I've changed my mind, I think resolve ambiguous and re-open with clarity. There's a lot more disagreement here than there should be.

@alexrjl Also, traviswfisher again confirmed time traveller.

I feel like if we want stronger operationalisation we should just go for something like

This question will not be resolved by any human. At the point at which an AGI has either joined the metaculus team as an admin/community moderator, or gained access to admin powers through other means, it can resolve the problem itself.

There's now a question-writing subchannel on the Metaculus discord. It's currently mostly me spamming it with questions I'm trying to write, but bringing some of your favourites from this thread over would be a great way to get them written!

@SimonM haven't checked but I'm guessing A-B list celebrity divorce rates are a better reference class and could be substantially higher.

I'm very happy to answer questions about what it's like to be a community mod from people who are considering applying. Feel free to ask here, but I'm also on the metaculus discord if you'd rather speak privately.

That [this](https://www.metaculus.com/questions/5973/transit-activity-in-dallas-tx-for-february/) question closes **after** every other question resolves poses quite severe problems for the scoring mechanism. In particular, predictors who are just "outside the money" have strong incentives to be extremely overconfident in order to maximise their chance of having a big enough points difference to overtake whoever's in third. People in the top 3 spots have a difficult decision with whether to try to react to this by also extremising (and risk then losing t...

I'm fairly unconvinced that this method for this kind of list question is a good idea, mostly because everyone can just predict a "comb" on the 5 people. Tagging other @moderators for thoughts, and @admins for an ETA on an option to have a "multiple choice" question like this with neater implementation.

The scenes when @traviswfisher deliberately makes bad predictions on a couple of questions which resolve on a particular date, drops to second, but has a super narrow spike in this question which bumps him back to first. :D

@SimonM I think the community median is likely to be predictive of the answer to this poll.

@Linch Of the top 20 countries by population, I think the descriptor "basically a straight line on a log scale since April 20" applies to (ordered by total deaths):

  • Brazil

  • Mexico

  • India

  • Russia

  • Pakistan

  • Egypt

  • Bangladesh

  • Nigeria

  • Ethiopia

edit: I'm currently at 90%

— edited by alexrjl

27 of the 60 nominated authors of the last decade have been women. In that time there have been 11 winners, 5 of whom were female, however this includes two women sharing the award in 2019, despite the rules explicitly prohibiting joint winners.

— edited by alexrjl

We are the 99%! Don't let the elites hoard points for themselves while screwing over the rest of us!

@Jgalt We've also come extremely close to positive resolution on a couple of occsions. In one case, a Metaculus user's girlfriend, who he lives with, tested postive. He then developed symptoms and tested negative.

I also recently developed symptoms and got tested, but it was also negative.

edit: @NoUsernameSelected also had a close call.

— edited by alexrjl

I'm a little worried what the resolutions in 2 days will do to the rankings. Rather than having been extra motivated to fight my way into the top 3, I've been a little lax with updating this month :(