Thread by @EduEngineer

1. Yesterday, I got a text from my favorite scaly statistician, @drrollergator and was all,

HAVE YOU SEEN HAAS ISRAELI DATA PAPER, OMG THIS IS DUM AND MISLEADING

So, I was like, "Saw it. Haven't had time to consume it. Fill me in."

2. So, the gator (who should probably be POTUS) jumped on zoom and started showing Bayesian analysis and he was like

FOR THIS AGE RANGE AND THAT AGE RANGE, THESE THINGS DON'T LOOK AS GOOD AS THEY SAY AND MIGHT LEAD TO OVERCONFIDENCE AND EVERYTHING IS DUMB

3. He further explained that even the relative risks took advantage of the declining infection curve since those vaccinated spent more time on average at the low end of the infection curve.

4. He said

I WILL WRITE THIS ALL UP IN A FEW DAYS HOPEFULLY THE WORLD DOESN'T ALL DIE FIRST

I encouraged the writing part, but not the dying part. His substack is here:

drrollergator.substack.com

5. Then he asked if I had anything to add, and I told him it would be interesting to see what efficacy rates looked like under a sensitivity analysis in which x/100,000 who are vaccinated die quickly, but counted as Type II COVID cases.

roundingtheearth.substack.com/p/inventor-of-…

6. I told him I could write my own article on that aspect if he sent me the outputs of his program. A little while later he messaged me again:

OH GOD THIS IS ALL WAY WORSE AND DOES NOT LOOK GOOD FOR ANY VACCINE RISK-BENEFIT ANALYSIS A LOT OF IT IS NEGATIVE. EVERYTHING IS STUPID.

7. To be continued next week...