I will not be updating the prediction model daily. But for those interested in keeping track themselves, here is the prediction for the next 14 days (starting 8-21). I may update it when it runs its course though, as a follow up.
The newest value from yesterday, 3 has shown a current deviation from the prediction
so for the last few days it has run below.
When using the model on the same data it was based on, there have been instances when the cases are in an exponential growth that the model appears to slightly over estimate, before coming back in line. Because this model is so deliberately nieve there could be many variables that account for that.
So we can revisit this in detail in a later review. Perhaps, for instance - when cases are lower, the ages are in higher risk groups (most easily to get infected + most at risk of death) but then as cases start to increase, the lower risk groups become more infected, lowering the average a bit.
That's a possibility.
As for the curve "flattening" or not - it is the last two values, 10 and 3, that send it far off. We will see how long it continues. Hopefully they stay low.
I see that new articles are promised. I'm tempted to subscribe to your substack. Here's the thing: I've already subscribed to a few others, and that $10 per month adds up pretty quickly as you start to add subscriptions. I'd love to see like-minded substack publishers come together in loose alliances in order to offer "bundles". Getting my top 10 favorite substackers for, say, $25 per month is a lot more appealing than having to pay $100 each month. I'm not disputing that the content is worth the money -- it definitely is. But many of us have families to raise and dollars don't grow on trees for us. Anyhow just a suggestion :-)
https://every.to is this bundling package that does this for newsletter writers. I don't know how MSM-biased they are, hope not at all, but you can consider this.
Continued Israeli deaths from https://www.arcgis.com/apps/dashboards/bda7594740fd40299423467b48e9ecf6
22: 55
23: 26
24: 24
25: 28
26: 29
27: 10
28: 10
29:3
30:54
31: 0
1: 43
2: 36
3: 7
4: 25
5: 51
6: 0
7: 0
8: 56
9: 52
10: 8
11: 17
12: 45
13: 23
14: 32
So the prediction has now failed? Curve flattened out?
The newest value from yesterday, 3 has shown a current deviation from the prediction
so for the last few days it has run below.
When using the model on the same data it was based on, there have been instances when the cases are in an exponential growth that the model appears to slightly over estimate, before coming back in line. Because this model is so deliberately nieve there could be many variables that account for that.
So we can revisit this in detail in a later review. Perhaps, for instance - when cases are lower, the ages are in higher risk groups (most easily to get infected + most at risk of death) but then as cases start to increase, the lower risk groups become more infected, lowering the average a bit.
That's a possibility.
As for the curve "flattening" or not - it is the last two values, 10 and 3, that send it far off. We will see how long it continues. Hopefully they stay low.
Thanks for that. Indeed we do hope they stay low! One can be both skeptical of the vaccines and hope that one is proved wrong.
What? I pointed so many people to this substack. They’re hanging on each day’s report.
I wonder if it would be possible to somehow crowdsource this effort? I expect he's busy and can't keep this up on his own...?
I see that new articles are promised. I'm tempted to subscribe to your substack. Here's the thing: I've already subscribed to a few others, and that $10 per month adds up pretty quickly as you start to add subscriptions. I'd love to see like-minded substack publishers come together in loose alliances in order to offer "bundles". Getting my top 10 favorite substackers for, say, $25 per month is a lot more appealing than having to pay $100 each month. I'm not disputing that the content is worth the money -- it definitely is. But many of us have families to raise and dollars don't grow on trees for us. Anyhow just a suggestion :-)
Send me an email at drrollergator@protonmail.com
https://every.to is this bundling package that does this for newsletter writers. I don't know how MSM-biased they are, hope not at all, but you can consider this.
Can you please comment on this study? https://www.nejm.org/doi/10.1056/NEJMoa2110475