Jagermesiter1871
I've stopped even bothering to look at confirmed cases as its such a poor and incomparable stat. Deaths is far more useful and that is still growing massively, although there is of course a lag so hopefully in a few weeks time we might start to see a drop.
Although different countries may have different testing regimes, the average is very indicative. So, for example, the percentage relationship of cumulative cases to cumulative deaths is "stable". It's ALMOST fixed but creeping up over time at the rate of about 0.1% a day. It was
4.02% 17-Mar
4.04% 18-Mar*
4.06% 19-Mar
4.15% 20-Mar
4.26% 21-Mar
4.36% 22-Mar
4.36% 23-Mar
4.47% 24-Mar
4.53% 25-Mar
4.53% 26-Mar
4.58% 27-Mar
4.65% 28-Mar
4.71% 29-Mar
13 Days, a growth of 0.69 or 0.053 per day. In other words tomorrow you can look at the cumulative cases, multiply by 5% and you will be quite close. Incidentally the non-China number has a very similar growth, from 4.04% to 4.8% (different because the Chinese numbers have been virtually static for weeks now) 13 days to creep up 0.076% at a rate of just under 0.06 per day.
THE Stat though is, take the number of "Completed Cases". These are Discharged + Died. The ratio of these two (Deaths:Recovered) is again creeping up, but noticeably faster than Cases
13.92%
15.69%
16.32%
17.01%
17.85%
18.35%
This stat means 18% of those going into hospital don't make it out alive. It looks likely top go over 20%. How much beyond 20% is hard to guess.
*Interpolated, messed-up data
EDIT: Incidentally the model I'm using predicted 720,000 Cumulative Cases at midnight tonight. Actual number was 720,416
It had predicted 790,000 for close of play tomorrow but the curve seems to be flattening, so I would say the cumulative will be 775,000-780,000