Some fun food for thought I learned about. You can attempt to guess true infections based off a set of estimated or known parameters. For example, with a known numbers of deaths alongside an estimate infection-to-death average you can plug in mortality rates to estimate infection counts.
For example, using the estimate of 15 days from infection-to-death I saw, and using the roughly 1k deaths yesterday, you plug in estimated mortality rates.
So at 0.5% mortality you'd estimate 1000/0.005 or 200,000 newly infected 15 days ago. Or at 2% there would've been 50,000 newly infected 15 days ago.
Mortality time varies and these are all estimates, but fun food for thought. Once deaths get high, you're looking at a response that massively lags what's happening at the presence. Making it utterly useless.
Meanwhile, people wanting to act when the daily death rate reaches a high value are subsequently planning action well into an unfolding catastrophe.