First, the assumption you bolded was my attempt to see what numbers would result if a critic of elimination of 3rd class medicals would make that claim. I think it useful to anticipate potential arguments to see where they lead.
Second, the idea of taking incapacitation statistics from car accidents and applying them to aviation accidents is not my idea, but comes from paragraph 4 of this document by the Aerospace Medical Association (AsMA) that was sent to the FAA opposing PBOR2:
http://www.asma.org/asma/media/AsMA...ical-Certification-Legislation-April-2015.pdf
Their math in paragraph 4 made the egregious mistake of multiplying an incapacitation accident rate per year by number of pilots, which yields a number with units that make no sense. They mistakenly thought the number was a pure "accident" count. The fact that it came out to 2503, twice as large as the last two year's annual average accident rate of ~1347, should have made them realize their math was hosed. My (hopefully correct) math yields about 16 extra crashes per year.
Interestingly, in paragraph 2 of their document they claim 5 of 180 fatal accidents (~2.8%) are due to pilot incapacitation [they don't provide the source for this number.] This accident rate occurred under the current medical standards, so by their own math and arguments the accident rate due to pilot incapacitation might be expected to rise from 2.8% to 4.0% (2.8 + 1.2).
Yes, that is one of the assumptions. My assumptions were deliberately chosen to be somewhat pessimistic relative to the position I hold. Since the alleged rate on aviation accidents due to pilot incapacitation is already ~2.8% while the car driver accident rate is ~1.2%, it may be that the third class medicals aren't preventing incapacitation accidents in any measurable way. In fact the AsMA document inadvertently admits that autopsies on pilots who died in accidents but were medically certified as "fit" found that 25% had moderate to severe "medical hazards." That's a large percentage of pilots! Either the autopsies aren't terribly reliable or the general pilot population is pretty dang unhealthy.