Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

metaculus.com

Post reply on HN
Print view View original

291–300 of 419 posts

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#291

Post by rdtwo »

s1artibartfast wrote: post #181
jessriedel wrote: post #43

Earlier quoted context omitted.

To be clear, Metaculus is a prediction aggregation platform, not a prediction market. There's no money at stake, only reputation. https://www.metaculus.com/about/ It's well recognized, including my Metaculus users and admin, that money can provide a powerful incentive for accuracy which they do not use (presumably because this would be a very different, more legally encumbered platform). However, they attempt to make…

IT looks to me like there is some money to be made: https://polymarket.com/market/will-nato-expand-in-2022

I’d bet on no just cause how long it would take to certify and ratify. I don’t like bets where I put up a lot to win a little.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#292

Post by pmoriarty »

otherme123 wrote: post #82
Chris2048 wrote: post #48

Earlier quoted context omitted.

> if accompanied with guarantees of no further aggressions Which Russia cannot provide since they already promised not to invade.

They can agree on a DMZ, like in Korea. Remember that Putin negated the invasion while amasing an army at the border. You can't invade if your troops cannot be piled up near the border.

A DMZ is a good idea, but it doesn't address Russia's demand that no nukes be stationed in Ukraine (or other countries on its border) either.

I'm not sure how much that matters in practice, since nukes can be delivered by submarine or through ICBMs, but regardless, that's what Russia wants and a DMZ won't address that demand.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#293

Post by time_to_smile »

oezi wrote: post #246
time_to_smile wrote: post #150

Earlier quoted context omitted.

> no confidence that we can really know anything for sure one way or another Not only is this true from a general sense of "fog of war" and misinformation, but even in a purely statistical sense, these numbers aren't compelling. As a stats person, both 0.9 and 0.26 are probabilities that basically reflect nothing more than a vague sense how the outcome will go. In a classical NHST this degree of confidence would not…

I am not sure I follow. Are you saying that a 90% chance of not needing an umbrella is the same as a 26% chance? These are vastly different probabilities which mean that you might be wrong 9 out 10 days, or right 3 out of 4 times. As another example: I bet most companies would die to turn up conversion rates up even 1 or 2 percent. When dealing in risk management where probabilities should start in 1 in 10 or 1 in 10…

> These are vastly different probabilities

When it comes to certainty about an outcome (i.e. a hypothesis), they're not really that different. Probabilities are not linear, and start at 0.5. In log-odds terms were talking about 2.19 vs -1.04... not really a shocking amount of information. Sure in terms of gambling odds this might be useful but not for making any concrete statements about the world.

To your point:

> I bet most companies would die to turn up conversion rates up even 1 or 2 percent.

This is talking about two very different things. I would also be happy if I had a known 2 percent increasing in my stock performance. We're talking about quantifying uncertainty (though these criticism hold even in a frequentist frame work). I certainly wouldn't "die for" 2% absolute increase in my confidence about something (unless I'm 98% confident already!)

We're talking about our confidence in a hypothesis. At no company I've worked at would I be anywhere near comfortable stating that a 0.9 probability that A > B is enough to call an A/B test on. I'm going to be very skeptical of any analysis I'm doing if mode coefficients have anywhere near that probability of a sign error.

If you have no other information, then sure if I am 90% confident that one choice is stronger than other I'll go with the limited information that I have. However anyone with a statistics background would view either of these numbers as being very uncertain. That's arguably why these numbers are so good for betting, you want there to be a lot of uncertainty when gambling or it's no fun.

But as far as statistics is considered both 0.9 and 0.26 are essentially uninteresting numbers, that lead to very low confidence in the conclusion being drawn.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#294

Post by lostmsu »

paxys wrote: post #114
jcadam wrote: post #13

Zelensky is starting to sound like he's open to considering some of Russia's stated terms (no pursuit of NATO membership and cession of eastern Ukraine). That doesn't sound like things are going especially well for him.

Cession of eastern territories and Crimea is probably not going to be too difficult a decision, considering Russia controls them anyways and did so even before the war. Demilitarization of Ukraine, NATO membership and relations with the west in general are all going to be very contentious, and it's hard to see a solution here that both sides will accept.

I bet you he is only prepared to give up Crimea. That's the only area where referendum to join Russia could actually be fair and passed.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#295

Post by pmoriarty »

mullen wrote: post #96
wackro wrote: post #65

Earlier quoted context omitted.

>call Putin's nuclear bluff. And what do you think would be the outcome?

Nothing. A Nuclear Bluff is just that, a Bluff. Putin can't win in a nuclear war and if he uses nukes anywhere, he is done for. No one is going to have anything to do with him or Russia on any level after that.

"A Nuclear Bluff is just that, a Bluff. Putin can't win in a nuclear war and if he uses nukes anywhere, he is done for."

We don't know what he would do. A desperate dictator might just use nukes.

There's also the greatly increased risk of accidentally starting a nuclear war -- something we've come close to multiple times, even in peace time. In war time that's just short of WW3, the risk is increased exponentially.

No sane person wants to gamble with the literal fate of the world. Yet both sides are doing this by getting this close to WW3... that doesn't lend confidence to the possibility of staying out of a nuclear war if NATO "calls Putin's nuclear bluff".

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#296

Post by maury91 »

cbdumas wrote: post #238
s1artibartfast wrote: post #231

Earlier quoted context omitted.

I assume you were talking about me. I think blame is a stupid and fruitless framing for discussion. I do think the Eastward advance of NATO was a strategic and moral mistake.

How do you think you would feel about that if you lived in Lithuania?

I live in Lithuania and I'm glad we are part of Nato or now we will be the ones invaded by Russia

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#297

Post by s1artibartfast »

testrun wrote: post #289
s1artibartfast wrote: post #255

Earlier quoted context omitted.

If I lived in Lithuania I would probably feel very scared of Russia, and not care much about what is good for NATO, the US, and the world. If I thought I would be accepted without repercussions, I would probably want that.

Lithuanians disagrees. This is why they applied for EU and NATO membership as quickly as they can. The Baltic states also was the first to leave the USSR. They always wanted to be free.

Reread my response.

I said if I was Lithuanian I would probably apply. Sounds like they think the way I said they do.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#298

Post by tim333 »

It's certainly been going better for the Ukrainians than I thought at the start - I thought the tanks would roll straight into Kyiv. Now they seem stuck and Russia lost 30 helicopters last night and today Poland agreed to give all their MIG29s. I guess this isn't really just Russia vs Ukraine but to an extent Russia vs the free world.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#299

Post by jcadam »

pmoriarty wrote: post #285
Steltek wrote: post #268

Earlier quoted context omitted.

In terms of human lives, it's the right deal. Past that, it's the worst possible scenario: Ukraine's territory is not rightfully and completely restored while Russia's economy is nearly irrecoverable, leading it to become nothing more than a resource vassal state to China.

This war has dire global economic consequences. China's not going to get away unscathed either, as demands for its own goods will decrease as the rest of the world's economies get hit by an enormous increase in energy prices and decreased profits as companies around the world stop doing business with Russia. The world is connected like never before and there are no unipolar winners or losers any more (if there ever w…

This could very well lead to the end of the global economy as it exists now, and the formation of a few large, insular economic blocs only doing business amongst themselves.

Re: Prediction that Kyiv falls to Russian forces by April 2022 goes from 90% to 26%

#300

Post by axiosgunnar »

kspacewalk2 wrote: post #168
dekhn wrote: post #121

Earlier quoted context omitted.

Eventually the question becomes "if these guest fighters are successful, at what point do they consider going north through the contested border and attacking Russian (or, more likely, Belurussian) cities?"

As a Ukrainian whose relatives are being bombed and whose friends are currently enlisted - I sure hope they won't consider that at all. The only actions performed outside the internationally recognized borders of Ukraine should be in aid of stopping the invasion of Ukraine. Bombing troops and launch sites - yes. Attacking Russian servicemen wherever they can be reached - yes. Occupying even an inch of Russia - hell n…

> and it won't end ASAP if Russians suddenly have a non-absurd reason to defend their land.

This is an important point. This is why we should keep reiterating to the Russian people that the sanctions are not to spite them and will be revoked immediately if Putin is gone.

Post reply on HN