It might sound counter-intuitive, but Russia's best interest now is to deescalate the conflict as soon as possible - to keep the gains, avoid long term damage and make some reforms in a calm situation. So, the purpose of the increased threats is to persuade NATO and Ukraine there is no way of winning and try get the best deal out of negotiations. The nuclear "red line" (e.g. attack on Crimea, Putin, critical infrastructure) is kept ambiguous for the same reason. Since leaders and people develop "tolerance" to threats over time, the rhetoric and action...

Is there a reason for further delaying the resolution of this question?

— edited by OpenSystem

@admins I agree with @citizen. Looks like 19.5k is the best number for resolution.

Looks like NATO support to Ukraine and the level of anxiety in the comments are both increasing.

Thanks to the new feature, adjusting from 2% to 2.1% :)

@(casens) If I understand the new criteria correctly: Ukraine casualties - at least 10k according to all sources (rus&ukr). Russian army has at least 5k confirmed by name. Civillian 5k dead confirmed by UN. DNR & LNR acknowledge 3k dead soldiers. This gives an absolute minimum of 23k. In such particular case it makes sense to wait a few more days for newer reports. However... Using such resolution criteria will most likely hugely underestimate the losses, it's clear that huge amounts of civillians were killed in Mariupol and that much more than 5k russ...

@Gaia I appreciate your kind answer. However, even after reading, Its still unclear which number to use for the resolution.

@aristophanes I tend to agree with this view. My "gut feeling" is that the continuous questions are responsible for the most fluctuations in my score, and the "luck" factor is much larger than in the binaries. I wonder if in reality the variance of gain/loss across all players differs between question types. Perhaps it can be checked by admins who have the entire picture.

@(ForkLeaf) I agree with most of this analysis. Currently, Ukraine can not upkeep the invasion of Crimea economically and militarily, perhaps only capture a small part with a lot of help. There is a VERY big threat of escalation with Crimea and US might be reluctant to continue supporting such thing as much as needed to achieve the result, especially capturing Sevastopol, which is in the south with nowhere to retreat. The current status quo or at least current reality is that Crimea belongs to Russia, so acquiring it through negotiations is not possible...

@(casens) My personal opinion - it makes sense to resolve the 25k total deaths positively, because most people predicted it before the discussion started and because the conservative criteria has been met. Others (50, 100) can be resolve ambiguously now, in my own experience and from tracking the discussion there was indeed lots of confusion, so the points might not represent the participants prediction ability correctly for those questions. I also noticed the general tendency of numerical questions to suffer from low "reliability and validity" of the re...

@casens Can you give a couple of concrete examples of a situation where question resolves true? e.g "Aljazeera reports 55k, India reports 45k, Russia 35k etc.." -> Resolves Yes/No

@admins Thank you for resolving. However, looks like the standings and the tournament score were not updated following the resolution, at least for me. In the past, I noticed that the score was updated immediately, while the standings used to change on the next day.

Could you please check for a possible reason for such behavior, perhaps a bug?

@admins Could you please elaborate on the status/conditions of this question's resolution? There has been a long discussion, after which this one was reopened, other similar questions remain ambiguous. I am still confused...

Does anyone know which number will be used for resolution?

1. Border crossings from Ukraine (since 24 February 2022) - currently about 7M

2. Individual refugees from Ukraine recorded across Europe - currently below 5M

3. Ukraine border exits - entries

Thank you

@(Homing_Sliver) @RyanBeck Wow, that's a lot of time that alternative and identical source was available. In current situation advantage will clearly go to whoever noticed that sooner and read the comments (activity and luck), rather than someone who predicted accurately a while ago (skill). As the spike is currenly growing, the point system will "punish" a lot of less active players, which predicted very closely to the real number, but didn't revisit the question. Alternatively, someone can notice the spike, someone who has never predicted the question...

@(Homing_Sliver) I totally agree that signals appear during the entire year. However, the prediction is of a composite indicator, which is not known for certain. There can be, in theory, a huge negative inflation in the last day of December. However, after Dec 31, no surprise is possible in the reality itself, so its totally reasonable to for users to stop predicting and just wait. Everyone in this discussion benefits from noticing the spike early and the question being left open till IMF graph is updated, but it still makes little sense to introduce mo...

@NMorrison and the team, I appreciate and thank you much for making those challenges possible, interesting and inspiring.

@RyanBeck Most people follow only a small part of discussions. Also, when an outsider looks at the graph, he thinks the crowd prediction changed, when in fact, it's due to "adjustments".

@(RyanBeck) @EvanHarper Thank you for clarification. However, I still fail to understand the gap between 358 (total points shown) a day before and 31 (on resolution). I was on 100% coverage, with predictions better (by density) than the crowd at all points. Log score seems strange too. Btw, it works both ways, there was a case that I got like +100 points and +0.1 log score for questions predicted seemingly much worse than the crowd average. Shouldn't the user interface provide us with a correct estimation? I mean, what's the point of having it then if...

@casens Thanks for resolution and analysis.

@Prodicus @tryingToPredictFuture I believe people get better at predicting, but also better at using Metaculus, and assume that there is a selection effect, but not sure. Curious to hear insights from the admins or someone that has experience with this kind of analysis.