@ImminentDownfall Welcome to Metaculus. This site is not a discussion forum, but a forecasting platform. See here for details and, in particular, note that

Metaculus aims at a high level of discourse. Comments should be on topic, relevant and interesting as much as possible. Much less welcome are highly (and especially irrelevantly) opinionated, spammy, aggressive, or low-information comments.

I found this podcast episode of the Ezra Klein show with Nate Silver ("Why 2020 isn't 2016") pretty informative. As a side note, when asked at the end of the show to mention a few books he liked, two of his three choices are Nick Bostrom's Superintelligence and Toby Ord's The Precipice.

There's now a petition to the New York Times requesting that they respect Alexander's wish to remain pseudonymous, signed by many prominent figures.

The methodology employed to generate this ranking seems very questionable to me. For Chinese cities, the authors relied on the most recent figures they could find (often going back several years) and adjusted them by a constant 20% rate of yearly growth. Thus, for the city of Chongqing, which tops their ranking, they used a figure from 2011 (500,000) and multiplied it by \(1.20^9\) to reach their estimate of 2,579,890. All they say by way of justification for picking that particular growth rate is that it has been "indicated by the Chinese Ministry of Pu...

@Jgalt McAfee has shown repeatedly that he has little concern for the truth. I don't see the point of signal-boosting his pronouncements, which are largely devoid of evidential value.

— edited by PabloStafforini

This is a proper occasion for reminding Metaculites that Travis Fisher holds the rare distinction of predicting the Covid pandemic back when few were even paying attention. See this Twitter thread for details. Fisher, to quote Slate Star Codex, will "g[o] down in legend."

Why follow the news when you have @Jgalt to keep you informed?

@nextbigfuture There are many more predictions than users, since many users make multiple predictions. Furthermore, some of these predictions are from users who registered after the question was posted, and who are as such ineligible to trigger positive resolution. With that said, I'm still very surprised that no positive cases have yet been reported.


When the facts change, we change our constitution. What do you do, sir?

[Gwern](https://www.gwern.net/Littlewood): >Online & mainstream media and social networking have become increasingly misleading as to the state of the world by focusing on ‘stories’ and ‘events’ rather than trends and averages. This is because as the global population increases and the scope of media increases, media’s urge for narrative focuses on the most extreme outlier datapoints—but such datapoints are, at a global scale, deeply misleading as they are driven by unusual processes such as the mentally ill or hoaxers. >At a global scale, anything tha...

Next Metaculus question: "By 2022, the mystery surrounding Metaculite wolfish_wrath's comment will have been resolved."

I'm somewhat confused: the title of this question says 'Metaculus vs. Experts', and the body mentions the Covid-19 Expert Surveys, but the resolution appears to compare the Metaculus algorithm with the Metaculus community, with the experts playing no role. Is my understanding of the resolution criterion correct, or am I misunderstanding something?

@(Jgalt) I recommend [re-reading](https://www.metaculus.com/questions/3688/will-the-2020-tokyo-olympics-be-cancelled-due-to-covid-19/) last year's declarations from senior members of the International Olympic Committee and the Japanese government to judge the weight one should give to these more recent statements. Back then we were told that "We will overcome the spread of the infection and host the Olympics without problem, as planned" (Japan's President); that Japan is "not considering cancellation of Tokyo olympics at all" (Japan's EconMinister); that...
@(Linch) >I'd also like the ability to add negative probability mass. For example, if I'm trying to predict that something's likely to happen tomorrow, it's hard for me to adjust the sliders such that I'm not accidentally making a prediction in the pass. I'd like a way to reduce probability mass of "will this happen in the past" occurrences (as well as predictions for mortality count that's not below the current level, etc). I think many of these situations could be easily avoided by allowing users to arbitrarily truncate the probability distribution. ...
The gap between Biden and Trump has widened by ~8% in the prediction markets (PredictIt, FTX). In my opinion, this seems like an excessive update given the weight of the evidence, and suggests that the discrepancy between the markets and the models was in part due to a prior irrational expectation by traders that the debates would cause Biden to significantly drop in the polls (an expectation that I've observed in several people in my social circle). One may object that the expectation can be seen as irrational only with the benefit of hindsight. But Me...
I lost 177 points on this question. My reasoning for assigning such a low probability to a positive resolution relied on two flawed assumptions: 1. I thought the clock wouldn't move to 1 minute to midnight, so I assumed it wouldn't move closer to midnight. It didn't occur to me that it could move in increments of less than a minute. This mistake was particularly inexcusable given that there is a precedent for a fractional move: in 2017 the clock moved from 3 to 2.5 minutes to midnight. 2. Like Jgalt, I thought it would be absurd to move the clock closer...

Alexa now says that only 17 sites are linking to Metaculus. I think these link counts are quite erratic and unreliable; not sure it makes much sense to have a prediction with this resolution criterion.