CURVE FITTING UPDATE

Below is an updated chart of the curve that I had fitted after about a week or two of data from the New York times.  Each morning around 6 am I go to their website and take the current figure and credit it as YESTERDAYS number. Since their number each day is constantly updated there is some error in this depending upon when they receive updates and when I check their website. That number goes in the NATIONAL TOTAL ACTUAL COLUMN. (Fourth one over)

I then subtract the previous days number and get the NATIONAL NEW ACTUAL first column of data. When the new numbers were small, less than 200 they had a bar graph that was easily readable so I was able to get  all the new cases back to Thursday March 5 while I do not have the NATIONAL TOTAL ACTUAL FOR THOSE DAYS.

I was then able to fit a curve NATIONAL NEW PREDICTED that closely approximated the NATIONAL NEW ACTUAL. The fit is indicated by the DAYS AHEAD/BEHIND COLUMN. Those numbers are zero for almost two weeks.

By Wednesday March 18th the NATIONAL NEW ACTUAL (2436) exceeded the NATIONAL NEW PREDICTED (1758) so that it slightly exceeded the predicted figure for Thursday March 19th (2312) We were ONE day ahead of prediction.

As you can see the situation became worse. For five days we were TWO days ahead of the curve. We want to be BEHIND not AHEAD of the predictions. HOWEVER, things change on March 25th. I am a little suspicious of the data for this day, but it does mark a consistent change in which THE NATIONAL NEW PREDICTED are either on target OR ONE, TWO, and now THREE days BEHIND target.

The same phenomena is also evident in the relationship between the NATIONAL TOTAL ACTUAL and the NATIONAL TOTAL PREDICTED. Since these are smoothed curves the changes are less pronounced.

Now part of the DAYS AHEAD data might reflect the availability of more tests in some parts of the country which were able to detect more cases.  However the total picture is consistent with the hypothesis that social distancing may be beginning to show its effects, which should become even more evident in the days ahead.

By the way the orange color in the NATIONAL TOTAL PREDICTED was near to the number of cases in China at the time the curve was fitted.

		National	Days	National	National	Days	National
		New	Ahead	New	Total	Ahead	Total
		Actual	Behind	Predicted	Actual	Behind	Predicted
March
TUE	3
WED	4						350
THR	5	50		50			400
FRI	6	80	0	66			466
SAT	7	90	0	86			552
SUN	8	110	0	114			666
MON	9	150	0	150			815
TUE	10	200	0	197			1012
WED	11	260	0	259	1269	0	1271
THR	12	400	0	340	1660	0	1611
FRI	13	440	0	447	2100	0	2058
SAT	14	659	0	588	2759	0	2646
SUN	15	843	0	773	3602	0	3419
MON	16	880	0	1017	4482	0	4435
TUE	17	1399	0	1337	5881	0	5772
WED	18	2436	1	1758	8317	0	7530
THR	19	4075	1	2312	12392	1	9842
FRI	20	5570	2	3040	17962	1	12882
SAT	21	6418	2	3997	24380	1	16879
SUN	22	8638	2	5257	33018	1	22136
MON	23	10481	2	6912	43499	1	29048
TUE	24	15353	2	9090	58852	1	38138
WED	25	9656	-1	11953	68508	1	50091
THR	26	16873	0	15718	85381	1	65810
FRI	27	17255	0	20670	102636	1	86480
SAT	28	20981	-1	27181	123617	1	113660
SUN	29	18378	-2	35743	141995	0	149403
MON	30	21580	-3	47002	163575	0	196405