This is the whole problem with your way of thinking.
There is no percentage that is being extrapolated here.
It is not like the dead are only coming out of the sample pool.
Right now in Ontario, there are 423 dead out of 8967 cases for a death rate of 4.7%.
So far Ontario has tested close to 1% of the population.
Thinking theoretically, if you could instantly test the other 99% right now, the number of dead would still be what it is but the number of infections would be many times higher, thus the death rate would be many times lower.
The reason I think this is important is because if 4.7% was even remotely accurate, it would be way too risky to open up the economy.
Without accurate infection and death rates, opening up the economy is going to be a shot in the dark.
The methodology used in calculating the death rate for Covid19 is no different than how it's used for the flu, SARS, MERS and so on. It's always the numbers of death divided by the confirmed cases.
Again you are assuming that we have to test the other 99% and with that test there will be "many more infections" making the death rate lower. We don't test the general population on mass for other viruses and we shouldn't for the Coronavirus. With what you are saying are you also claiming that the SARS, MERS, H1N1 and all other viruses numbers are not representative of the virus because we didn't test the masses?
What we can say without a doubt is as of right now there are more than 2 165 000 confirmed cases of Coronavirus around the world and that's increasing with more than 144 000 deaths attributed to the virus and that is increasing, which gives us a death rate of 6.65% at the current moment. where that death rate will end up is anyones guess. All professional estimates point to a minimum 1% death rate attributed to the Coronavirus and some say up to 2.5%. Comparing that to the death rate of the flu which is 0.1% it's safe to say you are at least 10X more likely to die from Coronavirus than you are from the flu.
Again your opinion is highly assumptive on your part and fully non factual, that you think if we could instantly test the other 99% of the population the number of infections will see a drastic spike as you put it "many times higher" but yet we will not see significant new cases of critical care or the number of deaths actually increasing in the future.
Facts are the Coronavirus so far has about 3% of it's cases ending up in critical care, from those 3% of cases ending in critical care roughly 40% will not survive. So if you can figure out what your "many times higher" number of infections is as a number you can actually figure out a death rate.
But the problem becomes when a community and their healthcare system is over burdened with the number of confirmed cases as that 3% of critical care needed can not be handled by the healthcare system and in affect drastically increasing the death rate, exactly what transpired throughout Italy, Spain and is transpiring in New York.
Like I've pointed out numerous times there is a direct correlation from the death rate to the measures each country takes to fight the Coronavirus. The countries who social distance or quarantine the masses specially at a very early stage or the ones who takes drastic health check measures will have a far lower death rate than those who do not.