Autocorrelation & Admissions
In the last couple of posts we’ve seen how techniques developed in the field of signal processing can reveal hidden patterns within pandemic data that are gathered on a daily basis. I’m hoping people raised an eyebrow when we discovered a hidden weekly pattern for COVID cases, for unless we can argue that the virus does the same strange periodic dance every 7 days we have to pin that pattern on the weekly testing regime.
Since dead viral fragments can linger for 90 days and trigger positive tests then we are forced to conclude it is the imposition of periodic weekly testing that is causing the patterns we are seeing. The alternative is to conclude that viral fragments (alive or dead) are phasing in and out of reality on a 7-day periodic basis. This may appeal to Trekkies.
This is not a new finding and members can thumb back to my entry dated 7/7/21 (Is Viral Testing Churning Out Random Numbers?) to discover you can predict 92.5% of the variation in COVID cases using a simple linear mode…