For those who have read my diaries on this subject, I keep track of the COVID numbers in some facet most of the time, and likely to be playing some math games with the data at any given time. In some of the previous surges I have felt exhilaration as my models did well for 2 weeks or so — WOW! Then my predicted peak was blown past. I hesitate to waste the time looking at the results...it is a deep wade, and frankly I feel of little use.
There’s evidence that Omicron can lead to a fast rising and falling wave — South Africa is the archetype
That rise begins about 11/21/2021 and ends 12/17/2021, about 4 weeks. Double that to account for the decline, it’s about a 2 month wave (waving my wands at the calculations and data...). This appears about twice as fast as the others rises/drops South Africa has seen.
Looking at the UK — it’s a little hard to see/guess when the Omicron surge starts, but 12/1/2021 seems to be a reasonable place. It’s now a month old, so it’s already as long as the South African rise, and no real sign of a peak visually.
So I decided to scrape a month’s data and look at the rate of change.
Even if the UK is getting set to peak and retreat, it still seems a few days away — putting the time frame at about 5 weeks. This is slower than South Africa it seems...whether by a week or 3 weeks is hard to say at this point.
The UK chart is a little hard to guess at a typical time frame for the valley to peak (or the opposite). The beginning and end of this year offer 2 periods where the rise/fall was about 2 months...so Omicron does seem faster. Again, maybe twice as fast, or not quite…
Here in the USA, we’ve seen enough rises/falls to get a characteristic time — I get about 50 days for a typical rise/fall, but it’s not very consistent. Currently our Omicron wave is about half that age — so if the half the time or so suggestion has merit, we very well might be close to a peak. But the rate of change is still high (8%) even in this holiday crash (I used a guess of 380K for the 12/31 average 7 day value for new cases)
For Florida, the peak/valley time is a little more consistent, averaging about 57 days
Interestingly, the 4 drops after these surges also average 57 days, but with about 1/3 the standard deviation...still — at this point we are again about half-way through this time frame with our Omicron surge — does the derivative of the new case curve look like it’s fixing to peak?
Sorry — that rate of change curve is not ready to slow yet, at least to my eyes. Of course things can change on a dime. I’m guessing the headline # of 56,865 is new data for yesterday, but I don’t see it yet at Worldometer. Assuming it is the most recent day it would be a little slower of a rate of change (dropping to about 9%), but still following the pattern of a holiday reporting upset. It would also pretty much be right in the middle of my 2 forecast models, albeit only 2 days in…
Predicting 1 day is trivial; 2 days is only slightly less so. A dozen days might show some skill, but the Holy Grail is to be able to see these peaks even a week out. My USA calculations are similar (I essentially assume FL will be about 10% of national cases), so if we start to see a decline in cases by 1/9/2021 (as there are some reports of), that seems like a best case scenario.
When have we had that type luck in this pandemic?
We’ll see where this is at the end of the week...in the interim, stay safe, be well.