@Jgalt thanks for the source for a higher res version! I found it most helpful to focus on the stars on the flag on the right.
But after seeing this version in addition to the Twitter one, it looks like these changes are regular and periodic, so they are then likely to be the MPEG intra-frames, i.e. an expected compression artifact.
@Sylvain I think my suggestion is only as outrageous as catching COVID-19 on purpose to win a question :D
Not that it matters, but if a state goes
this would not resolve positive, which is a fun hypothetical scenario.
- Political ad blackout
- Clampdown on voter dissuasion
- Ban on claiming ‘false victory’
- Lots more info for voters
@randallburns that's higher than his chances of winning the popular vote right now.
I can't find any numbers reporting only San Francisco cases on any of the sources listed in the question (JHU, CDC, or WHO.) Could someone point out where to find them?
@Tamay aw shucks, thank you! Honestly I don't know how this happened, I see everyone in the comments doing proper modelling and research, with a super deep understanding of epidemiology and the operations specific to Covid-19.
In the meantime I'm just here looking at numbers, eyeballing it, and somehow being all the way up there? It's honestly so confusing.
As of 19:07 GMT on April 1, which is seven minutes after 12:00 PT (the resolution time mentioned in the description):
Alameda: 313 confirmed
Contra Costa: 222 confirmed
Marin: 107 confirmed
Napa: 15 confirmed
San Francisco: 397 confirmed
San Mateo: 388 confirmed
Santa Clara: 890 confirmed
Solano: 49 confirmed
Sonoma: 78 confirmed
A total of 2,456 confirmed cases.
— edited by underyx
Thanks for the clarification. This made me flip to a much lower percentage, as I think the most likely outcome is broadcasting without a live audience. Precedent for this would be today's announcement that the Formula 1 Bahrain Grand Prix of 2020 is now going to be participants-only.
— edited by underyx
@underyx A sock puppet farm has been used and shut down on a Metaculus question for probably the first time.
@ninjaedit and I got -50, making it my worst prediction by far :(
Isn't this too easy to manipulate by flooding the site with 900k predictions? Not only does that endanger the integrity of this question, but even just one actor trying to manipulate the outcome of this question would introduce lots of noise to all other questions, as well.
@juancambeiro1015 awesome, thank you for the edit!
the DOB’s were all wrong, all being marked as 01-01-1900
The guy behind the US Election Project posted on Twitter last week that 01/01/1900 is set as DoB when moving to electronic systems if they cannot discern the correct DoB in manual records.
@underyx oh, I can answer my own question. If the only two possible resolutions are ambiguous or positive, the only reasonable prediction is 99%, however unlikely positive resolution is.
In this case, why not close the question retroactively, and wait and see if the resolution is ambiguous or positive by resolve date?