Everybody is very optimistic about Omicron.
I HOPE they're right.
I FEAR they're prey to a statistical error, Simpson's Paradox:
I HOPE they're right.
I FEAR they're prey to a statistical error, Simpson's Paradox:
Simpson's Paradox says that you can see a trend because you're mixing two populations, but when you separate the populations, the trend is the opposite.
en.wikipedia.org/wiki/Simpson%2…
en.wikipedia.org/wiki/Simpson%2…
For Omicron, we might be mixing two populations: immune (through vaccinations or natural immunity) vs. naive (ie they have neither).
Maybe Omicron is *more lethal than Delta* for both groups.
But because those immune have grown dramatically since Delta, mixing both groups appears to lower fatality rates
But because those immune have grown dramatically since Delta, mixing both groups appears to lower fatality rates
Eg: 10% of pop immune when Delta arrives & Delta kills 1% of naive and 0.1% of immune --> average fatality rate=0.91%
90% pop immune when Omicron arrives & Omicron kills 1.1% of naive & 0.3% of immune --> average fatality rate=0.38%
90% pop immune when Omicron arrives & Omicron kills 1.1% of naive & 0.3% of immune --> average fatality rate=0.38%
It appears that your fatality rate has done down by ~60%, when in truth it's up by ~20%. You just have more immunity.
Let's *hope* this is not the case, but from the data I've seen so far, this sounds like the most reasonable outcome?
Let's *hope* this is not the case, but from the data I've seen so far, this sounds like the most reasonable outcome?
• • •
Missing some Tweet in this thread? You can try to
force a refresh
