# beala Candidacy Hello Metaculites, I joined Metaculus in February of this year and have since found myself spending nearly all my free time thinking about forecasting. I’m drawn to it because I believe it’s one of the best ways to (1) check that my beliefs track reality and (2) encourage discussion of those beliefs in a way that is grounded in evidence and data. I’ve also found that it’s a great way to learn more about the world. My activities have so far consisted of [detailed write ups of my forecasts](http://www.usrsb.in/predictions.html), and mor...

@Tamay Wow! Another casualty of this conflict: my brier score.

I did a bit of research to try and find recent events comparable to what's happening in Minneapolis. These were all large riots triggered by racial tensions lasting days or even weeks with looting and fires. - [2017 St. Louis protests](https://en.wikipedia.org/wiki/2017_St._Louis_protests): [$10 million](https://news.stlpublicradio.org/post/protesters-claim-more-10-million-impact-economist-says-overall-thats-minuscule#stream/0) in "impact", 300+ arrests, lasted weeks - [2016 Charlotte Riots](https://en.wikipedia.org/wiki/Shooting_of_Keith_Lamont_Scott#P...

AP calls it for Biden.



The Associated Press declares Joe Biden the winner of a grueling campaign for the American presidency. He will lead a polarized nation through a historic collision of health, economic and social crises.


@godshatter SCOTUSblog agrees with this assessment:


For those dreading, or hoping, that a conservative 6-3 Supreme Court with three appointees of Donald Trump will overturn the results of the election and deem him to be re-elected: there is absolutely no chance of that happening, whatsoever. None.

Looking over the last ~100 years, back to the Red Summer of 1919, I count 7 events in the US that would qualify as a riot as defined in this question. So, I expect 0.07 riots per year, or a 7% chance of a riot in any given year. Given that this year is 30% over, my prior is a 5% chance for the remainder of 2020. I update 2:1 given how unprecedented the pandemic is and 10:9 for the recent lockdown protests. In general, I don’t see protests as a strong predictor. They happen all the time. The economic downturn is unprecedented, but I don’t see any riots c...

News outlets reporting intel that the plan is to divide Ukraine in two making an invasion of Lviv less likely:

Desired goal: Seize the leadership of the state (not specified who) and force them to sign a peace agreement on Russian terms, threatening large civilian deaths.

Result – Ukraine can be divided into two parts, like East and West Germany.


@Roko I disagree with this. High quality questions and an active moderation/admin staff to help review, adjudicate, and resolve questions is why I like metaculus. There are plenty of other places on the internet that have a more laissez faire attitude. If anything, I’d like metaculus to be more focused and opinionated rather than less, and I don’t think this scheme would achieve that.

Is there any interest in continuing to track these events? Any suggestions for what a good next question would be? Some ideas:

  • Will there be a China-India war before 2022? (Coming up with the the trigger is tricky, since neither side may formally declare a war.)
  • Will there be more than 1,000 total deaths as a result of clashes between China and India before 2022?

— edited by beala

The Economist keeps historical data on big mac prices for its Big Mac Index.

2017-01-01: 5.06

2021-01-01: 5.66

(5.66 - 5.06) / 5.06 = 0.12

3.79 * 1.12 = 4.24

This seems like the appropriate amount of effort to put into this question. See you all in 2025!

— edited by beala

@(j.m.) I mostly agree with this, but I think the final prediction is a bit overconfident. I posted Gelman’s recent blog post [below](https://www.metaculus.com/questions/5383/donald-trump-usa-president-on-2021-01-21/#comment-44411). He co-authored The Economist’s model and thinks it’s overconfident. It sounds like it would be closer to 538 if the tails were tuned properly. And [here](https://fivethirtyeight.com/features/im-here-to-remind-you-that-trump-can-still-win/) is Nate Silver on his model: > Here’s what it seems safe to say, though. In an elec...
[Rollout of AstraZeneca Covid-19 Vaccine Halted in South Africa After Study](https://www.wsj.com/articles/astrazeneca-vaccine-doesnt-protect-against-mild-and-moderate-cases-from-south-africa-strain-11612700385) > JOHANNESBURG—South Africa said Sunday it would halt a planned rollout of AstraZeneca PLC’s Covid-19 vaccine after a small clinical trial found that it doesn’t appear to protect recipients against mild and moderate illness from a fast-spreading new strain of the coronavirus first detected in the country. > > ... > > Of the 39 volunteers in the...

The Strategic National Stockpile (SNS) has stockpiled enough smallpox vaccine to vaccinate every person in the United States.


IIUC, the smallpox vaccine is effective against monkeypox.

Carl Bergstrom is skeptical of the 90% number.


I don't believe the 90% number. This "arm" of the trial was unplanned and resulted from a dosing mistake. Was it randomized equally across all subpopulations (ages, locations)? I believe not but don't know the details. If not, you have multiple comparisons+confounding.

Related tweet thread making the same point: https://twitter.com/ct_bergstrom/status/13310…

I’ve been waiting to see if there would be any new information on the death. So far, it looks like all we know is that a soldier associated with the Indian armed forces died after stepping on a landmine planted in 1962. We do not know any of the circumstances surrounding the death, such as what the purpose of the operation was, or if it even was part of a broader military operation. Thanks for pointing out the ambiguity here, @AvrahamEisenberg. One point I’d like to make is that I think it’s reasonable to assume that resolution criteria always apply to ...

Here is a simple linear regression on the OWID data with China and Russian subtracted out: https://docs.google.com/spreadsheets/d/18_lHz…

This gives a date of Feb 15th (Jan 22 + 3.5 weeks)

Hopefully the roll out will be better than linear.

I’ve written up my prediction here: http://www.usrsb.in/three-gorges-dam-2020.html

The tldr is that I don’t think the evidence circulating on social media is very compelling (satellite images and rumors), but the amount of rainfall happening right now is worrisome and has caused at least two other dams to fail (one had to be blown up, another “slipped”).

@juancambeiro I believe we hit this (10,955,906 doses) on Jan 6. Can someone double check my work? I subtracted out Russia and China: https://docs.google.com/spreadsheets/d/1-oplm…

If I did this right, it should resolve January 30th.