tried that in Wuhan and it miserable failed, why would it suddenly work now? And how do you explain they didn't have mass lockdowns for 2 years and now they are back? These lockdowns are super unpopular and wreck the economy so I don't see why they would do it unless really necessary.

belter 4y ago

"China hiding real Covid deaths by masking cause of death: Report" "Amid a massive surge in COVID-19 cases in China, Beijing is trying to hide the real death toll by masking the cause of death. Citing Financial Times, Taiwan News reported that if someone dies after contracting COVID-19 but had, for example, cancer, heart disease, or diabetes at the time, Chinese hospitals would not classify the death as resulting from COVID-19, but as a chronic illness instead. The faulty methodology was confirmed by Jin Dong-yan, a virologist at Hong Kong University." "The numbers are not accurate, but Shanghai hospitals are not necessarily doing this on purpose. From the start, China had this method of recording deaths," he added." https://www.business-standard.com/article/current-affairs/china-s-hiding-real-covid-deaths-by-masking-cause-of-death-report-122041700536_1.html https://www.business-standard.com/article/current-affairs/ch...

DiogenesKynikos 4y ago
First, taiwannews.com is not even remotely reliable for anything related to mainland China. This is the same outlet that claimed the Three Gorges Dam was about to break last year. Taiwannews.com loves disaster stories about China, and isn't afraid to invent them. Second, all these stories about China supposedly concealing deaths in the Shanghai outbreak came out just a few weeks after cases began to rise. Back then, there were only a handful of reported deaths. A few weeks later, there have been over 500 deaths (roughly in line with what one would expect, given the number of infections), and these claims of concealing deaths look pretty silly. Deaths always lag cases, and there can also be reporting delays.
- belter 4y ago
  "The pandemic’s true death toll" https://www.economist.com/graphic-detail/coronavirus-excess-deaths-estimates https://www.economist.com/graphic-detail/coronavirus-excess-...
  - DiogenesKynikos 4y ago
    I discussed The Economist's model a few comments up in this chain. It's a case of garbage-in, garbage-out. When a black box ML model spits out a highly surprising answer that contradicts everything that is known, your response should be, "Why should I believe this model?"