Just want to say that I moved from Kyiv to Lviv on Feb 13 /entirely/ thanks to this prediction thread and the Metaculus estimates. (Still in Lviv but leaving Ukraine later today.)

It’s hard to estimate the impact, but at the very least I have saved myself an enormous hassle trying to move all my possessions, finding an apartment in Lviv, etc on very short notice — which would have happened otherwise. Of course, worse outcomes were possible too if things had played out a bit differently.

Thank you, everyone.

— edited by availablegreen

@casens To see this in my inbox:

"Metaculus Notifications <notifications\@mg2.metaculus.com>

Resolved: Will at least one nuclear weapon be detonated in Ukraine before 2023?

Was probably the scariest moment of the whole year to me.

DeepMind: [Gato🐈a scalable generalist agent that uses a single transformer with exactly the same weights to play Atari, follow text instructions, caption images, chat with people, control a real robot arm, and more](https://twitter.com/DeepMind/status/1524770016259887107) DeepMind publication: [A Generalist Agent](https://www.deepmind.com/publications/a-generalist-agent) >Abstract >Inspired by progress in large-scale language modelling, we apply a similar approach towards building a single generalist agent beyond the realm of text outputs. The agent, ...

Came here to note this paper also. For the record, when I wrote the question this is the sort of system I had in mind to resolve it. I doubt this actual system resolves it, but a future version of this system that did perform well enough would satisfy the "unified system" criterion.

For reference, I surveyed battles/sieges listed https://en.wikipedia.org/wiki/2022_Russian_invasion_of_Ukraine not including ones involving Kyiv. Mauripol was first surrounded sometime in the period March 2-4 https://en.wikipedia.org/wiki/Siege_of_Mariupol, so it's been holding out for 8-10 days. Mauripol has about 500k population. Kherson, which is around half the size of Mauripol (250k population) fell about 4 days after it was surrounded https://en.wikipedia.org/wiki/Battle_of_Kherson Kharkiv, with 1.5 million people, has been the site of battles s...
# RyanBeck Candidacy Hi everyone, I'm Ryan. I've been forecasting since 2020 and have been active here since early 2021. I'm in the top 50 in [all-time points earned](https://www.metaculus.com/rankings/) on Metaculus, I've forecasted on 988 separate questions, authored [51 questions](https://www.metaculus.com/questions/?search=author:RyanBeck), and authored [4 notebooks](https://www.metaculus.com/questions/?status=notebook&search=author:RyanBeck). I'm interested in becoming a moderator because I think forecasting is important and I'd like to see it grow ...

@GameTheory both your comments are in violation of our Etiquette rules 1 through 7 and our Moderation rule 3. I believe that's a record. To be explicit, your comments are:

  • Low quality.
  • Disrespectful.
  • Insulting.
  • Conflictual.
  • Strongly political.
  • Intolerant.
  • Calling for violence.

As a consequence I will ban you for a week. Please change your behaviour when you are back.

my apologies for the mistaken resolution notification. no points have been awarded, predicting will continue.

As of right now, today is the first day since GJI launched their question on Jan 28 that the superforecasters are above the Metaculus community when you aggregate their options, 80% vs 75%:

  • 64%: Donetsk and/or Luhansk oblasts + other Ukrainian territory
  • 15%: Donetsk and/or Luhansk oblasts only (includes separatist- and Ukrainian-controlled areas)
  • 1%: Only other Ukrainian territory

NathanpmYoung Candidacy

I like forecasting and want to make it easy for people to do it. I spend 2-5 hours a week on here anyway and I think I could happily look over and moderate questions in that time. I run a weekly collaborative forecasting call on the discord.

@predictors I again apologize for the mistaken resolution email sent today. The fault is my own; and essentially comes down to me having more than one tab open and not seeing which question I was resolving, until it was too late. I know tensions are quite high, and this error caused significant distress. The Metaculus team will be working on safeguards to prevent this kind of error in the future.

Emails are (in a small way) a bit like bombs -- they can't be un-delivered.

Moscow Bureau Chief for the FT on Twitter:

Putin speaking now: he says Russia has recognized the separatist's claim to the whole Donbas. This is as close as you can get to declaring war.

For context, Ukraine controls ~70% of that territory.

— edited by ClayGraubard

@liljaycup Hodges seems pretty unimpressive to me - at the end of that interview he says he signed onto a letter calling for a no-fly-zone because “we have to do something”, which doesn’t inspire confidence in his reasoning. His comments about the huge Russian convoy were also contradicted by Michael Kofman in yesterday’s War on the Rocks podcast. Wouldn’t be that surprised if he’s just become a lazy talking head since leaving the military.

— edited by occamsbulldog

We want predictions on Metaculus to be valuable and informative and we take the threat of manipulation seriously. This question has received a significant amount of discussion regarding the potential for prediction manipulation as well as the character of predictors. We hope that we can alleviate some of these concerns: we have investigated the predictions and comments and have not found evidence of prediction manipulation or coordinated bot behavior. Here are two of the factors we've investigated: * There are only a few percentage points difference be...
Prior (Poisson process): Russia has invaded three times in the past 20 years (Georgia 2008, Ukraine 2014, Ukraine 2022). Call that \(\frac{3 + 1 invasions}{22 + 1 year} \approx \) 20%/year. Remove a 1/6th because we are 1/6 of the way through the year to get \(\approx\) 17% In the past there's been 3-month+ lead times where Russia's amassed troops around another country. If we say Russia does not have such a troop build-up now and a 3 month lead-time is required 80% of the time, we're down to 13% chance of an invasion this year. But perhaps Russia _...

@EvanHarper

I went 18% -> 14%. The bombings and the Russian watniks' reactions indicate there is no need for Putin to use nukes. A large-scale conventional bombing is sufficient to please the hawks, even after major humiliations like the Crimea bridge explosion.

Resolving positively today, February 23, 2022 at 6:49pm Pacific Time based on: 1. Russian President Vladimir [Putin's emergency speech](https://twitter.com/The_Real_Fly/status/1496718133746405377?s=20&t=8Y5JHlLf4bdKX2yyaefk5Q) stating ["We have decided to launch a special military action"](https://twitter.com/DAlperovitch/status/1496678639911510024). 2. US President Joe Biden's [statement](https://www.whitehouse.gov/briefing-room/statements-releases/2022/02/23/statement-by-president-biden-on-russias-unprovoked-and-unjustified-attack-on-ukraine/): "The ...
My predictions in the last few weeks (~80%) have been quite a bit higher than the community's for some time. So I thought it would be worthwhile to write out where I'm at. Thoughts/critiques are always welcome. I think it's worth trying to take a step back from the daily (or hourly) updates and think about what the likely endgame scenarios are. I see four major ones, starting with the clearest and ending with the muddiest.  1. **Ukraine folds**. Ukraine agrees to implement the Minsk agreement according to Russia's interpretation, and swear off applying ...

@mumpskin But the US IC often makes bad predictions, such as, e.g. the prediction that the Afghani government would be able to keep control over Kabul for some time after US troops left.