Why the Weekly Community Numbers Reported are Lower than Reality

With the map looking as bad as it does, you would think that it was likely as bad as it really is... but no, the reality is worse than the Baker administration is reporting.  They're doing their math right, and I don't think I have sufficient access to check the figures to confirm that those figures are in fact the proper ones but I also have no reason to believe they would lie.... and they don't need to in order to skew the results a bit in their favor.  You see, math can lie depending on what you're using to do the calculation.

I have long assumed that they were comparing like to like to do their comparison to determine the quantity of new cases in the last 2 weeks.  Like to like would be the known total for a city/town from on a particular day vs the same value 2 weeks later.  Like to like could be other options as well, provided those numbers were equally stable.  This is not the case, however.

The Baker administration appears to be taking, on a single day, the currently known total from the past Saturday and the currently known total from 3 Saturdays ago (2 weeks prior to the past Saturday).  Why is this important?  Because numbers don't stabilize for about 7 days.  The first few days after a given date, the new cases uncovered that started on that date continue to increase by triple digits.  Thereafter for a few more days it increases by double digits.  Then we're down to single digit increases (and, more rare but not never, decreases).  Why is that important?  Because you're saying the new cases in the preceding 2 weeks are based on a number that is fully known vs a number that is definitely not yet finished increasing.  The outcome is that we're always seeing fewer new cases in the report than we actually had, because some of the new cases are as of yet unknown.  

A more proper analysis would be to track like for like.  Equally stable numbers, or at least more equally stable numbers.  For instance, the case count total known on a given day vs the case count total known as of a date 2 weeks prior.  

How different are these two approaches though, really?  Well, let's do a side by side comparison (note, the rate is based on the state report.  There is no population of MA that I can find that comes to this so I will use 6.99 million people for the population for the next calculation as this matches the outcome within the state's calculations):

• State option of using the much better known figure and subtracting a known-to-be-lower-than-will-end-up-being figure: 149,074 - 137,517 = 11,557 or 11.8 per 100,000 per day 
• My option of calculating known on a given date and subtracting the quantity known 2 weeks prior: 150,498 - 138,083 = 12,415 or 12.7 per 100,000 per day. 

Sure, this leads to only a difference of 1 per 100,000 per day, but in a situation where the green is <4 per 100,000 per day, yellow is 4-8, and red is >8, a difference of 1 seems pretty significant.

Make no mistake, the Baker administration is doing everything in their power short of giving actual false numbers to try to convince you that everything is better than they really are.

Stay safe. Stay sane. Stay informed. Stay understanding that Baker does not have your safety as his top priority or anywhere near it.