still don't think modelling is possible based on the data we have but a few comments on forecasting in general
1. Forecasting's usually a game of second derivatives - the change in the rate of change. Is the number of new cases speeding up or slowing down? When will that change?
1. Forecasting's usually a game of second derivatives - the change in the rate of change. Is the number of new cases speeding up or slowing down? When will that change?
2. Following on from that, death happens at the end of an illness not the beginning. It's a noisy filter at best because it's affected by age structure and healthcare system stress, but even adjusting for those, deaths are a lagging indicator.
3. We are still in the early stages so IMO it still makes sense to ignore cures and think about gross new cases rather than net. Similarly, dividing things by populations is probably throwing information away; the infected population is small compared to the uninfected
4. Which says that the interesting thing to look for is a change in the sign of the rate of growth of daily new cases. (Reminiscence: A lot of good investors used to look at changes in new NPL formation to understand a credit cycle!)
