closes Jan 1, 2024
Will a nuclear weapon be launched in combat by the end of 2023?
4%
chance

Market resolves to YES in the event that a nuclear weapon is confirmed to be launched or detonated for combat, non-testing purposes before Jan 1, 2024, anywhere in the world.

Sep 29, 6:02pm: "launched" was changed to "launched or detonated" - see this comment for rationale and how to be refunded if you feel misled.

Sort by:
SneakySly avatar
SneakySlyis predicting NO at 12%

@IsaacKing What a bizarre tweet!

jack avatar
Jackis predicting NO at 11%

Interesting, there's some specific predictions in there that I disagree with. I made a market for one:

Vs their prediction of the NATO response to a Russian nuke in Ukraine:

My estimate is quite high (80%) that NATO's response will be forceful enough to include a non-nuclear military strike against Russia, because key NATO leaders have already made strongly worded statements to this effect.

And:

Vs:

My most likely (70%) scenario after that is Russian counterstrikes followed by rapid escalation via retaliatory actions from both sides, culminating in execution of the all-out nuclear war plans that both sides have spent decades preparing.

AndyMartin avatar
Andy Martinis predicting YES at 11%

There's a "10x Amplified Odds" derivative market here: https://manifold.markets/AndyMartin/amplified-odds-10x-will-a-nuclear-w

SG avatar
S Gis predicting YES at 12%

Compare with

AndyMartin avatar
Andy Martinbought Ṁ100 of YES

re: https://www.thetimes.co.uk/article/putin-orders-nuclear-military-train-to-ukraine-front-line-tswzv2v50

> President Putin is set to demonstrate his willingness to use weapons of mass destruction with a nuclear test on Ukraine’s borders, Nato is believed to have warned its members.

Open to being convinced to do something else on any of these, but for resolution clarity, my current thinking is:

  • this will resolve YES if a "test" like what's mentioned above happens anywhere within the pre-2014 boundaries of Ukraine

  • a test that detonates within the testing country's own borders will not trigger a positive resolution, even if harmful levels radiation are detected elsewhere

  • a detonation in international waters will trigger a positive resolution if it's judged (by me) to be more than a "test" -- in general "Demonstration detonations" designed to intimidate would count as more than a test.

MartinRandall avatar
Martin Randallis predicting NO at 11%

@AndyMartin Maybe you could make it more concrete by substituting NATO judgment about whether it is a test or combat?

In my opinion, intimidation is not combat. The US nuclear arsenal is very intimidating just sitting there casually threatening the end of the world. Similarly China's attempt to intimidate Taiwan this year would not count for me as combat.

MartinRandall avatar
Martin Randallis predicting NO at 11%

I think the relevant definition of combat is "fighting between armed forces". That could plausibly include "strategic" targets like cities full of innocent civilians but not running a nuclear test in international waters with no military or strategic targets affected.

AndyMartin avatar
Andy Martinis predicting YES at 11%

@MartinRandall

In my opinion, intimidation is not combat. The US nuclear arsenal is very intimidating just sitting there casually threatening the end of the world. Similarly China's attempt to intimidate Taiwan this year would not count for me as combat.

I think I totally agree on ^, but the US/China doesn't perform "intimidation detonations" on other countries' land or in international waters - this is the only thing covered by the points above.

MartinRandall avatar
Martin Randallis predicting NO at 11%

@AndyMartin Topical

https://www.cnn.com/2022/10/03/asia/north-korea-missile-suspected-intl/index.html

I don't think these non nuclear missiles were launched "in combat".

MartinRandall avatar
Martin Randallis predicting NO at 11%

@AndyMartin Also, I think that the USA nuclear tests on Bikini Atoll were done on another country's land.

AndyMartin avatar
Andy Martinis predicting YES at 11%

@MartinRandall that's a great point, but:

  1. NK and Japan are not in the middle of an active war

  1. The missile was fired without a nuclear payload - seems more "a missile which could hold a nuclear weapon was launched" than "a nuclear weapon was launched"

MartinRandall avatar
Martin Randallis predicting NO at 11%

@AndyMartin My point is that they were not launched "in combat".

(also NK and SK are officially at war)

It turns out that there US blew up several bits of international waters and space during the cold war, for testing and intimidation. Nobody called these combat uses at the time.

https://en.m.wikipedia.org/wiki/List_of_nuclear_weapons_tests

AndyMartin avatar
Andy Martinis predicting YES at 11%

@MartinRandall

@AndyMartin My point is that they were not launched "in combat".

Thanks - misread that

@AndyMartin Also, I think that the USA nuclear tests on Bikini Atoll were done on another country's land.

...
It turns out that there US blew up several bits of international waters and space during the cold war, for testing and intimidation. Nobody called these combat uses at the time.

https://en.m.wikipedia.org/wiki/List_of_nuclear_weapons_tests

Maybe it would be better to draw the line at "an area in immediate proximity to the combat" than covering all "international waters"/"another country's land"?

Treldman avatar
Treldmanis predicting YES at 11%
MartinRandall avatar
Martin Randallis predicting NO at 12%

@AndyMartin I think it would be best to not change the resolution criteria for a second time in a market where you own shares.

An area in proximity to combat is not in combat. An area in proximity to a house is not in a house.

AndyMartin avatar
Andy Martinis predicting YES at 12%

An area in proximity to combat is not in combat. An area in proximity to a house is not in a house.

I don't agree with applying this analogy here. An action taken in proximity to combat can directly affect the combat itself, and this is the bar I had in mind when creating the question.

In terms of the house metaphor: it is possible to perform an action in proximity to the house which can make the house itself unlivable/unsafe/inaccessible/etc.

@AndyMartin I think it would be best to not change the resolution criteria for a second time in a market where you own shares.

I'd rather hand off resolution than appear to have a conflict on interest or go back and forth on this more - will ask in the discord if anyone from MM is willing to take the market over.

AndyMartin avatar
Andy Martinis predicting YES at 12%

@AndyMartin In discord, @jack agreed to take over the resolution responsibility - thank you!

jack avatar
Jackis predicting NO at 12%

@AndyMartin Chatted on discord about this, I agreed to take over resolution here. tl;dr: I think "in combat" should probably be defined the same way as "non-test" in https://www.metaculus.com/questions/2797/no-non-test-nuclear-detonations-by-2024-01-01/.

Andy already said in an earlier comment that they'd use Metaculus's criteria on when resolving this barring any objections. The main difference was around "launched or detonated", and that has already been clarified well thanks to your earlier discussion. I think it makes sense to treat "in combat" and "non-test" as equivalent unless we notice that they carry substantially different meanings?

Unfortunately, I was not able to find a clear definition on Metaculus either, but if a nuclear detonation does happen then the Metaculus admins will have to decide whether it counts as test or non-test or ambiguous, and we can use the same judgement here. I also asked on Metaculus to see if they could clarify their thinking on the definition.

Here are some scenarios I was thinking about:

  • If Russia detonates a nuclear weapon within their own borders as a demonstration detonation, with the explicit intent of intimidation, is that a test or no? (Probably a test, non-combat in my mind)

  • What if it's detonated in international waters, causing no damage (and still intended as a demonstration/intimidation)? (Still seems like test, non-combat)

  • What if it's detonated on a small Ukranian island with a few buildings but no people? (This seems like pretty clearly non-test, yes combat to me)

  • What if the detonation occurs in international waters or within Russia's borders but the radioactive fallout affects Ukraine? (More borderline, but I'd lean towards non-test, yes combat)

Thoughts?

Disclaimer: I do have a position in this market too, but I also trade in almost all of my markets and am pretty used to that.

MartinRandall avatar
Martin Randallis predicting NO at 12%

@jack For the last case I'd call that a test unless the fallout is large enough to deny access to an area or cause immediate casualties. For comparison various nuclear tests in the past have caused increased cancer deaths without being considered combat.

I think deferring to Metaculus is a great approach, hopefully they'll answer soon.

jack avatar
Jackis predicting NO at 10%

@MartinRandall Yeah I roughly agree, I think if the nuclear detonation has large immediate, direct effects (e.g. casualties or forcing nontrivial evacuations) that seems like it should count as non-test, but otherwise if it's just a small amount of fallout drifting over Ukraine (and over the rest of Europe for that matter) then I'd likely call that a test.

LivInTheLookingGlass avatar
Oliviais predicting NO at 10%

May I make an amplified odds version of this market?

AndyMartin avatar
Andy Martinis predicting YES at 10%

@LivInTheLookingGlass if you're asking me, please do!

SG avatar
S Gbought Ṁ28 of YES
AndyMartin avatar
Andy Martinis predicting YES at 8%
AnselFreniere avatar
Ansel Freniereis predicting YES at 10%

I see that this was already posted below.

jack avatar
Jackis predicting NO at 10%

Readers should note that the market structure of prediction markets (both in general and Manifold in particular) means they tend to be very inaccurate at forecasting events with probabilities near 0% and 100%. E.g. the market price being 5% is pretty consistent with the predictors on this market believing the probability is anything from 0% to 5%, depending on a wide variety of factors. There are a couple reasons for this, one is that it takes more and more M$ to correct mispricings close to the extremes, and the profit for putting in all that M$ gets smaller and smaller.

For evidence that you should not trust a number like 5%, see e.g.

This was posted on https://marginalrevolution.com/marginalrevolution/2022/09/will-a-nuclear-weapon-be-launched-in-combat-by-the-end-of-2023.html without any commentary on that issue. It is worth noting that Metaculus's forecast isn't too far different, but take this as a general cautionary principle about interpreting prediction market data.

ManifoldMarkets avatar
Manifold Markets

@jack This!

MartinRandall avatar
Martin Randallis predicting NO at 6%

@jack Both linked markets are now at 1%. Thank you for pointing out the handful of mana lying on the floor. I picked it up. This market is more liquid. At time of writing, someone bidding down to 3% could make m2338 (USD $23.38) if the market resolves NO. I'm not doing that myself because I think it may resolve YES.

I think the Marginal Revolution take that the risk of nuclear combat was "worryingly high" and now is "very, very bad" is spot on. I'd much prefer to see this market at 3%, and I do not find your explanation for 6% market odds at all reassuring.

jack avatar
Jackis predicting NO at 10%

@MartinRandall I actually believe the probability is in the right ballpark currently. But if the facts on the ground changed to make nuclear risk drop to say 1% or 0.1%, I don't think the market would accurately reflect that - maybe it would go to 3% or something, which is a pretty big difference. So my general claim is that even if the market seems roughly accurate now, I am skeptical of its robustness in general and in particular I think the market price is a very biased estimator near 0% (biased away from 0%).

https://manifold.markets/LarsDoucet/will-china-invade-taiwan-in-2022 is another pretty liquid market that I think is pretty wrong (and I just put in a bunch of M$ to correct it some, but I think it's still wrong).

IsaacKing avatar
Isaac King

Related market:

Gigacasting avatar
Gigacastingis predicting NO at 8%
Gigacasting avatar
Gigacastingis predicting NO at 8%

Based on the Taiwan markets and the Trump markets—I don’t think these are “serious predictions”

Would take this at 100:1

MP avatar
MPis predicting NO at 8%

@Gigacasting I think that most smart people here have markets with better expected values than betting NO @ 6%.

MP avatar
MPis predicting NO at 8%

@MP The only way that MM would solve this is by implementing leverage so you could take this bet 100:1

AndyMartin avatar
Andy Martinbought Ṁ10 of YES

@MP The only way that MM would solve this is by implementing leverage so you could take this bet 100:1

there's actually been a lot of experimentation/work in this area - https://news.manifold.markets/p/above-the-fold-anyone-want-a-loan

Without loans, there was very little incentive for users to bet on long-term markets. This prevented an array of interesting questions from receiving engagement and accurate forecasting. We hope as activity in long-term markets increases there will be more room for profit and more accurate probabilities generated.

betting on https://manifold.markets/EricJang/will-a-nuclear-weapon-be-launched-i (ending in a few months, currently trading at 6%) might be another option

AndyMartin avatar
Andy Martinis predicting YES at 9%
AndyMartin avatar
Andy Martinis predicting YES at 10%

I think Kremlin may view nuclear strike on Ukraine (with an American retaliatory strike) as a rational move. It may not make much sense in the context of foreign policy, but it does in the context of domestic policy. Meanwhile foreign policy is just domestic policy by other means

My argument is based on three premises:

1. Foreign policy serves domestic policy goals

2. Keeping power is the top priority of domestic policy

3. Kremlin is looking for a way out of the conflict

Launching a nuclear strike and getting a retaliatory one may be seen as a way out

If Putin is looking for a way out, that probably means he is looking for a way out that would allow him to keep the supreme political power. Which may be incompatible with suffering a humiliating military defeat from a supposedly inferior force. Like Japan in 1905 and Ukraine now

Many argued that Putin should be allowed to "save face". The thing is: it is nearly impossible for him to save face at this point. Inability to crush Ukraine is already a loss of face, no matter how you frame it. For a simple reason that Ukraine had been considered an inferior

The very idea that Ukraine can stand its ground against Russia would have been considered totally insane at the start of this year, both in Russia and outside of it. The fact that it does means that Putin has already suffered a major loss of face, and will continue suffering it

To keep power, you need to save face. And how can you save face? One obvious solution is: try to engage with the U.S. Make a nuclear strike that would inflict an American retaliatory strike. If you and your power survives that, that gonna be total win. Absolute victory

Russian public opinion considers Ukraine as absolutely inferior. That is why standards for not losing face in this conflict are pretty high. Since Ukraine is so ridiculous, inability to crush it is humiliating, no matter how you frame it. You can't defeat it, you lose your face

But Russian public opinion does not view the US as inferior, it's the other way around. That is why standards for not losing face in a conflict with the US are so much lower. I would say that inability of the US to crush Russia would be seen as American loss and Putin's win

Compare two scenarios:

1. Russia engages Ukraine. Then Russia not crushing Ukraine is Ukrainian victory. High bar

2. US engages Russia. Then the US not crushing Russia is Russian victory. (Seemingly) low bar

He absolutely may choose 2 as the bar for not losing face is lower

I would even say that provoking an American retaliatory strike may boost the regime. The US tried to defeat us, used the deadly force, but we still exist - that's the story of honour, heroism and stoicism. For the Putin's target audience I mean. That's a great and powerful myth

Meanwhile, mobilisation may stabilise the regime by simply getting rid of young males that could otherwise create problems. For example, if they were recruited by anti-regime forces, should the supreme power weaken. The less young males in the country, the stronger is Kremlin

Engaging with the US directly may sound suicidal foreign policy wise. At the same time it may sound totally rational domestic policy wise. Simply because the bar for saving your face (=keeping power) in this conflict would be much lower. They didn't destroy you = you win. The end

PS For example, the US destroying the Black Sea Navy as a response for a Russian nuclear strike absolutely may count as Putin's victory. It is very, very much easier to frame this scenario as honourable, than let's say an obvious military defeat from Ukraine

PPS Withdrawing before the US is very much preferable than withdrawing before Ukraine. In the first case, you can frame it correctly, save your face and prepare for a round 2. If you withdraw before Ukraine though, there may be no round 2

https://twitter.com/kamilkazani/status/1575510055369293824?s=12&t=arChNdyjJGcUOCZrGdbfrg

AndyMartin avatar
Andy Martinbought Ṁ0 of NO

unless there are any objections, I'll plan to use the resolution criteria from https://www.metaculus.com/questions/2797/no-non-test-nuclear-detonations-by-2024-01-01/ when resolving this

MartinRandall avatar
Martin Randallis predicting NO at 9%

@AndyMartin I think there are some detonations that would not have been "launched", eg a cyber attack or an accident that causes a missile to blow up on the ground.

Conversely a missile could be launched but not detonated, eg because it was sabotaged or intercepted or recalled or didn't work.

AndyMartin avatar
Andy Martinis predicting YES at 9%

@MartinRandall that's a good point. Another corner case could be around delivery systems where the payload is not "launched".

Will a nuclear weapon be launched in combat by the end of 2023?

Market resolves to YES in the event that a nuclear weapon is confirmed to be launched for combat, non-testing purposes before Jan 1, 2024, anywhere in the world.

I had originally taken the title/description from https://manifold.markets/EricJang/will-a-nuclear-weapon-be-launched-i , and I think an "as literal as possible" interpretation would resolve like:

  • deliberate missile-based nuclear attack - yes

  • missile-based nuclear attack for "demonstration detonations" - yes

  • a missile that's launched but not detonated (sabotaged, interception, didn't work, etc) - yes

  • deliberate detonation on the ground - no

  • nuclear bomb dropped from a plane - no

  • a cyber attack or accident that causes a missile to detonate on the ground, related to a war/conflict context - no

  • a cyber attack or conflict-related accident that causes a missile to launch - yes

  • an accident that causes a missile to launch or detonate, unrelated to any war/conflict context - no

I think the average person would joined would have viewed "launch" as something closer to "triggered" or "detonated" and would end up being surprised by some of these.

The best thing I can think to do is to keep the original text but switch over to "launched or detonated":

Market resolves to YES in the event that a nuclear weapon is confirmed to be launched or detonated for combat, non-testing purposes before Jan 1, 2024, anywhere in the world.

  • deliberate missile-based nuclear attack - yes

  • missile-based nuclear attack for "demonstration detonations" - yes

  • a missile that's launched but not detonated (sabotaged, interception, didn't work, etc) - yes

  • deliberate detonation on the ground - yes

  • nuclear bomb dropped from a plane - yes

  • a cyber attack or accident that causes a missile to detonate on the ground, related to a war/conflict context - yes

  • a cyber attack or conflict-related accident that causes a missile to launch - yes

  • an accident that causes a missile to launch or detonate, unrelated to any war/conflict context - no

I'm going to go ahead and update the description, but if anyone feels misled, please sell your shares and reply to this comment - I'll tip you back any $M losses.

MartinRandall avatar
Martin Randallis predicting NO at 9%

@AndyMartin Does anyone have pure gravity nuclear bombs any more? I thought plane launched nukes were all missile based so they could be launched from further away from the target.

MartinRandall avatar
Martin Randallis predicting NO at 9%

@AndyMartin I do feel slightly misled, this update makes YES more likely, right? If I sell all my shares now at market price I have a m150 loss. But I think that's a bad price, so I won't sell. Feel free to tip an amount you think is fair.

AndyMartin avatar
Andy Martinis predicting YES at 9%

@MartinRandall I'll push it up to M$11 to fill your M$1000 limit order and then sell it back down - does that seem fair?

MartinRandall avatar
Martin Randallbought Ṁ1,000 of NO

@AndyMartin blinks. Sure, wasn't what I was expecting.

AndyMartin avatar
Andy Martinbought Ṁ160 of YES

I was thinking that might be preferable since it'd show up as positive profit, not just mana

MartinRandall avatar
Martin Randallbought Ṁ0 of NO

@AndyMartin Just don't launch any nukes to increase your profit number.

AndrewHartman avatar
Andrew Hartman

@MartinRandall Uh, iirc some of the nuclear bunker busters are still gravity-based, albeit designed to be used from a substantial height. I am less certain that they're still in the US nuclear inventory, though - there weren't that many of them in the first place and the US has been winding down a lot of the non-strategic nuclear weapons over the last two decades, but afaik Bush (second one) had pressed for an updated design of them for some reason, so I thought there were still a few floating around in inventory.

MartinRandall avatar
Martin Randallis predicting NO at 7%
IsaacKing avatar
Isaac King

Out of curiosity I made a market for failed detonations. Seems people are pretty pessimistic (or optimistic from our perspective I guess) about the quality of Russia's arsenal.

MartinRandall avatar
Martin Randallbought Ṁ1,000 of NO

https://www.metaculus.com/questions/7404/nuclear-detonation-fatality-by-2024/ says 5%, unless folks are placing a lot of weight on nonlethal nuclear weapon use.