So it looks like you're saying that not many in group 1 die in 2020 because only a small amount of them will become infected? — Punshhh

Well, I'm relating this group to my discussion with Isaac.

If this group was very large and most deaths from Covid came from this group, then Covid deaths displace near-future deaths.

If this this group is small and people are dying of Covid outside this group then this effect is small. If, for our purposes of decision making now, this group has not even gotten Covid much, then the effect so far is even smaller and so even less likely this group is displacing near-future deaths.

However, if your question is simply if we can be 95% sure people who are destined to die in 2020 would die of Covid if they get it, then no we can't make that assumption. It doesn't really matter what we assume, as they are going to burden the health care system either way, but we'd have to know more about these people; maybe they have some terminal disease that doesn't affect their resistance to Covid (they experience Covid as just a cold and go onto die from a heart attack this year anyways).

So you are happy with there being a group (1), which is a small group, who are destined to die in 2020 due to another medical condition, comorbidity. With an overlap of 95% or more, who have contracted Covid, dying due to Covid. — Punshhh

No, because not everyone in this group gets Covid.

It works in reverse, 95% of people in this group who get Covid we might reasonably expect they die of Covid, and certainly all of them within 2020 (as we don't expect Covid to somehow cure them).

But, the whole group doesn't get Covid. So far estimates are 1 - 15% of the population actually has Covid ... or exposure to Covid, even in hotspot areas (exposure, even if real and not a false positive, may or may not provide good immunity). So, even with the upper bound of 15% got Covid, 85% of people in this group don't die of Covid and do die of something else, so their burden on the medical system remains 85% (and they are a small group to begin with).

As more people get Covid, more of this group also get Covid, but more people in other groups get Covid too. The ratio remains more or less the same (for decision making purposes about projecting health care burden) absent an extreme bias for this group to get Covid (for which there is no plausible mechanism, considering geographic constraints alone).

You are happy with a group (2), who have an underlying medical condition, comorbidity, but who are not destined to die in 2020, they may die in 1, 2, or 10 years of these conditions. That this is a large group, and that a large proportion of these patients will die in 2020 if they contract Covid. I estimated that 60% of these who contract Covid will die. — Punshhh

This is not justified by what we know so far.

The basic pattern is Covid doubles your risk of death this year. Most people who have a risk of death "within 10 years" don't have 30% risk of death this year and therefore 60% risk of death with Covid this year (which is still not 60% chance of death from Covid). If a person of high risk of death with in 10 years has 5% risk of death this year, then their risk of death of Covid seems to be also 5% (therefore 10% within the year).

Why we see Covid deaths overwhelming medical system is that large numbers of people have a low risk of sever complication of Covid.

Most individual deaths within a year are not well predicted individually at the start of the year. Being in a risk group of 10% chance of death this year is a very high risk group.

Actuary science and medical science places a large importance on random variation of environmental factors, internal factors, life choices, life events, medical intervention, accidents, family support, etc. in causal determination of who will actually die within a year time span. In short, there are large groups of which a small portion of them will get "unlucky" within a year time-frame (with a fundamental inability to make a better prediction at the start of the year regardless of amount of tests, measurement, modelling, crystal balls or any other predictive device).

Definitely, these groups from which deaths happen "randomly" are heavily weighted towards being old and / or having underlying conditions, but these groups are still very large for our statistical purposes here of estimating overlap between the set of Covid deaths and the set of "otherwise would have died this year".

The reason that the group of people we are pretty certain will die this year is small, is because (for most people in wealthy society not at war) to get to ~90% chance of death this year (such as 90 year old with dementia and failing heart) meant having a ~80% chance of death last year (due to slightly less sever dementia and failing heart), and ~70% chance of death the year before that, and so on (though these numbers will depend on disease, there is not a large group that had 0.1% chance of dying last year but 90% chance of dying this year). So most people "getting old", the group they are a part of was already reduced significantly each year prior to getting to 90% and in a accumulative way: starting at about 5% chance of death this year, these risk-groups thin out very rapidly in a 10 year time span (each year they lose members and the chance of death of remaining members increases); during this time span most deaths are not well predicted individually (of 20 people, one of them dies the first year; maybe in a way that makes sense in hindsight, but there was no way to predict which individual would die at the start of the year).

This is why from 60 to 100 years of age the demographic chart is nearly a straight line to almost 0 population at 100, but it's a fat bar until 60 (variations in birth rates and immigration can dominate death rates below 60).

Also, why I keep coming back to the fact the discussion is about a year time frame.

If we were talking about dying within a 20 year time frame, the overlap can easily start approaching 90% for the exact same reasons (relatively high-risk groups rapidly thin out on decade time scales; just not 1 year time scales). Overlap between dying of Covid or "otherwise would die within 20 years", which is not to say people who will die within 20 years are "very likely to die of Covid".

In a 1 year time frame -- which is relevant for estimating health care resource needs and other policy choices -- overlap is low if a disease affects large risk groups (such as people in their 60s with hypertension). If a random (otherwise benign) genetic difference is also a good predictor of death from Covid, then the overlap is even less strong as Covid doesn't tend to select for "worst heart" within these risk groups but it's random genetics that dominates chance of death from Covid between risk category peers, leaving survivors to be just as likely to die of heart failure as before (perhaps more so due to long term lung injury), and also explains why Covid can kill completely healthy people, as perhaps they just have bad genetic luck (maybe Covid exploits particularly well 3 uncommon gene variation; then it could be if you have all 3 genes you have a 90% chance of death even if healthy, and it's quite rare to have them all, but happens) so is consistent with "gene variations matter hypothesis" and this hypothesis is consistent with the dominant medical theory.

Unless there is a very clear pattern that would be obvious by now (you only die of Covid if you not only have hypertension but have already had a heart-attack, or if you are on immune suppression therapy or otherwise severely immuno-compromised); absent such patterns, "risk-factors and genetics" is the go-to explanation for why some populations survive a selective pressure and some don't, without good individual predictors available at the start: it's how evolution usually works, so no reason to assume it's not happening with Covid; i.e. it obviously doesn't help to be obese or have hyper-tension or smoke or be old and frail, but the virus maybe only particularly lethal with certain particular kinds of proteins on cells; i.e. certain genes or particular epigenetic gene expression or specific immune system history (such as getting or not getting some particular common cold in the past by random chance).

Therefore, it's more reasonable to assume there is large random variation determining individual deaths from Covid from among large risk groups (as this is the pattern we actually see) and subsequently assume that deaths from other causes will continue, perhaps even increase (for the purposes of decision making), than to assume actuary and medical science is wrong (foundationally, not just some specific issue) and there are some hidden variables that dominate the real determination of both individual death each year, for instance not just heart disease but having "the weakest heart" (but in a way we can't measure), and individual death from Covid (and these hidden variables would need to be the same in both cases to boot; the hypothesis is implausible, and even if true, it's still implausible) resulting in Covid deaths tending to lighten the burden on the health system from other causes of death (as those deaths are now dead from Covid, no longer available to die of something else).

Of course, there can be second order effects that actually do reduce deaths (people drive less and therefore there are less accidents) but this has nothing to do with the statistical overlap discussed here but a consequence of our response to Covid (it's entirely reasonable for modelers to estimate less traffic due to lock-downs, and therefore less accidents and therefore position less traffic accident resources; this was an obvious lock-down health-care hypothesis that has already been proven true; but another hypothesis that people who need care for other things don't get care and therefore die at a higher rate also seems to be proven true).

if we look at a group in the population who are ill with one of the comorbidity diseases who would be destined to die in 2020. Some of those will die prematurely due to a Covid infection. I would find it hard to believe that many of these patients would survive Covid, only to die later in the year, so the overlap will be large, say around 95% ( of those who become infected with Covid) — Punshhh

I agree if you are destined to die in 2020 you are "even more" destined to die in 2020 if you get Covid.

This group I have been calling "would otherwise die within 1 year" or "terminally ill" interchangeably.

The first problem is that not all terminally ill people will get Covid. For instance, if only 15% of people have Covid so far, then there's 85% of these terminally ill people out there, absent a selector that makes these terminally ill vastly more likely to get Covid. If we look simply at the fact Covid progresses geographically then we already know the selector to get Covid of a "well mixed" sub-population is weak, because they are not all clustered geographically; yes, they may cluster around hospitals within their individual regions, but the disease still progresses geographically.

Furthermore, we know who the at risk populations are and we take additional measures, so this also weakens the selector.

Therefore, if 85% of terminally ill people are still out there, they will still die in 2020.

And that's an upper-bound of total infected. The lower bounds is as low as 1% infected (confirmed Covid infections) in which case 99% of terminally ill are still out there and will still die in 2020.

So, although we can assume terminally ill people who get Covid will more likely die even sooner than Covid somehow having the opposite effect and curing them of their underlying condition, for this to create a big effect of simply moving deaths around within the year then we need to have reason to believe this entire population gets Covid with extreme bias (and there's solid reason to not assume that's no so).

There is a second group who are ill with the same illnesses, but who are not destined to die in 2020. A proportion of thes patients will die in 2020 after contracting Covid. I would expect the overlap here to remain high, but not as high, say 60%.(of those infected with Covid) — Punshhh

I read this to mean that 60% of total Covid deaths are from ill people, just not ill enough to die in 2020.

Although I agree most people who die of Covid have underlying conditions, the reason to believe most Covid deaths are not from people who "would have otherwise died" within 1 year, is because the vast majority of deaths each year are not from people who doctors are certain will die within the year.

Lot's of people who have 1% chance of dying from a heart attack have just that, a 1% chance. At the start of the year you can test them however you want, but you couldn't have done any test to determine a greater than 1% chance. The reason is that random things (from the perspective of the start of the year) happen: stressful life events, poor response to treatment due to genetic variation or "bad batch" of pharmaceuticals (quality control exists because processes aren't perfect, including quality control), taking up drinking, unlucky torque on an artery, immune system "learning something" by a lucky stochastic result.

If someone with a 1% chance of death has a prognosis of 1% chance of death from Covid if they contract Covid (what the evidence Isaac cited broadly indicates), then this is the "large risk" group situation I have been talking about and overlap with "who would otherwise die within 1 year" is small if Covid proliferates in these risk groups.

If you look at the risk groups people are in, they are these very large risk groups with around 1% dying per year (increasing with age and severity / number of conditions).

Very, very few people are in a risk group of 90% chance of dying this year. So if everyone got Covid, yes, all these people would die, but there are few deaths because it's a small group.

Very, very large amounts of people are in groups with < 1%, 1%, 2% up to about 10% (with decreasing total numbers). Most deaths per year are due to a very large number of people having a small chance to die, resulting in still a large number. Nearly all these people are above 60, but the are still in large groups that are not otherwise expected to die within 1 year.

The prevailing theory of medicine and actuary science is not that there are hidden variables within the body that actually explain who dies and who doesn't, but rather that variations in environment, disease progression, immune response, life choices, doctor actions, timing of intervention and dosage, etc. that determine who lives and who dies without any ability to predict these things much better than we currently do on a 1 year time frame.

Isaac's position relies not only on these hidden variables, but furthermore that these hidden variables are the same between determining "who actually die from Covid" and "who actually dies from underlying condition like heart disease".

It makes sense on the surface that "the weakest" would die in each case, but this is exactly the opposite idea actuary science is premised on; it's not true that the "weakest hearts" die each year (lot's of other factors involved), and even if it was true (which is isn't) it can easily be something else that drives Covid deaths within risk-groups, such as an otherwise benign genetic variation (subtle protein differences that don't have any difference until now, but Covid exploits that difference particularly well) that helps the virus proliferate faster (epidemic resistance is a classic reason to explain why genetic variation is a good thing). A genetic sub-group particularly susceptibility to Covid explains very well why we still see deaths even in seemingly healthy people.

There is a third group who were destined to die of a disease in 2020, but who presented as quite well, but who will die unexpectedly in 2020. Of this group there may, or may not be an overlap, if there is I expect it is quite low, say 10, or 20%.( of those infected with Covid) — Punshhh

This is statistically impossible to reach 10 - 20%; that would be a huge overlap for a group of people who's "otherwise death in 2020" shares no causal mechanism at all to their Covid death (as they were simply not expected to die).

For instance, I think we could agree that dying of Covid won't somehow preferentially select for people who would otherwise die in a car crash.

This is simply the "base case": let's say a person has 0.5% chance of dying from Covid and 0.5% chance of dying from something else had they lived (so the people that do die from this group represent "unexpected" deaths), so it's a simple "choose 1 out of 200, choose another out of 200, what are the chances the choice is the same item?" which is simply 1 out of 200, a small effect of Covid deaths overlapping "would have otherwise died deaths" in this case (and an effect easily compensated for by increase in death probabilities due to lung damage from Covid, interruption of quality care for many risk groups while the medical system deals with Covid, or other things that can have a small forcing on large groups; decrease in pollution and changes to stress patterns may push things the other way).

Of course, the probabilities don't need to be the same (but they remarkably line up pretty well with chance-of-death by year by age group), but if they are small then overlap is very small. If they are probabilities that apply to large groups, "like all 60 year old's that seem healthy and not expected to die this year" then you still have large numbers because these are large groups.

Furthermore, if Covid deaths would be happening in a short period of time, instead of over a year, then even groups that have low probabilities of dying from Covid, they would still overload the health system as they arrive in a short period of time (why the idea of trying to protect over-70s and letting everyone else live normally made no numerical sense).

No! Who the hell thinks people over 60 are at the end of their lives. I bloody hope not. — Isaac

The point is, you don't know.

It's completely reasonable to say people at 60 are closer to the end of their lives than people at 20, more so people above 70 or 80.

It's entirely possible to talk about people "at the end of their lives" without meaning "people that will be dead within 1 year", just meaning that old people, by definition, are usually closer to the end of their lives than anyone else.

For a technical expert, it's entirely accurate to talk about old people as "close to the end of their lives".

Yes. In the context (and supported by David Spiegelhalter, who specifically referred to 2020). I'm quite confident "end of their lives" meant they they were close enough to death to fit mostly in the year's mortality. — Isaac

Quite confident based on nothing.

Read his papers, if he had statistical evidence for this, he would have included it in his model, as his model papers are all about health care capacity based on his best use of the statistics available to him.

For instance, he discusses the possibility that there's a very large amount of asymptomatic infected, and explains why (despite a large potential variation in this factor) it can't possibly be high enough to change the main conclusions of his paper and the requirement of social distancing to keep within health care capacity for a significant amount of time.

If he thought a large portion of people who were dying would have otherwise been dead within the year, that's very significant, and he would have included a model or at least some discussion of what that would mean.

The statement you're referring to is also clearly in the context of social distancing working to keep deaths below 20 000 for the whole UK; so, we can understand it to be a feature of that specific scenario, not a feature of Covid if left to proliferate uncontrollably (which his model of a unmitigated spread cannot possibly be interpreted to kill everyone who otherwise would have died within 1 year anyway, not even close). He also just says "maybe" in the sense that it hasn't been completely excluded yet, an upper bound without any reason to assume things will be anywhere close to that upper bound in reality.

So, he is not lending support to your position, just didn't completely exclude it yet within the context of a social distancing scenario that the UK has already passed.

You can't take one statement (not even in a paper but an interview) of an expert, out of the context of where they said it, interpret it wrong (confuse pre-modelling guesses of upper bounds and "likelihood"), not consider their published papers on the same subject, and call it evidence supporting your position (well you can say it is, as you've been doing, it just isn't actual evidence).

What? If I can't cite evidence he meant within exactly one year then that somehow counts as evidence supporting your position? — Isaac

The evidence you cited is that 90% of cases have comorbidities, that is not evidence that 90% (or anywhere close to that) "would have died within 1 year". You've cited risk factors for large groups.

The larger the risk-group Covid is affecting (where both the preexisting condition risk and Covid risk of death is low), the smaller the overlap between people who die from Covid and those that would have died anyway. You are citing evidence that supports my position.

You do not have the technical ability to understand your mistake. You don't want to be taught by me; fine, but your unwillingness to learn doesn't impact my willingness to defend my position.

You make sense (most of the time, and even when I do not agree, you've provided something meaty to disagree with - but I wish most of your posts were shorter). — tim wood

Thanks you for appreciating my arguments; unfortunately, it takes a lot more work to make an argument shorter. I too appreciate people who disagree with me but are being honest with their views at any given time; not just how we learn, but also how we learn what humanity is really like.

My own view is that Joe Biden might just be the second worst possible candidate, but he's running against the worst, and between them there's no comparison. — tim wood

The problem with the lesser of two evils argument (why Hillary didn't easily win) is that it is the nature of evil to be deceptive, so it's not logically possible to "know for sure" how evil compares to evil. So, you can always imagine the "second worst option" has hidden things making them actually worse.

So, as soon as you admit to using a lessor of two evils approach, there's no solid argument to make that it's true. It's entirely consistent to suspect being just better deceived by what appears as the 2nd option.

What happens is that people with a bias one way can just go with that bias, and people with a bias another way go with their bias. A Democrat saying to a Republican "look, I've got the lessor of two evils here", it's completely reasonable for the Republicans to worry it's a trick.

To be clear, most Americans agreed Hillary was the best of two bad options, but for US elections we need to contextualize things in the fact minority popular vote can win the office.

Republicans have been playing the "win with a minority" game really well. For Democrats to win they need to overcompensate this disadvantage, and this is a difficult game to play; to succeed when the odds are stacked against you, requires accumulating every advantage possible: using the primaries to get to the strongest candidate available; even if that means a brokered convention where progressives have some power.

This is why I am so harsh on the DNC here; there was no need to orchestrate all the candidates dropping out to rally around Biden, use Warren to split the vote, in a panicked backroom horsetrading coup, there were other strong candidates relative Biden; and if a brokered convention would select Bernie, maybe he's just the strongest candidate and Bernie in the white house is not the end of the world. Furthermore, Bernie has serious problems too, mainly being super old now, so may have been willing to support a younger compromise candidate. In otherwords, the DNC could have chosen to engage with politics.

Indeed, had the DNC not orchestrated a premature end to the race (there's only a "clear winner" from everyone else dropping out), then there would have been younger candidates in this time of Covid. Both Bernie and Biden are in the high risk group for Covid of nearly 80. How do you campaign in isolation? How do you campaign without isolation if it may kill you?

Younger candidates wouldn't have had to worry so much, giving rise to the possibility that both Joe and Biden agree on a compromise candidate and "pass the torch". A moving moment that brings the Democrats together. Instead, the DNC orchestrated a strategic catastrophe.

Not that I'm saying the election is already decided, but it's really depressing to see a genuinely vibrant primary with lot's of good points of view and candidates, narrowed to just Biden for no reason ... then Covid happen (which was already inevitable for anyone paying attention) and Biden in even worse position (the weakest on healthcare, easiest to attack, old and very vulnerable to Covid itself).

hat is striking, i.e., worthy of notice, is how the attacks coincide with events. Biden a viable candidate for president of the USA? Time to run ads accusing him of being venal, corrupt, a serial rapist/sexual harasser/abuser/pedophile. The irony is that's almost Trump's exact curriculum vitae. — tim wood

Definitely, the irony can't be more palpable.

Yes, but your premise is not true. Having a comorbidity of sufficient severity to class as a cause of death is not a "large risk-group" it is, as the country's leading expert in the field has said "people at the end of their lives". — Isaac

"End of their lives" as in over 60?

Or, "end of their lives" as in will die within 1 year?

You can't just substitute meanings all over the place to pretend your position has been my position all along.

All my arguments have been about this 1 year time frame.

So, please show where this expert clarified their meaning of "end of their lives" as to mean "would have died within 1 year". Otherwise, again, you are citing evidence that supports my position, not yours.

That's how I understand it.

Also note, neither @ssu nor I characterize all Americans as extreme partisans, but rather that extreme partisanship (where small extreme groups hold disproportionate power) is a feature of the US electoral system.

The relevant issue vis-a-vis Joe Biden is that Trump has more extreme partisans than Joe Biden, and the reality is that criticism of Joe Biden will more likely land simply because less people on the left are an extreme Joe Biden partisan. In otherwords, Joe Biden is a terrible candidate for the current electoral mood and for the fact the internet is very much a thing now (people can watch Joe Biden with children; the mainstream media ignoring the issue doesn't magically make the issue not matter).

And more generally, the extreme partisan game is simply less powerful on the left. Lot's of people on the left had legitimate corruption concerns about Hillary, because there's simply important evidence about her foundation, purpose for even having a private mail server in her basement, and "private and public position" ideology (private position ... benefiting who?). There were extreme Hillary partisans that engaged in the same reality denying games as Trump supporters do now, but they were a minority; so, for a significant part of the US left the argument for Hillary is "a lesser of two evils", and maybe analytically correct (who know's; certainly Hillary would have been better for US empire than Trump, but US empire may not be ultimately a good thing and Trump is doing the Lord's work by dismantling it through inept management and extracting value from it for personal gain wherever possible), but, in terms of winning elections, "lesser of two evils" is not a motivating reasoning, so even if it's correct it may not help you actually win.

The right doesn't have this problem. A larger proportion of Trump supporters believe he's great, and there's lot's of positive reasons to vote for him on top of the democrats being crazy socialists, more swampy, or pure evil.

It's a big advantage, and the best way to compensate Trump's advantage is with a good candidate that genuinely can deal with criticism due to clearly not being corrupt, womanizer (touching lots of women uncomfortably as the Times reported and then deleted), bizarre child "doter" (why Trump was most afraid of running against Bernie; these criticism don't land on Bernie, they do land on Biden because there's lot's of smoke and thus reason to suspect fire).

Apologists for Biden already arguing "yeah, maybe he sucks, but he's better than Trump"; maybe they're right, maybe wrong (maybe Biden is king of some pedophilia cult and Trump is, despite incredible moral failings not also a pedophilia king, and is, incredibly, the lesser of two evils), but what I think we all agree on is that running a weak candidate is not a good strategy to win.

If I criticize Trump I have TDS? Criticism is manifestation of craziness? That's some generalization. — tim wood

I think @ssu is saying that polarization and "post-truth" politics leads Trump supporters to dismiss all criticism of Trump with the euphemism of TDS.

Yes, I agree.

The DNC leadership is old and simply hasn't got the feel to the pulse of the nation. It genuinely lacks vision and understanding of it's voters and the situation. (Neither did the GOP either actually: Trump was just a train wreck that suddenly caught the party by total surprise with even a bigger surprise that he won.) It lacks ability to get people excited. — ssu

Both parties (pre-Trump) continued to function on the idea the "internet just doesn't exist", and they simply ignored it.

Fox News has fallen in line now, and internet Trump groups now drive Fox news rather than the other way around, but there were multiple times Fox tried to bury Trump and just not talk about him anymore.

Conservative media have accepted their place on the totem poll Trump > Trump online networks > Them.

Why they accept it is that, though they don't like Trumpian politics (they prefer a polite and "respectable" mascot), there is no actual alternative governing ideology on the right; so it's not really a threat to have Trump and online Trumpians drive the discussion. The slogan for this group has become "I don't like Trump, but I love his policies! Tax breaks, woohooo!"

The problem with the progressives is that they do have a coherent alternate governing ideology ready to go. This governing ideology has been proven to be workable in dozens of countries (still lot's to debate, lot's of policy variations, but it clearly can work).

The "job" of democrat centrist is to keep this ideology out of government.

With Hillary, progressives were split on the subject of whether the DNC is ready to lose power to fulfill their donor mandate to keep progressive ideas out of real power.

Yes, Hillary had a lot of flaws and legitimate examples of (highly likely) corruption such as her foundation shenanigans. But she did have a lot going for her, such as she won a majority of women voters and won the popular vote.

But already, nearly all the progressive American voices I listen to are unanimous that the DNC would rather lose with Biden than let progressives share power in any meaningful way.

A key point of course is that they orchestrated a "rally around Biden" right before super Tuesday (but keep in Warren just long enough to split votes with Bernie), not to mention Ohio had clear admitted to vote rigging (DNC lawyers just claimed it was fair vote rigging to make simple math errors in summing votes, as it's in broad daylight).

Not to say Bernie had no chance of winning despite these odds, but DNC preference is clear they don't want a situation where progressives have any power (they'll change votes, and do whatever necessary to avoid a contested convention with progressives).

The writing's on the wall of course, younger generations use the internet, but it's clear they will hold on tooth and nail to power, even if it means playing second fiddle to Trump.

I guess the only way for Joe Biden to win is to pick a progressive vice-president nominee, perhaps Elizabeth Warren or even another geriatric, Bernie. Otherwise they really can loose. — ssu

I'm not sure.

A good running mate (the left actually likes) helps, but the general wisdom is that it mostly comes down to the candidate. A good running mate adds more momentum to a good candidate, but doesn't really help a bad candidate.

Of course anything can happen (even replacing Joe somehow), but as it stands, Joe seems like a long shot candidate. It's repeating Hillary 2016 but somehow worse.

300,000 people die each year (from disease). These deaths are drawn, in the overwhelming majority, from the exact group of people who would have the comorbidities listed in the ONS figures as having a 90% overlap with Covid-19 fatality. I've supported that assertion for heart disease and cancer by providing studies of risk factors and prognosis. — Isaac

This is the error in your analysis I've pointed out like 5 times already.

We're talking about deaths within 1 year, so talking about overlap with comorbidity in larger groups than "likely to die within 1 year" supports my position.

The "heart disease" risk group is larger than "people going to die of heart disease within one year from heart disease".

For your claim to be true, there would have to be little overlap with this group. — Isaac

No, my claim is completely compatible with these facts.

Lot's of deaths (most deaths) each year are not predictable at the individual level.

At the start of the year, we cannot predict with any degree of certainly in the sense of individual identification who will be dead by the end of the year.

Most deaths within 1 year do not come from groups with 90% chance of death this year. There are such groups, but they are small and so even 90% of such people dying within the year is not a big number.

Lot's of effort has gone into this, as doctors and life insurers would like to know, but they don't know. What we know is that everyone has a chance of dying, that chance varies and can be statistically investigated, our understanding always improved, but the pure element of chance (relative our knowledge at the start of year as well as just the nature of reality) is also at work.

If statisticians put someone in a group of 1% risk of death due to heart disease this year, they are not saying that they were just too lazy to analyse further and see which of these people with heart disease have actually quite strong hearts (and so many 0.1% of dying) and which have "the weakest heart" (and so 90% of dying); they are saying "of 100 people in this group we expect 1 to be dead by the end of the year, but we don't know which one". Further analysis can make some progress, but does not fundamentally change the fact that most deaths are from groups with small chance of death within the year, but they are large groups and so result in lot's of deaths.

Statisticians of these sorts of things are constantly doing analysis to see if there are other predictors, and sometimes new predictors are found and new risk-groups created, but things are no where close to predicting "who's going to die within 1 year".

That's why your argument depends (depended) on some hidden variable we do not know, such as assuming the people who would die from heart disease this year have "the weakest heart" and the people with heart disease who die from Covid too have "the weakest heart". This is not what statisticians believe. Certainly, such a hidden variable is there that might be uncovered by better medical tools or perhaps is fundamentally hidden for ever, but there is also a large amount of random chance that goes into who dies or doesn't within a year of heart disease.

No. "Complicate the figures" is not anywhere near "replenish the entire cohort". Again, there is no evidence that lung damage will cause future deaths in these numbers. This is just your speculation and needs evidence to support it. — Isaac

You have not bothered to understand my argument.

It's you that has been claiming that the overlap is big, so big that Covid maybe just a problem of reallocating resources and does not require new net-resources.

I have been arguing that the overlap is small with "people who would die this year anyways"; small in the sense that Covid is not just a resource allocation problem even if you could reallocate without friction from other health resources that having nothing to do with respiratory disease (which you can't), and even if you could reallocate from the future to the present (which you can't, even with loans).

I have also mentioned, that even if you are right about overlap, the risk group might be replenished due to long term lung injury.

No where have I said it's guaranteed or I know it to be true. It is, however, a risk, a big risk, and therefore no reason to change policy even if your overlap hypothesis was true (which it isn't).

Furthermore, the effect of replenishing the risk group can be very small but still result in replenishing the risk group (a small thing that affects a large amount of people). If a risk group is 1% risk of death of heart disease and 1% risk of death of Covid (if infected with Covid), and they all get Covid and 1 person dies of Covid, then there's a decrease in expected deaths in absolute terms within the year based only this, due to that 1 person no longer in the group, so now there's 1% risk of death for 99 people, so 0.9 expected deaths from this group within the year.

Long term lung injury only has to increase the risk of death in this group by ~0.1% to replenish the risk group back to resulting in 1 expected death within the year; so 2 deaths within the year (1 from Covid and one from the other risk, such as heart disease) instead of 1.9 deaths due to the slight culling effect Covid had on this group.

This is why overlap has to be with small groups that have super high chance of death to not only see a culling effect but also for the long-term injury effect to also need to be very large. 0.1% increase of death in a group that has 90% chance of death within the year changes little in absolute terms.

If everyone, or most people, gets Covid, and most deaths arise within large risk-groups, then a very slight increase in chances of death due to surviving Covid can easily replenish all the risk groups to result in the same amount of deaths in absolute terms within the year.

You've been basically wrong at every level of your argument, and now that you're beginning to realize this, you are trying re-interpret things to arrive at my position.

Welcome to my position.

I agree with the major points.

However, the Trump camp really is made up of people with beliefs that seem caricatural. Granted, they have conditions and concerns motivating these beliefs that we can empathize with, but there simply isn't any significant amount of Trump supporters that can't be described as "the most craziest most eccentric view" around.

Trump is simply crazy and eccentric, from a governing point of view -- not necessarily from enriching himself and cultivating a cult following point of view. For instance, tweeting out support for insurrection against the government that he is the highest official of. For his cult though, it's just bad ass and a great move.

But I only wanted to mention this to point out Joe Biden doesn't have a cult following and so the mudslinging is not symmetric.

Joe Biden and his supporters simply don't have the ability to neutralize genuine criticism. Republicans respond well to "well, their corrupt too!" without realizing that if they don't do anything to remove corruption on their side they are de facto supporting it. People on the left simply don't respond well to the argument that "corruption on the other side excuses corruption on our side", they don't want to see corruption on either side and they can do something about the corruption on their side.

Furthermore, the ideologies that led to simply legalizing most of what corruption means originated with the Republicans; it's Republican judges that ruled money is speech, that gerrymandering is a "political issue" that judiciary can't remedy, and that bribery cannot be implied with a wink and nod, or giving money while having a public campaign making it clear what you want, but requires explicit recorded quid pro quo. So, when Republican politicians take advantage of legal corruption it doesn't create a sense of frustration for Republicans, these were all "Republican victories"; it does, however, create frustration on the left regardless of who is doing it.

So yes, the right is also frustrated with corruption, but they no longer really have an idea of what corruption is and why it's bad. Corruption on their side is easily viewed as "winning". An example outside of politics is the multi-millionaire pastors who raise money to buy private jets; all of these pastors are on the right, essentially campaign for Trump, point out Trump's just winning like their winning, as they both have God's blessing.

Whereas on the left, not only are there no multi-millionaire pastors herding leftists around -- though it would be a mistake to say there are no Christians on the left -- but there's pressure on politicians to "not take corporate money"; this was a big part of Bernie's identity and argument. Corruption in a reasonably defined way, legal or not, is a big issue on the left.

So, why did the establishment choose Joe? Why can't the DNC find a candidate that is centrist but not easily accused of corruption, perhaps legal corruption, but still corruption?

This is the heart of the DNC problem, they are the party of "can't we have bit of the corruption" and so they, basically by definition, can't easilly find politicians who want to defend the corrupt setup of the status quo but who aren't themselves corrupt. The Republicans realized that, corruption long term is only sustainable if you build ideologies where the corruption is seen as a good thing, whereas the Democrat establishment have not done so, they are just an elite aristocracy detached from their constituents; but they can't just step aside, that's just not how corruption works, so they are content to just lose thousands of seats in State level legislatures, lose to Trump, lose the supreme court, let the republicans play hardball in every inning and just whimper around in response, and content to lose to Trump again.

3 days ago, I made my position very clear:

I'm not sure you're even arguing / implying something against what I emphasize above, or are just compiling all the statistical minutia of relations to consider.

In terms of adding to the list, a big one that can not only nullify the affect of high-risk groups decreasing in absolute size (due to dying), but actually reverse that tendency, is that the virus may cause long term lung damage.

So, if every 70 year old got the disease, all else being equal, we may expect that demographic cohort to have less deaths post-pandemic, simply due to their numbers being smaller or perhaps particularly weak breathers being culled from the heard. However, if long term injury increases the risk-of-death factor for survivors of Covid, then you may end up with more deaths in absolute terms next year due to lung injuries or other long term Covid treatment complications. (likewise for every other demographic cohort)

Long story short, some Covid deaths would have died anyways, but expected overlap is small (extreme bias towards this group getting Covid would be needed for a significant overlap), and long-term injury may compensate, even significantly over-compensate, this overlap by increasing the risk-of-death factor for these risk groups (indeed all risk groups). — boethius

I gave you the benefit of the doubt that you were saying the same thing.

If you're saying the same thing now, you've wasted your time and made yourself look like a fool. Though you haven't wasted my time fortunately, for it is true, as @Punshhh suggests, that I enjoy debating on a debate forum, which is why I come to a debate forum to debate.

For instance if you mean by:

What we know is that the vast majority of fatalities (over 90%) had other comorbidities which were "mostly likely to be the underlying cause of death for a person of that age and sex had they not died from COVID-19". so this is referring to cause of death at the time of death. — Isaac

That 90 percent of Covid deaths had comobidities, but that does not mean 90 percent will die within 1 year. Then you agree with my original position!

If by:

I agree that the complicating factors of system overloading and long-term lung damage make the figures difficult to say with certainty, — Isaac

You mean that regardless of overlap, long term lung damage may simply replenished those "at risk of dying within 1 year" then you agree with my original position!

Maybe you aren't here to debate, which sometimes means recognizing a change in position and simply saying so. Maybe you just want to have pat you on the back all day. That can be done in private, why bother us about it?

Please do stick around to discuss this, your contributions are valued. I think you unfortunately chose to dig a little deeper with the wrong interlocutor. Boethius is quite argumentative, he seems to enjoy it. But this might result in a failure to reach consensus. — Punshhh

Poor, poor @Isaac, comes to a debate forum, engages in a debate, get's served with a debate.

None of the points I have made is motivated simply to disagree with Isaac.

I genuinely believe the effect of overlap of deaths of Covid this year with "otherwise would have actually died" group this year is small and can be defended with reasoning. By small I mean the effect is there, but not big enough to change policy response.

If you agree with my position, which you seem to, then why would it be unreasonable to defend a position you view as correct? If you disagree, then join the debate -- maybe I'm wrong and you can explain where and why -- instead of complaining about others debating on a debate forum.

Professor Ferguson and Professor Spiegelhalter are referring to the yearly mortality in their comments, as have I been. — Isaac

Ok, so we're talking about a year.

What evidence is there that the effect of overlap with people "who would otherwise die this year" is a big effect as opposed to a small effect?

I.e. big enough to support the idea that:

1. High overlap undermines certain arguments against social distancing measures because there should be little net excess in treatment requirement, focusing the main problem even more in the height of the spike of cases. Without overlap there is an argument that flattening the curve will not help because it pulls staff from other vulnerable cases in the long term so providing no net gain. In other words, with overlap we only need to re-assign resources (which everyone agrees is doable), without overlap we need to produce a net increase in resources (which many think is not doable, so why bother >> herd immunity bullshit). — Isaac

My argument with boethius is mainly about his ridiculous assertion that the overlap will definitely be small because there's no significant overlap in factors. This despite the fact the the only recorded factors affecting prognosis thus far are exactly the same as the factors affecting prognosis in other conditions, as the four articles I cited demonstrate. — Isaac

By small, I have been clear that the effect is there, the effect is measurable, but the effect is not so large as to essentially balance out deaths over the year, or come anywhere close to that.

Yes, people have problems that will likely kill them with time. We've been talking about a 1 year time frame. No where have you presented any evidence that most people dying of Covid would die within 1 year.

Him saying the overlap 'is not the point' of the graph has somehow become him saying that there is no substantial overlap (oh, sorry I forgot 'substantial' now means 'very small' - I will have to get the hang of this newspeak) — Isaac

Substantial for a statistician can easily mean "a small but statistically significant effect".

I have said the overlap is small in the context of your initial assertion that Covid deaths and "otherwise deaths" may completely balance out to have no, or hardly, and net increase in deaths. That's a big effect.

I've been arguing that the effect is small, because we have enough information to know it's killing people in large risk groups. And furthermore, I've been arguing that the overlap maybe not only small but not as big as long term injury replenishing those risk groups.

I agree that the complicating factors of system overloading and long-term lung damage make the figures difficult to say with certainty, but there is not any evidential support for the position that the overlap with those who would have died anyway will be statistically very small. As Professor Ferguson says, this is primarily a condition which causes death in those who are already very ill. — Isaac

You complain about me being a dishonest debater, and yet you don't mention that this has been one of my major points, that I said from the beginning.

And now you say "very small" is what you've been disagreeing with, to give room for you to have the small v very small part of the debate. I never used there term very small.

Yes, we know people that die of Covid are usually ill, maybe very ill, we've been discussing the overlap with people "who would otherwise die this year", not the overlap with "ill people".

If you want to discuss 10 year, 20 year time frame, then I would agreeing with you. But we've been talking about a year.

You've basically changed your position to my position, but you're so cranky about being wrong and citing evidence that supports my position, that you want to pretend my position was your position all along based on substituting meaning of words. But that ambiguity isn't there.

1) A spike in the death rate is only a snapshot at a particular moment. The 6000 extra people who died last week are not now available to form the pool of people who will die next week. This would be irrelevant if Covid-19 did not preferentially target those with underlying problems, but it does. — Isaac

This is your initial position in your disagreement with @I like sushi. I like sushi and I have been saying such an effect, is there, but is not big.

You've been arguing that Covid targets "the weakest" hearts and so on, to support your position that the effect is large, or at least likely large. So large as to change policy response or create narrative risks on the left of some sort.

You're right, and of course, the timescale matters. Thinking about overlap with deaths this year is a fairly arbitrary cut off point (why not the next two years or five). — Isaac

If you think we've been talking about some vague timeline and therefore, your position is correct given more time, you are wrong.

3) 2000 cases from respiratory conditions is not far off normal. It's the amount of cases with underlying health problems being pushed over the edge that is the real problem here. The key thing there being that we don't know how many of them would have died anyway, nor will we until the year's figures are out. — Isaac

Yes, the reasoning is based on the empirical data that the virus seems to simply double your chances of death this year, whatever your risk group; that this is the best predictor for most people. — boethius

However, if long term injury increases the risk-of-death factor for survivors of Covid, then you may end up with more deaths in absolute terms next year due to lung injuries or other long term Covid treatment complications. (likewise for every other demographic cohort) — boethius

We know with great precision how many of those people were going to die this year anyway, its about 300,000 (the death rate minus deaths from accidents). So until the death rate from Covid-19 exceeds 300,000 you can't possibly say that the victims were not going to die anyway, simply on the basis of the numbers, you additionally need data on the overlap - or you need to wait for deaths occurring over a longer timescale - say a year, or you need a plausible mechanism of fatality which does no coincide with underlying health conditions. — Isaac

My argument is a counter-argument to the idea that Covid is shaving off a population from these risk groups that can be in some sense said to "about to die anyways"; I've been using a year as a baseline time frame for the meaning of "about to die".

Covid doesn't kill enough people to have an obvious and noticeable statistical effect of this kind, such as non-respiratory disease going forward making up for, or nearly making up for, Covid deaths and arriving at some equilibrium. — boethius

We've been talking about a year.

Obviously, if you make "die anyways" to mean any length of time then the overlap is 100 percent as @Benkei mentions. Since that's obvious it's necessary to discuss some specific time frame.

We've been discussing the time frame of a year.

Why don't you just take it up with the experts, they both have blogs. I can't be bothered with this condescending "I'll teach you where you've gone wrong" crap. — Isaac

You just stated in your previous comment that you've been asking for the basic statistics all this time.

But yes, explaining a position requires explaining it. If you don't want to debate, probably a debate forum isn't a good place to be.

Prof Sir David Spiegelhalter {Professor of Mathematical Statistics at Cambridge), - "there will be a substantial overlap, Many people who die of Covid [the disease caused by coronavirus] would have died anyway within a short period," — Isaac

This statement depends on what he means by "substantive". Most people might think it to be "a lot", but in this context it is just some measurable effect.

That does not mean there will be no extra deaths - but, Sir David says, there will be "a substantial overlap".

"Many people who die of Covid [the disease caused by coronavirus] would have died anyway within a short period," he says.

Knowing exactly how many is impossible to tell at this stage. — BBC

This is the context. The statement here does not support the idea that this effect is large, just that we will see it.

If he thought it was the majority or "nearly all" he would have said so.

The basic math he uses is "Nearly 10% of people aged over 80 will die in the next year, Prof Sir David Spiegelhalter at the University of Cambridge points out, and the risk of them dying if infected with coronavirus is almost exactly the same".

I'm not sure he's trying to mislead lay people on purpose, as many will interpret that statement to mean your risk of dying this year is the same with or without Covid. Rather, coincidentally, risk of death of Covid happens to be similar to your yearly risk of death. The missing key element he or the reporter leave out is that based on this information the risks are commutative. So, with an unmitigated spread where all ~80 year olds get Covid, 10% of them die, but then another 10% of those remaining go onto die within the year; so 19% total deaths this year.

Now, if we include other risk factors, yes, we could expect less due to this; but unless the disease somehow targets those 10% that would normally die (10% of 80 year olds die and then significantly fewer die for a whole year), the the change is small.

Furthermore, it's simply irresponsible to not mention that maybe surviving Covid increases the risk profile in these groups, so you get more deaths due to long term lung damage instead of less deaths due to overlap.

In otherwords, this expert does not support your position but has made an ambiguous easily misunderstood statement about a lack of knowledge.

It seems the journalist paraphrased a longer rambling explanation of the details with "knowing exactly how many is impossible to tell at this stage". So, it's clearly not a prediction in any case.

Prof Neil Ferguson, the lead modeller at Imperial College London, has suggested it [the deaths of those who would have died anyway] "might be as much as half or two thirds of the deaths we see, because these are people at the end of their lives or who have underlying conditions. — Isaac

I read the article where this citation comes up. I'm going to give the benefit of the doubt for Prof Ferguson that he was temporarily hallucinating about statistics at the time, as he himself was recovering from Covid. And again, it's a "who knows" kind of statement.

But if you read his modelling paper Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand then you see an unmitigated death toll from Covid of 2.2 million in the United States from May to July.

Total US deaths in 2018 was 2.8 million. There's simply no way Prof Ferguson is saying you could have 2.2 million deaths in 3 months and then going less half to a third of otherwise expected deaths happening for the rest of the year.

Likewise, if you read the paper of the expert you're citing, you'll see a massive overburdening of the health care system in unmitigated spread as well as the problem of it happening again and again and again each time a lockdown is lifted.

When he talks about "half to two thirds would have died anyways" he seems to be talking about a situation where social distancing is done really, really well and the disease is limited to being like a hospital disease that kills terminally ill people since they cluster around hospitals.

And again, without mentioning that the effects of Covid could replenish these risk groups that would likely die within the year, it is simply a mistake on the part of this expert. Experts can still make these kinds of errors in interviews, which is why citations of published papers are usually preferred supporting evidence.

I don't know if he's specifically revised this statement, but 2 days ago he was interviewed and said:

“What we really need is the ability to put something in their place. If we want to reopen schools, let people get back to work, then we need to keep transmission down in another manner.

“And I should say, it’s not going to be going back to normal. We will have to maintain some level of social distancing – a significant level of social distancing – probably indefinitely until we have a vaccine available.”

He said that despite the “billions of pounds per day” cost to the economy – by putting in place infrastructure to tackle the virus – it was a “small price to pay” to tackle the outbreak of the virus.

Pressed on whether the government was moving towards an exit strategy, Prof Ferguson went on: “I’m not completely sure. I would like to see action accelerated. I don’t have a deep insight into what’s going on in government but decisions certainly need to be accelerated and real progress made. — independent reporting a BBC 4 interview

This doesn't seem the position of someone who thinks up to two thirds of people dying of Covid are those that would have died within the year.

No, you gave a small number of minor factors without any citations to back them up and nothing to counter the cited evidence I provided of the major factors which do overlap. — Isaac

You didn't disagree with future contingencies affecting heart disease outcome; therefore, I need not cite it as you seem to agree and I used common knowledge examples.

So, if you don't point and say "I'd need proof to assume stressful events that might happen in the future could impact heart outcomes" then I can't know you don't also accept this common wisdom.

Instead, you cited evidence that supports my view, that risk factors are very large groups such as heart disease, obesity, etc. not evidence that it is only the "severe risk subgroups doctors say will likely die anyways soon" nor any evidence that such groups exist despite doctors not knowing about them.

If you provide citations that support my point, why do I need to do anything?

And again, no where have I said factors don't overlap, the point at issue is whether these overlapping factors are coming from large groups that have small factors (many with a small chance of death in both sets; i.e. risk of death from both selectors) or small groups with big factors (few with a large chance of death from each).

Yes, and you've yet to demonstrate, with evidence, that the selection is small (relative to the group {at most risk}, nor that there is 'some other' selector rather than exactly the same one. — Isaac

This is your hypothesis, and there's no evidence for it, therefore "there is no reason to assume it", therefore it's not a reasonable risk to take.

Furthermore, it wold overturn the prevailing view in medicine that most deaths in a given year are fundamentally unpredictable (there are very, very small groups in "90% risk of death this year" but there are very, very large groups with "10%, 5%, 1%, 0.5%, 0.1%" and and most deaths come from these groups in a given year because although probabilities are small membership is large).

But if you are not aware just how sensitive "selector collisions" are to the summed probabilities of selectors being small, then it's difficult to intuit these large number statistical theorems (which are clearly at play with the evolution of Covid, otherwise nearly every country wouldn't be experiencing the same simultaneously, just managing better or worse, as there would be dominating small number variations that result in very different outcomes due to pure chance).

So I will explain these theorems and why chances of the overlap you are talking about become so small as risk groups become large and chances of death from both Covid and underlying conditions are small that the broad facts about the pandemic support my position without the need for laborious analysis taking into consideration all sorts of subtleties.

There is a clear dominating driver of events which is Covid, once containment fails, kills enough people in a short period of time as to overwhelm health systems; society (relatively rich societies anyway) simply can't function without a health system and so, even the most "initially downplayed to the max" governments take social distancing action and try to ramp up resources to deal with the situation (with the exception of Brasil, so a convenient, although tragic, control case for this analysis).

There is. Loans, postponing leave, postponing retirement, postponing investment plans. There's all sorts of ways of borrowing from the future. — Isaac

Not physical resources. A loan doesn't help a doctor today treat a Covid patient, only real material and human resources (which cannot come from the future).

Right. So you're wrong when you say that governments can deal with these problems through rationing then aren't you? — Isaac

Government have already implemented rationing successfully, even some chains voluntarily implemented rationing.

For instance, in Japan there is a limit to masks you can buy and a large fine for buying more (because rationing obviously works to prevent perverse distribution of resources in an emergency).

Yes, that's what I've been repeatedly asking you to do. — Isaac

You have not been repeatedly asking me to do this, and I have not volunteered as calculations don't actually solve the disagreement.

Your claim is that within these risk groups such as heart disease, there is a hidden risk group of "weakest heart" and these are those dying of Covid and are subsequently culled from the heard and don't die / burden the health system later.

Demonstrating how my point about these groups being large makes the overlap small using calculations doesn't solve the above difference. And, it seemed you accepted that my point about large groups held, so there was not need for me to demonstrate it.

I gave lot's of reasons why there is no such subgroup of "weakest heart" as lot's of factors affecting real death from heart disease are in the future and therefore Covid cannot select for.

My whole position is based on a well known statistical fact that as selection from a group becomes a small, the chance of colliding with some other small selector is small. If we are both picking 1 out of a hundred hats at random, it's unlikely we'll pick the same hat. Now, it's not random given the whole population, there are risk groups, but if these risk groups are large (such as heart disease or obesity) then collision remain small (1 out of a thousand instead of 1 out of million). The set we're selecting from needs to become very small or then the number of selection very big of one or both selectors for there to be collisions.

However, for your benefit I will bring out all the theorems and things we "certainly" (highly, highly likely) know about the pandemic so far, to demonstrate why overlap will be very small of all kinds; both confounding causes of death and overlap with "would die soon anyways".

In other words, with overlap we only need to re-assign resources (which everyone agrees is doable), without overlap we need to produce a net increase in resources (which many think is not doable, so why bother >> herd immunity bullshit). — Isaac

For instance, if you want to debate your arguments even assuming your premise, the above simply doesn't hold. Even if there's overlap, I've been mostly talking about overlap of a temporal nature (people dying now that would die in the near future, such as within a year), and there's no way to re-assign resources from the future to the present. Furthermore, lot's of health-care resources simply don't apply to respiratory infection, and therefore can't be reassigned; therefore, for both reasons, even if your overlap theory is true, there needs to be "net increase in resources".

True, it is not doable to scale in parallel to a unmitigated pandemic; yes, the fascists say "why bother, let's keep the economy humming and the dividends flowing". But this is a false dichotomy. We can lower infection rate by social distancing to something that is below health care capacity (such as many countries have demonstrated) or then at least not so far beyond capacity to reach a total health system collapse (every country late to the game is doing). We can at the same time scale resources as best we can.

5. We will have to come out of lockdown soon (partially) and continued promotion of the idea that Covid-19 is some random reaper stalking the land takes resources away from those who really need them as the hysterical-selfish (by far the largest population group) panic-buy themselves their ppe/food parcel combo (Disney-themed, Bluetooth-enabled version, only £9.99 on Amazon), — Isaac

Again, nor I, nor anyone else has here has claimed it's a "random reaper", just "random within risk groups that are large enough and existing risks low enough that Covid deaths do not coincide with 'weakest heart' deaths type hypothesis". Furthermore, people aren't panic buying due to such a fear in any case, most young people and even a lot of elderly people don't fear getting the virus, but people are panic buying because they think other people are panic buying or then not even panic but in the hopes of price gauging, because of the lock-downs not the virus itself.

while doctors make do with paper towels and some sellotape. — Isaac

Both the panic buying in itself and doctors going without due to panic and hoarding are false dichotomies. Government can easily solve such a problem through rationing, and many governments have. In the US this would be socialism and "a republican" administration doesn't want to set precedents that socialism can help on some issues (except giving corporations as well as 2 week old companies created by sycophants money).

No one is debating these facts, but once numbers get large the ambiguities get small, as a large amount of people dying of pneumonia in a region during respiratory epidemic is very likely.

The graph I just posted (if true or close to true) demonstrates the basic problem, as it would mean (if everyone got the disease) about double total deaths in the year (there is less infants that die, but there's some excess in the 60s range) happening in a short period of time would be a total disaster.

Furthermore, many people that recover still needed hospitalization and care not just at an alarming rate, but also for a long duration of time (2-3 weeks).

There's simply no way statistical coincidence with other causes of death explain the phenomena of overloaded health systems, nor any reason to expect we'd have some large effect of abatement of those other causes of death after the initial overload.

If you really want me to go into the calculations that explain my position, I can do so.

I'd rather hope to be discussing things with people intelligent enough to know the difference between a fact which is unhelpful and one which is wrong. — Isaac

If you think this applies to me in this discussion, please argue the point. I said maybe you are affected by attempting to overcompensate your bias; maybe not. Either way, you are still wrong about believing the real overlap that really does exist could be big enough to result in effects you think are possible.

The source I previously mentioned.

And you elected Trump, so what's the problem? — ssu

The problem is that the right has all sorts of dogmas that excuse Trump's corruption, even make it a good thing.

For instance, Randians may see Trump's corruption as simply "following self interest" which is a good thing in their system. Of course, the immediately use the left's standards of corruption and the idea "greed is bad" when criticizing the left or government in general. So Hillary is a corrupt greedy and basically evil politician, like all bureaucrats and politicians, but Trump is a savvy business person who knows how to make the system work for him, and if he does it corruptly and gets ahead that's just winning, and of course he is bringing that savviness into his new life as a politician to enrich himself. This world view basically reduces to the dogma that business people being naturally greedy creates good but bureaucratic being naturally greedy not only creates bad but makes them evil; without any attempt to resolve these two groups are coming from the same homogeneously greedy humanity that is postulated nor resolve what happens when "good greedy business people" get involved in politics or bureaucratic processes.

Likewise, the evangelical right have created a dogma that Trump is literally or metaphorically the reincarnation of an old testament king, which God worked through to do the good. Trump is behaving as they'd expect an old testament king to behave (an old testament king does what they will, takes the women they want, the money they want, isn't accountable to anything) so therefore this is evidence of the theory working.

However, the left doesn't have similar dogmas to justify corruption on the left. There is a legitimate self criticism in the desire to reduce and remove corruption as well as a legitimate response to criticism from the right. The right can therefore use the left's standards of right and wrong against the left, because the left tends to believe those are real legitimate standards (not just politically expedient value signalling that can be resolved by the talking points of the day regardless of even a surface scrutiny of those talking points nor if they contradict yesterday's talking points).

Pointing out the right is hypocritical in using "greed is bad" with regard to left politicians and not right politicians, and hypocritical of assuming Trump can be the reincarnation of an old testament king (to do good mysteriously through doing bad, basically) but not Obama (who's literally the anti-Christ) doesn't help win an election.

What helps winning an election is nominating someone who can be easily defended against charges of corruption (due to a long track record easily explained by a lack of corruption) and who doesn't have super cringe videos of interaction with children (the argument Biden is simply not self-aware doesn't help, nor the argument that it's actually normal stranger-child interaction that "happens to have bad optics for a politician" as people can verify in their own lives that no one interacts with children in these ways; the constant touching and prodding of children even fathers don't do, much less grandfathers much less strangers; normal touching is holding hands, a pat on the head or some sort of fun game that involves touching like tag; the idea that these videos aren't a major problem for the Democrats is crazy; they'll be circulated and recirculated and the emotional impact will be enormous; the idea he's an out-of-touch, not self-aware, hyper-doting pan-grandpa to a creepily cringe level isn't a good argument, which is basically what apologists for Biden about these videos argue; nor the argument he's just trying to "get a good photo" stand-up to 1 second of thought, as he's not the photographer nor actually looking at what he's adjusting, and no one normal thinks "oh, this child may ruin the picture, I need to be continuously on this problem" to begin with). What's the apology? "Oh, no, no, no, Biden's not a pedophile, he just loves children so much!" Yeah, sure, maybe, but choosing a candidate where scrutiny starts at such a statement is sub-optimum strategy.

True, on some level "what about Trump" can justify supporting Biden in a pick-your-poison analysis, even if the worst accusations about Biden are true, but whataboutism doesn't win new converts, it just helps keep existing converts in the cult, and Trump win's handily if it's about who has the bigger cult following, and so the progressive left concludes that nominating Biden demonstrates the DNC rather lose to Trump than win with Bernie (or any candidate remotely credible, due to legitimately being not corrupt); Trump's given all the donor class huge tax breaks, he's pretty good if things are looked at objectively from the donor class point of view. And progressive leftists have been saying this for a while; Biden is just an ultimate confirmation for this point of view.

Am I too pessimistic in thinking that I should throw out Mills "On Liberty", and buy a new copy of Hobbes "Leviathan"? — graham hackett

There are two meanings of liberty.

We can mean liberty in the sense of individual freedom of action as you have been considering.

However, there is another meaning of liberty which is "equal political participant"; arguably the original meaning. The concepts naturally go together, as an equal political participant and first class citizen naturally has lots of liberty of action and expression, as much as anyone else (otherwise, by definition, there would be some higher political class with such liberties that actually manages things).

For instance, serfs did not have liberties of a political sense of equally participating in politics compared to aristocrats. They also didn't have liberties of the physical actions sense in terms of where they could go, what they could sell, etc. So although serfs "becoming equal" really only matters in terms of political participation (as there will always be rules) the visceral experience of "not having liberty" is being subject to those rules that apply to second class citizens and not first class citizens and so the two ideas are easily confused one for the other.

Getting rid of rules beyond the ideal of equal participation has no expected bearing on both kinds of liberties. For instance, if getting rid of a rule creates endemic corruption and subjugation by a new aristocratic class that has all the meaningful levers of power, then one has "given up essential Liberty, to purchase a little temporary Grift and deserve neither Liberty nor Grift."

So a situation like a pandemic which obviously requires changes to our general liberty of action, need not change political liberties of being a equal political participant.

Countries that can be said to have genuine political liberty of this sense (Switzerland, Scandinavia, New Zealand and a few others) the pandemic creates no "authoritarian anxiety", as there was none to begin with. Of course, there's lot's of discussion of what's best to do, as with any subject of government policy, but there is no fear that the pandemic will remove this liberty of political expression and participation and that the government can somehow "take advantage" (they will simply be voted out if they don't manage in a way generally approved).

However, countries that don't have much political liberty in this sense (where people objectively aren't equal political participants; either due to electoral design, endemic corruption and disenfranchisement, or simply being a straight up totalitarian state not even pretending to be democratic) then the pandemic amplifies authoritarian anxiety, as well as the real risk of even more of it, as the people don't have a sense of "being in control of a transparent and accountable government" before the crisis, and so the emergency powers and chaos of the pandemic simply bring this fact to the fore with very real reasons things may go even further away from genuine political participation (that the first class citizens that have the real liberties in both senses of the word, don't let a crisis "go to waste").

In other words, those who have liberty shall be given more and those that do not have liberty shall lose even that liberty which they had.

Note the repeat of hypertension, CVD, diabetes... — Isaac

This data supports my point not yours. Hypertension, diabetes, etc. are very large risk groups from which my analysis follows.

If risk groups are large, then the "people who we would expect to die this year from the existing pattern" are unlikely to intersect "people who actually die from Covid". There is some intersection, but it is small; there is also a small change in simply the absolute numbers that makeup the group due to people dying from Covid. But both these small effects would only be relevant with a large portion of people actually getting infected, and in such a scenario it is a very real the possibility that long term lung damage or other treatment complications replenishes the risk groups (this maybe a small amount too, but the effects under consideration are also small).

Furthermore, you've simply ignored the other reasons we shouldn't expect Covid deaths to be displacing near-future-otherwise-deaths even if there was unknown "weakest heart" kind of groups within these groups, such as diabetes, hypertension etc. such as the simple fact we're early in the outbreak.

Yes, people dying tend to have underlying conditions. But no, people dying are not "going to die soon anyway" in any meaningful sense. Covid does not select for "going to die anyway" nor "the weakest part of the spectrum within these groups" (and such a spectrum does not exist in well ordered nor static sense; a lot of chance is at work).

The reason I argue this point, is because it is a widespread misunderstanding that Covid is "culling the weakest". It lines up with certain political ideologies that want laissez faire survival of the fittest, which I understand you don't empathize with; nevertheless, not emphathizing with a political bias does not automatically protect oneself from misgivings propagated or that happen to line up with such a bias (indeed sometimes we can be so concerned of our own biases that we jump on inconvenient impressions to convince ourselves we are managing our biases).

So, I have no problem continuing to argue the point.

To support your position you have to demonstrate that the vast majority of factors defining the most vulnerable people in the group suffering from heart disease, lung conditions, cancer etc are not the same as the factors defining the most vulnerable people in the group of Covid-19 sufferers. — Isaac

First, I've already explained why those factors can't be the same as some of those factors are in the future. So I guess deal with those first.

Second, I've already explained why it's highly unlikely for those factors to be the same even in the present; for instance, inoculant load is a factor highly suspect to be a big factor in outcomes for Covid but cannot be a factor in any other underlying health condition as it's Covid specific.

The more the factors don't overlap, the smaller the effect of "Covid killing those that would die soon anyway" becomes; though, to be clear, no where have I stated it disappears, it just becomes small.

Smoking, obesity, even age, are risk groups where people can still be expected to live decades, but die from Covid despite such odds.

But it does kill people now who are likely to die soon that aren't likely to die now otherwise, right? — fdrake

Agreed, but this is a small group of both the known terminally ill risk groups and postulated hidden terminally ill groups, and within this group not all have gotten infected at this point (and the initial conversation was about UK numbers essentially next week or a few weeks from now). So the effect is small because these groups are small.

Covid definitely is a disease likely to kill the terminally ill, but it also kills people in very large risk groups that have an average life expectancy far beyond a year. Covid kills a small percentage of these people, again they need to get the disease first also, and so the effect is small on reducing future deaths because not many die and they continue to have the same risks as before.

We know that Covid is not killing only people that doctors expect to die shortly anyways, and it's implausible that there is some hidden extreme-risk sub-group within larger risk groups that Covid happens to kill (and implausible such a sub-group even exists that explains all, or even most, deaths in those risk groups).

since the dead people won't be in that group any more, and certainly can't die from other causes if they're already dead from COVID — fdrake

This is my position, which I have been very clear about since the beginning.

I also gave Isaac the same benefit of the doubt, as I wasn't sure if he meant "significant" in the sense of "big" or in the sense of some measurable statistically significant effect (which can be very small, but still measurable). Why I have stressed I'm talking about some short term observation, such as within a year.

But if you read Isaac's recent comments, he has clarified that he means Covid kills some hidden subgroup of for instance the "weakest heart" within the relatively large "heart disease" risk group. An effect larger than simply reducing these groups by whoever dies of Covid, but that they otherwise continue to have the same risk profile going forward.

I believe we agree that's not the case; that there is some effect of culling the terminally ill (whether known or unknown) but it's not a large effect.

What are these factors then (presumably ones which don't also overlap with factors making death from Covid-19 more likely)? — Isaac

I say "can be easily be other factors", I mention otherwise benign genetic differences (that benign differences can have a significant outcome difference given some new threat, is exactly why we have evolved to have such differences; epidemic is the classic case for why evolution goes this way), but there's also initial inoculation viral load that is highly expected to have a big effect on outcome, and of course timeliness and quality of care, but even with similar care there is variation in response to treatment.

Really? In what way? Presumably proximity to medical services is the key variable in time and place (those more remote will have more difficulty). How is that different with Covid-19? — Isaac

I am talking about the variable of proximity to medical services when one has a heart attack (or ability to get service before said heart attack). This is in the future and totally independent of Covid. These future contingent events that have an effect on heart disease outcome mean that the "weakest hearts" is an oversimplification of who exactly dies of heart disease in a given time; "weakest heart" maybe a subgroup, but there's also a large group that then get's filtered (in the future relative Covid) by proximity to medical care.

If you want to talk about the subgroup of people far from medical care as a constant risk factor; it's not symmetric as a heart attack is much more acute.

Again, how do these categories differ from those which relate to vulnerability to Covid-19 fatality? Stress, for example, suppresses immune response. — Isaac

Again, some future stressful event is a filtering mechanism that is independent of getting Covid today. The person that has heart disease but not the "weakest heart" today, may live to encounter some future event, such as acute stress, that puts them at acute risk of heart attack.

These future events that filter for who actually dies in the risk category is simply the strongest example of why "Covid kills the weakest in these categories" is not sound reasoning.

I mean, COVID is more likely to kill people who are more likely to die anyway. This complicates whatever attribution of death to COVID you do. — fdrake

I don't disagree here. Where I disagree is this effect will be big in terms of reducing deaths from these risk factors going forward.

The original claim I have contention with, is that respiratory deaths may go so low after Covid as to balance Covid deaths. So a incredibly large effect.

There is definitely some overlap, but my contention is it is small; small in the sense that it could be completely ignored in calculating likely deaths in these risk groups (though there can be other causes of big changes, such as the lock-downs). I.e. if you calculate the people who die of Covid that are at risk of heart disease in likely scenario of your choosing, and then calculate the people who will die of heart disease, you can essentially ignore the the fact some people died of Covid in this category other than there simply being slightly less people. That there is not a big culling effect.

It's not big now because most people haven't gotten the disease yet.

It may never be a big effect because likely Covid simply doesn't kill those who are about to die in the short term. I've defined short term as a year and explained all the reasons why we wouldn't expect there to be an overlap with "people who would otherwise die this year" and there's no reason to assume such a category even exists now in a predictive sense.

It may be an effect that that is not only not big but is dominated by something that goes in the other direction, such as a large amount of long term lung injury that replenishes the at risk categories, or even increases them.

requires an explicit model of how COVID interacts with the comorbidities, and can't be immediately read off the risk of death of those people who have those characteristics (comorbidity + age) who have confirmed cases and died in hospital (that group selects for comorbidity severity already!) — fdrake

Yes, this is my point. If the comorbidity groups are large, we can't simply assume Covid is killing off some unknown sub-group who are very likely to die in some agreed short term anyway; and once such comorbidity groups are large it is implausible such a subgroup even exists that explains many, much less all, deaths in that group: Some people in the risk group die due to simply being in the risk group, without some hidden mechanism that explains why they in particular died, and even if there are hidden subgroups there's no reason to expect they overlap with the subgroup of people more likely to die of Covid -- some otherwise benign genetic difference may dominate here, exactly why species store up diversity in case some otherwise benign difference is no longer benign given new conditions.

If you take out the entire risk group, then it's a different story, but Covid is not remotely lethal enough; therefore, there's no reason to expect Covid deaths intersect with "would die anyways from underlying conditions in the short term". In a refutable form; Covid discovering for us there are such hidden "weakest, about to die" subgroups in otherwise large risk-groups with little success so far in a finer grained differentiation, would be the greatest medical discovery of all time.

The exact same factor. — Isaac

Yes, there is the same factor.

But having a factor of risk does not mean you will die within a short period of time, it is just a factor.

You are making a completely unfounded addition to the risk group observation that Covid somehow selects to the weakest members of those risk groups -- essentially the terminally ill but we don't know it. There's no reason to assume such a thing. If Covid was a disease of the terminally ill, such as a hospital disease that kills only those with essentially failed immune systems, we would know it by now. There's no reason to assume that there's some hidden group of "true terminally ill people" that make up "actual expected deaths within a year" that we don't know about but will discover because Covid kills them and then they do not live to be killed of their other risk factor. Furthermore, I don't think any doctor would agree such a "hidden terminally ill" group exists, but would say there's a large element of chance within these large risk groups, such as heart disease or smoking or cancer.

Therefore, the only way to have a culling effect is if a disease killed a large proportion of such people, otherwise, with a small amount of killing that Covid does, small in the sense of seeing such an overlap between "Covid deaths" and "people who are about to die anyways", the same "elements of chance" continue to operate in determining who has a turn for the worse, who responds well to treatment and who doesn't, who encounters deteriorating life conditions, develops an addictions, has an accident or some other complicating factor, with respect to Covid as well as whatever other conditions such people continue to have.

In other words, there is not some well ordered spectrum from "good to worse" of heart disease, or lung disease, and that people simply move progressively towards "worse" and then fall off the edge and die. There is a large element of intrinsic randomness and re-ordering due to contingent events, which is why these groups are large and not already separated in a finely grained way with excellent predictors of who in particular is going to die in a given time frame. It's an extremely small group of people who doctors are "certain" will die in any relatively short time frame, such as a year.

Those most likely to die in the "heart disease" group are those with the weakest hearts (for various reasons), those are the same people who, within that group, are more likely to die from Covid-19. It is the inability of the heart to support recovery which causes the fatality, not some dice-rolling random factor. The exact same factor. — Isaac

No, this simply isn't true. You are taking a truism too far.

Also, note that you also have to deal with the fact that right now it's a small group of people who have been infected, and so the effect you describe would not be noticeable much anyways simply because the vast majority of "those with the weakest heart" have not yet gotten the disease. And even if you deal with that, you still have to deal with the fact that long term lung injury or treatment complications may simply replenish this supply of "those with the weakest heart".

However, even the middle part of assuming Covid kills "those with the weakest heart" within the heart disease group is not a sound argument. Other factors can easily dominate in selecting for death within the heart disease group.

Furthermore, a big determining factor for surviving a heart attack is time and place, and this is a completely independent variable to Covid; likewise, people may improve or deteriorate their lifestyle moving from this "weakest" category to "ok" or vice versa, or a really stressful life event has an acute impact on heart disease likelihood; failing to seek timely treatment etc.

Point being, even simply considering these future events that determine heart disease deaths, in other words make your "weakest heart" category not static but a dynamically changing group, significantly enlarges the category of "weakest heart"; i.e. Covid may kill someone who is in the "weakest" category now, but would otherwise have gone on to make life changing decisions and moved out of the weakest category, and so would not have appeared as a heart disease death in the short term.

So, you also have to deal with the dynamic nature of your "weakest" category, in addition to deal with the fact "weakest" is not a given and other factors may dominate who dies and who doesn't of Covid with a risk category. After solving these, there is the fact not enough people have gotten Covid for such an effect to be large, and the fact that if that does happen, Covid does kill off the "weakest", that those that survive may now have long term comorbidity effects due to Covid, thus replenishing this "weakest" group.

The question is: who are those people who are most likely to die of COVID? That's people who are elderly and have comorbidities. — fdrake

I'm not disagreeing that people with comobidities are more likely to die from Covid.

My argument is a counter-argument to the idea that Covid is shaving off a population from these risk groups that can be in some sense said to "about to die anyways"; I've been using a year as a baseline time frame for the meaning of "about to die".

Covid doesn't kill enough people to have an obvious and noticeable statistical effect of this kind, such as non-respiratory disease going forward making up for, or nearly making up for, Covid deaths and arriving at some equilibrium.

If Covid killed everyone who smoked, everyone with heart disease, and every cancer patient and all the old people, then it would have such a very noticeable effect, but the disease doesn't behave in this way and we cannot assume that "particularly unhealthy anyways" dominates in determining who in these risk groups actually dies from Covid; other factors could dominate the death selection process within these groups.

Now, maybe this effect of less respiratory disease deaths does actually happen, since the lockdowns we can easily expect to have a significant change to peoples relationship to pollution, to stress, and to influenza and other infections -- that we can assume are also depressed by the lockdowns. But these effects are due to the lockdowns, not due to Covid culling the "otherwise would have died anyway" group in a big way.

This is trivial compared to the disproportionate risk having heart disease, lung conditions or undergoing treatment for cancer has on your risk from dying of Covid-19. — Isaac

But these are the large groups I've been talking about.

Yes, heart disease is a big predictor of Covid outcome, but it's a large group and there's simply no reason to assume deaths from Covid overlap with some unknown "particularly unhealthy" sub-group of heart disease.

There are definitely known subgroups such as those already experience heart failure, or have had a heart transplant which is being rejected, which I would assume Covid is an even bigger risk. But there is no evidence that Covid is only affecting this known extreme risk subgroup.

Your argument is that we can assume Covid selects for the "particularly unhealthy" within these groups, but there is no reason to assume that is the case. We cannot expect heart attacks to lesson substantially after Covid because the "particular bad cases of heart disease" were culled from the heart disease risk group.

Now, maybe heart attacks do decrease due to systemic effects such as people de-stressing in lockdown, but this is totally different than a "cull effect".

And, even if there was a substantial cull effect, many survivors may have long term injury that simply replenishes the "particularly unhealthy" end of their risk group.

It isn't necessary that those within a risk group that at "least healthy" will die, it's just more likely. If you found any factor or variable which contributed to risk, and it wasn't aliased with * the risk group already, those in the sub group of that risk group that have the extra risk are more likely to die. — fdrake

This is simply not true, as I've explained above.

First of all, at this stage with a small percentage of the population that has gotten Covid, it just doesn't matter because 90-99% of these "least healthy" are still out there and will continue to die due to whatever they are at risk of.

Second of all, even of the people that die within a risk group, the factors determining death compared to one's peers could be otherwise benign. A particularly unhealthy smoker may survive Covid due to some completely benign genetic difference, such as exact shape of proteins on cells etc.

An example, it seems blood type O is particularly resistant to Covid, but blood type O does not provide a similar resistance to smoking. So O blood type's who smoke and survive Covid due to this genetic advantage, there's no reason to assume that they were a "particularly healthy" smoker.

It's an over simplistic assumption to postulate Covid deaths is selecting for "least healthy" within a risk group. It seems intuitively correct, but is not correct.

As I've also mentioned previously, the survivors of Covid, but with long term lung damage or treatment complications, may then replenish the "least healthy" category even if there was such a selector for "least healthy" to begin with.

No they don't, because if everyone is equally likely to be infected then the liklihood of infection can be removed from the equation. — Isaac

No it can't, if you're trying to support the idea that Covid kills the "particularly unhealthy". Lot's of "particularly unhealthy" simply don't get the disease, so there will remain lot's of these "particularly unhealthy" around since they didn't get infected.

Unless you're suggesting that there's some gene specific to the defense against Covid-19, then the only genetic component which might be relevant is one which affects the immune system in general. Such as defect would put you in the cohort from which the 300,000 yearly deaths are drawn. — Isaac

I'm saying there's no reason to assume the variation of death and survival within a risk group is due to being "particularly unhealthy" within that group. It could be some other mechanism such as otherwise benign genetic differences, or then simply random variation such as where exactly the virus begins replicating in the body, that then dominates chances of death within a risk group.

Yes we do. It will (disproportionately) be the least healthy. Same as those most likely to die from Covid-19. — Isaac

As I've explained above, this is not a given assumption.

We cannot assume those that die from Covid are "least healthy" within their risk group.

Then where is the random mechanism? — Isaac

The random mechanism is that we don't know who within a risk group is actually going to die this year, so taking people out of the group by another mechanism, such as Covid, doesn't change significantly the expected pattern of death from the existing risk. The randomness is due to a lack of knowledge at this stage; but there's no reason to expect Covid targets "the particularly unhealthy members of a risk group".

No it isn't. Even within a risk group, the least healthy members of that risk group are more likely to die than the most healthy. — Isaac

This is not true.

For this to be true, the "less healthy members" within a risk group need to somehow be far more likely to get infected to begin with. That is certainly not the case so far.

Furthermore, it would need to be the case that being particularly at risk of Covid within a risk group, means being particularly at risk of whatever makes up your risk group. It could be random genetic differences that make a person in a risk group, such as smoking, particularly at risk of Covid.

In other words, one could be a on the "healthy side" of a risk group, but particularly vulnerable to Covid due to some genetic difference that has no bearing on one's underlying condition.

Covid-19 kills people either by the lungs filling with fluid as a result of a failure of the immune system (sometimes from comorbid bacterial infection) or by exacerbating the effects of other conditions, particularly heart disease. Every single one of those mechanisms relies on an underlying health problem. — Isaac

Nothing I have said contradicts this.