Yep

edit: Maybe should wait for it to fossilise in the trendline at a 5min interval

edit: Wild - crashed to 410.00 from 475 literally just as the clock struck 10.

edit: Done! It's in stone

— edited by KrisMoore

@(AngraMainyu) wrote: > I agree with tom_sittler's suggestions. > Also, I'd like to suggest allowing less than 1% and more than 99% in yes/no questions. For example, I would put the odds of a US nuclear first strike on NK in 2018 are significantly less than 1%, but the system does not allow it. > > *— edited by AngraMainyu* If this were implemented, I think many new users, especially those not used to numerical forecasting, would be too tempted to make extreme predictions on events. Due to their inexperience, quite a few of those events would inevitab...
@(Barbarossa) wrote: > As expected, China wins this one. The "top 25 media outlets" link doesn't work for me anymore, but I found it here https://www.apnews.com/4997bb7aa36c45449b488e19ac83e86d > Or should we wait for confirmation of the claim? Confirmation definitely needed. It may well turn out upon further analysis that the editing simply didn't work. A key line in the AP article states that "[s]everal scientists reviewed materials that He provided to the AP and said tests so far are insufficient to say the editing worked". We should also remember t...

Note that the second semi-final will take place this evening, and its results could have a significant impact on people's predictions. For instance, it's not impossible that some of the current favourites on @PabloStafforini 's excellent spreadsheet – e.g. Netherlands, Sweden, Russia – get knocked out tonight.

Track record update is wonderful! Thank you. Suggestion: I want to directly compare my accuracy to that of the community + of the Metaculus algo at various points during the question's lifetime. (e.g. "Is my average Brier score 50 days before resolution better than the community's?") However, when I go to look at the average Brier scores at various time points, the "community" and "Metaculus" values are calculated including questions that I *hadn't yet predicted on*. This obviously means my displayed average Brier score is *not* directly comparable to...

Icecuber is now at 0.794. Do we resolve this retroactively by at least a few days before this point once the competition ends? I know the text of the question implies otherwise, but an earlier resolve date would prevent the slightly uninteresting behaviour of everyone just dogpiling on 99% and getting the exact same Brier score.

— edited by KrisMoore

I compiled a list of countries which recently gained independence from https://en.wikipedia.org/wiki/List_of_sovereign_states_by_date_of_formation . I thought some of you might find it useful when making your predictions. I restricted my reference class to countries whose "date of acquisition of sovereignty" was after 1992-12-31. I chose that date because it's a rough end-point for both the era of European decolonisation and the break-up of the USSR, and countries which gained independence as a result of those historical processes are probably less rel...
I found out recently that the antibody test on my blood sample has still not been run, so I suppose there is some chance that I'll eventually get a positive result. The biotech company I work at will be running many of these antibody tests and they have spent a lot of time evaluating different kits; I think they're going to go with one from Pfizer which from what I've heard has around or above 99% specificity and sensitivity. They may also combine results from multiple tests conducted on each sample. However, the [known cases here in Iceland represent 0...

@PeterHurford Article says that both parties have "agreed that their teams work urgently to complete the extension by Feb. 5", which means that the extension hasn't actually been ratified yet, right? Is there a chance that it won't be completed before the deadline?

— edited by KrisMoore

@DanielG wrote:

This is the last time I trust the media coverage instead of my intuitions.

I don't think the community prediction (which I can only assume yours was close to) was wildly off the mark here, given what was known at the time. Maybe there should have been a bit more upwards hedging to account for the brashness of the Trump administration, but if there is a shutdown, it'll feel to me like a "1-in-10 events happen 10% of the time" kind of deal, and that's fine.

I don't (yet) own a copy of Toby Ord's The Precipice, but I know that he tallies up estimates for the probabilities of all existential risks over the next century. Would anyone care to use those to come up with an estimate for this question, bearing in mind that not all of Ord's listed existential risks would necessarily entail human extinction, and that this question's timeframe is 80 rather than 100 years?

@(Anthony) wrote: > It has probably occurred to some predictors that there is an opportunity to gain lots of points by predicting a large number for this question, then making a contribution so as to cause that prediction to come to pass. I just want to assure everyone that this is [perfectly acceptable](https://www.metaculus.com/help/faq/#windfall). It's not impossible to imagine such a person pledging $500 for a month, and then cancelling that pledge a month later. Might it be worth setting the resolution value as the minimum of the dollars pledged 2...

Trevor Bedford, a good source of COVID information throughout the pandemic, says he is now 80-90% "convinced" that the B.1.1.7 strain is more transmissible.