All the people I've ever seen calling the 3.4% figure "flawed science" only seem to ever attack it for being too high. There's very rarely ever voices from the

All the people I've ever seen calling the 3.4% figure "flawed science" only seem to ever attack it for being too high. There's very rarely ever voices from the other direction of calling it too low. People underestimate how many cases Hubei actually managed to treat. They built two purpose built hospitals where every bed was an ICU and converted stadiums and hotels to centralized quarantine facilities so the less acute cases could be near but not in hospitals but could be rapidly shuttled there if the symptoms got worse. They flew in over 40K+ medical workers from all over China to treat less than 80K cases, 80% of which were "mild" which means 2.5 medical workers per non-mild case. What we really need is the cohort from China who had hospitalizatable symptoms and were denied any sort of treatment, what the range of outcomes in those cases? Even then, it's not a natural experiment because China is practicing triage, same as Italy, and so we need to compare like to like as much as possible. If the models are correct at assuming that our hospitals are going to get more than 10X overwhelmed, then this is the real number that matters. What's the rate of death for people who have 0 medical intervention. The reason why is, if the healthcare system isn't overwhelmed, then the CFR is not an important number because the total number of deaths is small enough that variances in that range are trivia. If the healthcare system is overwhelmed, then any data we get from non-overwhelmed systems, which is all people appear to be using, is utterly bollocks. If people are suggesting that 70% of people in their country could be infected in a year, then we should be using a CFR closer to 10 - 15% as the true number and that means we could be imagining a world where 7 - 10% of all people are dead by this time next year.