THE MATHEMATICAL-LOGICAL METHODS PROVING THE OFFICIAL GROUP OF 2020-COVID-19 DEATHS WAS A MULTIPLE OVERESTIMATION OF TRUE COVID-19 DEATHS, THE U.S.

EXTENDED ABSTRACT plus SUPPLEMENT

BACKGROUND: The objective information in the CDC tables named 'Deaths involving Coronavirus' is: "xxx thousands of people died, probably infected with Covid-19". But how many of these people would for sure still live if not Covid-19? The main aim of this paper is to present a method that makes it possible to reveal the real number of lethal Covid-19 victims. METHODS: The ideas for solutions are original, mathematical – logical; there were used constructed equations; a few riddles had to be solved. Calculations are in some places somewhat simplified, to chase calculations. There were used the CDC, NSC, SSA and other agencies/institutions' databases. Doubts are resolved to favor overestimation, not underestimation of the analysis results. The number of chronic conditions is considered a reliable indicator (on average) of a pre-infectious worsened health state compared with the average for a specific age. FINDINGS: Including into the 2020- official "Deaths involving coronavirus" (DIC) group was based on irrational mechanical assumptions. The structure of this group in terms of age was nearly indistinguishable from natural death patterns (adjusted by eliminating infant mortality, injuries, and by illness rates -the main reason of falling illness rates among aged 50+ hardly influences those already in a terminal state, proven in the essay -p.7+), which had to give an average age of 76.6 - 77.1. Only up to 10% of those officially reported as Covid-19 victims in the U.S. in 2020 could have died from Covid-19 complicity; the rest would have died at the same or a nearly identical (plus some weeks) time regardless, because they were in a terminal state before getting infected (or positively tested) and their death resulted from the normal age-structure of deaths in the United States, and from causes already existing before Covid-19, creating the expected average age of natural death actual in the given year. It is also very probable that some people in the official DIC group had their death-date accelerated (forced to die in 2020, while otherwise would die much later) by wrong/harmful procedures and so the total number of those whose death was visibly accelerated could have been meaningfully bigger than 30K (of the DIC group), but they can be also in part already included in exchange for the Covid-19 deaths. With no additional avoided risks ('m'), the average adjusted total (at-birth) life expectancy of genuine Covid-19 victims should have been 77 or less (p.21); their otherwise residual life expectancy (LEa) should have been 10 or less (if an average age was ⩾66), not about 12.5 years (/no increase in the number of conditions visible), as would be assumed under an irrational mechanical approach that all decedents of the 2020-DIC group were not those who died in 2020 naturally due to "aging". The equation 'ADcs + LEa1 = timely-LEWIIfmS' (p.20+) must be fulfilled for every external burdening factor provoking premature quick deaths (/e.g. Covid-19) and its real victims must be (on average) weaker simultaneously due to their higher age and due to their worsened, against an age-average, health-state (with a strongly increased average number of chronic conditions for low mortalities) -otherwise the equilibrium never exists ...unless there is a 100%-mortality. And, if a number of chronic conditions was not a perfect reflection of a worsened health-state against the age average (or of risks) it would not mean much for an average age of 73, becauseas the result has a significant safety margin (although there would not be required an initially supposed number, but calculated, using SD of LEa, lower ~14.5  -'Corrections'-f); for an average age of 67 it means yet much less, as ISD, not a number of conditions, is the main result-limiter there. ...Independently, even just the average number of chronic conditions protects the main result, which number must have been strongly increased among living individuals (but not already in a terminal state) and then really killed by the virus against the number in a comparative group of just alive ones with exactly the same age-structure (!), because Covid-19 out of infected ones at every specific age kills very few (of older ones) or very very few (of young ones) of them -usually weaker/weakest ones; the age-adjustment eliminates any meaning of older ones dying more often than younger ones from Covid-19. The CDC knew there had to be a very clear correlation between 'Covid-19 Death Risk Ratio' and a number of conditions, but in the DIC group none increase in the average number was visible (<4 conditions on death certificates) even with limiting the list of conditions in the comparative group to the conservative CCW list (/the comparison only of the prevalence of major heart diseases shows it is not increased in the DIC group -more in 'Additional Notes' -b). If older persons have a high number of CCW-conditions then this pure number matters very much to residual life expectancy (LEa); the marginal decline in LEa increases with an additional chronic condition, when a number of conditions is low or moderate, but this decline starts with low values -first 2 conditions of 2008-CCW ones (and so approximately first 3 of current CCW ones) sum up to about 3 times less negative effect on life expectancy than the next 2 conditions (3+4) of 2008-CCW do in 75-year old ones (= approximately 4+5+6 of current CCW ones). /E.g. normally, still alive ones at age 67 and 75 with a low number (0 - 3) of current CCW-conditions should otherwise live, on average, for the next approximately 22 and 16 years respectively, but with a very high number (15 - 30) of conditions they would otherwise live, on average, for <5 years only./ [18 +CCWdata]. Natural deaths with chronic conditions (ones 100%-mortality) are, in contrast to quick premature deaths due to the infection, realization of risks originated in ones' past, with almost none predictability of a specific year of falling into one's terminal state (read also 'Additional Notes'-c, which are below). INTERPRETATION: The official number of "Covid-19 victims" is a very untrue number and means, in a vast majority, 'the double counting' of those who would die whatsoever at the same (or very close to identical) time without Covid-19, because they were in a terminal state already before. The 2020-DIC group's construction is based on irrational mechanical assumptions, resulting in natural 2020-deaths being not removed from it at all.

SHORTENED SUPPLEMENT: THE ALTERNATIVE METHOD II (summary)

Complete disregard of LEWIIfmS will little weaken the proof, because we can base mainly on [a], [b], [c], [d], [e], [f] instead; any average age of death for true Covid-19 victims of ⩽73 y. proves the heavy Covid-19 mortality-fallacy by ISD (defined in the essay). At the same time, maintaining a high age is impossible without a strongly increased average number of chronic conditions. ...{the Supplement is almost fully 'Restricted'. Ask for it}... ...The method II, starting from the assumption that the official age structure of the 2020-DIC group was a true structure of premature deaths, proves this structure is impossible. This method was in 2025 validated by ChatGPT, the received numerical values were confirmed; it also agrees a high number of conditions is more important and says comparative risks for younger ones, but with a high number of conditions (diminishing LEa), in terms of real biological response to infection, are, on average, much higher than for simply older ones with the same LEa, and so the initial result concerning the required by the age-structure average number of current CCW conditions as >9 is an underestimation. ChatGPT basing on available sources suggested the death's Risk Multiplier (RM): 1.15 - 1.25 as a conservative path. If to take into account also severity of conditions, giving a bigger polarization of LEa among those with its high number, it would have to be yet higher (RM_total of 1.32++). Grok suggests a higher RM_total (1.50++). ...Adjusting the required (by the structure of the 2020-DIC group) 'mortality difference' (MD), between the sub -age-subgrups of the middle age subgroup, for a LE ratio of 1:2 gives huge initial 'mortality difference', close to 14.5 = MD1 = {[(MD - 1) /(LEaB/LEaA - 1)/(2.0 - 1)/] + 1} by the proportional method; finally the whole structure collapses. /The theory of particular combinations of conditions would make the structure collapse yet faster (+'Remarks' -f.). The av. number of conditions would be yet higher, as the higher number the bigger chance any combination is among it. /THE STRUCTURE OF THE 2020-DIC GROUP IS BY FAR IMPOSSIBLE FOR PREMATURE DEATHS (even regardless of condition counts and ignored LEWIIfmS), because among the highly-multimorbid ones, for a standardized LEa ratio of 1:2, even the highest possible here average age would require MD to be only 2* (*later validated by Grok); MD for the lower LEas' pair must be reasonably correlated with MD for the higher LEas' pair, being only a bit smaller (e.g. 2.5 vs. 3.5), not plenty of times (not 2.0 vs. 14.5 and more). The higher LEa from the lower LEas' pair is importantly higher than the lower LEa from the higher LEas' pair. ...This structure strongly resembles natural death patterns (including average ages in the age-subgroups of the 50++ age-range), with small deviations (+ p.4-7). Only with a very strong decrease in the share of the highest of our age subgroups (by its ~40% -still for a weak virus), both MDs can achieve the correlation (mainly due to the initially huge higher MD normalization). ...P.S. 1) The construction is very conservative; to make it yet better we could additionally take into accout 'a share of men seeking care (IfR x severe fraction) to a share of women seeking care' (M/F), which 'M/F' was ~1.15 for aged 85+, ~1.30 for aged 75-84, but huge ~1.55 for aged 65-74 (quite similar among highly-multimorbid ones), and ~ 1.25 for aged 55-64 [COVID-NET, JAMA, CDC]. The effect of a smaller share of men having high numbers of CCW conditions is weaker than the contrary effect of men having a smaller LEa (for the same condition count), even if the relative rise (with age) in prevalence of higly multimorbid men is a bit steeper [pooled MEPS analyses 2008-2012]; it will result in a redefining the highly-multimorbid ones. 2) The main script covers only 'Method I'. All materials are archived/digitized.

REMARKS: a) If specific numbers are provided in the text, e.g.: the average age of death in 2019 due to an injury (52 -p.3), the norm of 2008-CCW conditions for a group of alive ones with that of the DIC group age-dispersion (<3.5 -p.12) or of the current CCW ones for of-the-strong-dispersion av. age 67 (>4 -p.21) -they were calculated using sources given in the text and 'Discussion' [4, Injuryfacts; 18, 20, 26, CCWdata, CMS]. b) Raw deductions/recalculations-effects (p.6) mean basing on middle ages of age-subgroups. c) True Covid-19 deaths should have only lower than ADC average age and both must finally make 76.58 y. The flu serves as an example of external burdens; if with a proportional ADC's reduction then only with its <68 y. a share of deaths at age under 50 could exceed the one that the flu officially has (almost 8%); we have almost everywhere the considerably increased weight of the low age-range strictly against that of the higher age-subgroup (50-64), except for the 2011/2012 flu season, with illness rates always increasing between the age subgroups in the 2010-2020 period. The illness rates for Covid-19 decrease at advanced ages, strengthening a reduction in the average age of decedents compared with ADC. ...The above relates to avoided risks ('m' -a part of LEWIIfmS); no additional risks are considered in 'Next Steps' (p.19+) too, because it is a general epidemiological expectation that mortality shares among younger age groups should increase proportionally more than among old ones. {The flu's Ilness rates at age 65+ were usually small, but comparing those for 0-49 with those for aged 50-64, and comparing numbers of deaths at ages <50 to 50-64 to 65+, there were numerous deaths wrongly attributed to the flu at age 65+ (even 70% by Grok, to cease contradictions) lowering the av. number of chronic conditions, there (65+) rarely happen only one-two conditions, with a sharp picture, difficult to wrongly attribute to the flu. When two strong contradictory tendencies meet in one place (aged 50-64), something is false.}. /External burdens expected to kill with an underrepresentation of low age ranges are artificially created and selectively applied (to the oldest ones) by a man ones (p.22). d) ADcs =a supposed average death- age of real victims of Covid-19 (p.20). e) Assuming the final ISD of up to 0.8 and so a true Covid-19 deaths subgroup's size up to 7%, for the variant with the av. age 67 (p.22), is due to the assumed strongest behavioural factors (+ ACE2 gene expression in nasal epithelium declines in old ones, but it was still unknown by how much, so it was assumed that insensibly). A yet higher ISD would require comparative illness rates unreasonably strong increased for 65+ old ones, all in their terminal state. f) The average pre-infectious health-state must have been meaningfully worse (and so the strongly correlated with it average number of chronic conditions must have been meaningfully increased) for any specific age in a subgroup of true Covid-19 deaths, against that at the same age in a big comparative group of alive ones. The shift in the average age, strongly above 50 (mid-40s being the mean age of all infections in 2020), implies a significant shift in the number of conditions (+ remember about 'Superimposed Cumulative Effect'). After adjusting by illness rates, the share of people at age 65/67+ in the society of the U.S. is meaningly smaller than the share* of old ones (*among the old ones) having 10 - 30 of current CCW-conditions. The decline in LEa is much stronger with only the rise in the number of conditions from the age-average to 10+, than the decline while only aging from this 67 to 75 years; with a high number of conditions it means little for LEa if a person is e.g. 75 or only 67 instead [18], while the share of old ones in the population of the U.S. violently (>6-folds) declines with age, from those at age 60+ to those at age 80+(<4%), with the prevalence of high numbers of conditions rising 3-folds between somewhat under 60 and about 80 years of age [20(adjusted)] and <+50% between 60+ and 80+. [2, 6, 10, 18, 20]. Bigger than normal severity of conditions cannot noticeably limit their total number; first there must be picked up true Covid-19 deaths from the DIC group, then Covid-19 cannot choose only those with more severe one-two conditions and not with a total high number, also because the higher number the very disproportionately bigger risk a person has 'severe' conditions amongst it (it is clear, e.g. first, second and third of CCW ones all have a minimal, on average, negative impact on LEa) and this fact is already included into the av. LEa expected for different condition counts. Even if there were some cases where a high number did not result in an additional one being 'severe', then still one 'severe' condition plus many others make a person much more fragile than someone having one 'severe' plus few others. ...Later (2025) ChatGPT asked about it said: "The 'low total count but high severity' profiles are too rare and too low-risk individually = total count remains a necessary 'base layer' — severity does not offset or slow down the expected increase in total conditions among epidemic decedents." ...The required average number of conditions in the variant with an av. age of 73 as ~14.5 will not make the variant more probable, it needs enormous R = 0.975; we cannot assume Covid-19 killed those with av.LEa of only 5 years, as the share of old ones (60+) with LEa ⩽3 (excluding injuries and those in a terminal state, and adjusted by the illness rates) was <1.5%* for 2019/2020 (and their average age 86 y.), and no epidemiological practice show mortality among the rest of old ones could be ≫100 times smaller, it should be 3-9* times smaller for a severe epidemic (15-20* times for a mild one) [*Grok and ChatGPT]. g) The equation: 'ADcs + LEa1 = timely-LEWIIfmS' should be used on already dead ones. There is an extra up-correction of an 'initial LEWIIfmS', depending on Covid-19's partial and general mortalities; ...{this fragment is 'Restricted'}... the model itself explaines why without a high mortality among young ones a group like the DIC group must be composed mainly of those who were already in a dying state before getting infected, if their actual average age plus otherwise hypothetical LEa was much higher than LEWIIfmS. //ChatGPT agreed with us (2025), saying: "The overlap of average age and illness levels with pre-pandemic natural deaths indicates mass misattribution, strongly supported by your framework. If a virus were truly causing excess deaths, we would see a corresponding distortion in the average profile of decedents: many more conditions than normal (indicating it targets the vulnerable) and a wider age spread (indicating it kills across health profiles)". ChatGPT also agrees that true Covid-19 deaths should show a strong younger skew and a disproportionately increased share of younger fatalities (/due to a few reasons: the fact all deaths must be premature and due to the 'Superimposed Cumulative Effect', because of a much lower baseline among younger subgroups and due to the reason from the end of 'Shortened Supllement' = Risk Multiplier -counteracting the younger skew's diminishing, only increasing it, all next strengthened by falling in older adults with age illness rates -for epidemics with not rare fatal cases). All deaths must be premature by otherwise expected: from 2 to many more years; LEa ⩽2.0 y. (with excluded those already in a terminal state) is hardly predictable, unless among 90 and yet older ones, however younger of them usually must additionally be extremely multimorbid and the rest multimorbid, thus their share was <1 per mille.

CORRECTIONS: a) One thing is double-counted, and there is not fully correct reasoning about an initial ISD; but, if without ISD a pre-initial ADC is smaller than 76.58 then all gets very simple, because a positive difference between initial-ADC and 76.58 must then be a minimal one (as explained in last several lines of p.7), not as big as 0.09 year. b) The prevalence of 5+ conditions [20] rises in a nearly linear way from those aged 50-64 to those aged -79-, but for 2-4 conditions the rise slows considerably (p.12). c) A minor assumption concerns R2 (p.12) -a supposed average number of conditions in the 10+ subgroup was taken; if the theoretical av. age of 67 was stronger a result of the falling illness rates, or of a disproportionally increased share of younger fatalities (the rule, especially for systemic and respiratory viruses -ChatGPT), R1 of 0.93 (p.21) could be slightly lower. d) The sentence "AAADP concerns the value that should be, if the DIC group was not distorted by unreal Covid-19 deaths that happened in 2020" (p.13) is incorrect (mistake), it should be "...if without true Covid-19 deaths". e) The value on p.20 better to still call LEWIIfmS (not timely-) as it concerns age-normal health states. f) The average number of conditions for the variant with an average age of 73 is lower ~14.5; previously we looked there only from the end of condition counts what was not so correct (unlike using Standard deviations). Even if for an average age of 67 there should be ~10 conditions of CCW-like ones (p.21), its estimation is not necessary as matters very little, in contrast to ISD. g) The prevalence of 'Hyperlipidemia' was underestimated, while the prevalence of 'Non-Alzheimer's Dementia' was strongly overestimated; thus the newer 9 of CCW conditions weight a little more than 0.5 of 2008-CCW, but it only increases the required number of current CCW conditions.

LICENSE: Reuse (of the article, 'Supplement', 'Remarks' or of 'Additional Notes') needs a permission (until 01-2026, unless the date is changed) = confidentiality obligation. [drayse@proton.me]