You didn't make any math errors, but I'll point out that it's really best to compare deaths to the total case number - unless you're certain that most of the population has already had CoVid.
Please also be aware that in the US, Senior Living Apartments (for age 55 and older, sometimes age 50 and older) are a thing and they are included in the "care home stats." These are places where perfectly healthy, often working adults live.
Since you seem to be wanting to subtract the numbers of deaths attributable to "would have died in a few years anyway" group, that number is less than our total care home population.
People move into these facilities (sometimes buying into them) well in advance of their ultimate demise. 20-30 years ahead. They aren't nursing facilities, but are usually owned by the same companies. They do provide "care" (but see below)
There is one very near me. It's expensive, it's cute, it's classified under CA policy as a "care facility" but really only offers 24 hour concierge-like services (and a better chance. at a bed in a related nursing facility). There are also other senior living facilities that do not have any "care" and are not so classified - they too have been hit hard by CoVid. This is very popular with couples, as a big fear that people have is that "something will happen" to their spouse and there will be no one immediately nearby to help. Each apartment has a buzzer. There are always at least 2 employees on duty, one of whom has to be at least an LVN.
Anyway, we have about 200,000 deaths in the US with about 7 million known positive cases. Both numbers are under-reported, according to all the science. But, the ratio is roughly the same - so to be more meaningful (since CoVid isn't going away and frankly, the US isn't really locked down in any real sense - except for New York and New Jersey and Hawaii), we can expect the remaining 320,000,000 Americans to get CoVid at some point - as I posted yesterday. Actually, if we exclude the very old and the very young, one can argue that we're worried about approximately 220,000,000 Americans getting Covid, of whom...some will die (and we don't know exactly how many - for 50 somethings, it would be 1%).
To be really precise about our math, we'd use a total case number from about 3 weeks ago (since that's the pool from which 95-99% of the deaths are coming). So take your pick: 200,000 dead out of 6 million or out of 7 million (not a major difference).
My math says that of all Americans getting CoVid, almost 3% will die. If I were trying to back out the old people who would have died in the next 5 years anyway (as you're doing), I'd probably get nearly the same statistics as the scientists at the major medical research institutes (1% of 50 somethings, .03 percent of 35-49 year olds, .005 or less of 20 somethings, 3% for 65-75, etc. 20-25% of over 85, etc.
