I have found that some are following this very closely and appreciate regular updates. So I will continue to update until the end of the prediction timeline (the 20th).
If one were to evaluate the model, the model is losing accuracy. The last four data points have avoided the 95% prediction bounds. But unfortunately it just seems like the model is off by a day, which means the number of people predicted to die over time is fairly accurate.
Yesterday’s (8/16) Israel recorded 19 deaths, under the lower bound of the prediction which was 22.
What do moving averages look like? Even if the model is holding as a trend, it should not be surprising that there is not day to day precise accuracy, especially with smaller numbers
Pfizer's own 6 month Safety and Efficacy study appears to agree that protection against death isn't a thing. See page 12 of the supplementary materials. Summary: 15 dead in the vax portion, 14 in the placebo (N ~ 22k for each side):
Are some of these misses due to reporting timelines? For example, deaths actually occurring on Friday, Saturday and Sunday might be reported on Sunday.
I wonder if there might be a regular spike on Sundays because that day is after religious observances on Friday and Saturday, when some people don’t use tech? I suppose it doesn’t matter much unless cuts in the study are made just before Sunday. Or?
What do you make of the idea that time between jabs is less than optimal in Israel while it is better I.e. longer in the UK? Of course, which vax was used is also generally different.
Hard to say. Pretend instead of two doses to get you to 98% for a while with a decline of some kind over time, you had single doses semi regularly to keep you at 85% (or whatever the regular single dose would be). The difference would have to be thought out - modeling what would happen could be difficult. And those models could not match the reality of doing so.
may only matter in so much as it pushed back the the timing of the vaccine degradation. so if the last shot was in May/June for most people in UK, the fall and early winter will be explosive.
What do moving averages look like? Even if the model is holding as a trend, it should not be surprising that there is not day to day precise accuracy, especially with smaller numbers
91-divoc.com allows you to do a 7 day moving average. It's a steep incline at the moment.
Pfizer's own 6 month Safety and Efficacy study appears to agree that protection against death isn't a thing. See page 12 of the supplementary materials. Summary: 15 dead in the vax portion, 14 in the placebo (N ~ 22k for each side):
https://www.medrxiv.org/content/10.1101/2021.07.28.21261159v1.full.pdf
https://www.medrxiv.org/content/10.1101/2021.07.28.21261159v1.supplementary-material
Are some of these misses due to reporting timelines? For example, deaths actually occurring on Friday, Saturday and Sunday might be reported on Sunday.
It is very possible - the predictions are a hard 14 days out, so if for some reason deaths on Monday are reported on Tuesday I cannot distinguish.
I wonder if there might be a regular spike on Sundays because that day is after religious observances on Friday and Saturday, when some people don’t use tech? I suppose it doesn’t matter much unless cuts in the study are made just before Sunday. Or?
At the very end, would you mind fitting a trendline to the Real Deaths?
We will review the predictions and underlying details more fully in the near future.
What do you make of the idea that time between jabs is less than optimal in Israel while it is better I.e. longer in the UK? Of course, which vax was used is also generally different.
Hard to say. Pretend instead of two doses to get you to 98% for a while with a decline of some kind over time, you had single doses semi regularly to keep you at 85% (or whatever the regular single dose would be). The difference would have to be thought out - modeling what would happen could be difficult. And those models could not match the reality of doing so.
may only matter in so much as it pushed back the the timing of the vaccine degradation. so if the last shot was in May/June for most people in UK, the fall and early winter will be explosive.