Nuclear war probabilities are useless - counter to the nuclear gambler’s ruin | Hacker News Reader

Nuclear war probabilities are useless - counter to the nuclear gambler’s ruin

unfashionable.substack.com

79 pointsSvenSchnieders121 comments

In safety we try to prevent rare events (such as death) and these are rare enough that “past deaths” are not a useful way to predict the likelihood of a death.

A crucial metric that occurs more more often and does help predict deaths is “HPI” - high potential incident. A worker eight metres in the air dropped a spanner. It thudded on the ground, harmlessly. No one was hurt. No problem, right? That’s a HPI. A workplace that has a death will have had many HPIs. The dice-roll is not “it’s a new day, we roll the dice”. When a HPI occurs, that’s when we roll the dice.

Or in nuclear war estimates: a HPNWI occurs (high potential nuclear war incident). The Cuban missile crisis springs to mind as a particularly big example. But many many smaller HPNWI’s have occurred. Border skirmishes between nuclear powers. Sabre rattling by nuclear powers. Elect a president who uses threats and brinkmanship. Do this year in, year out. The dice is rolled not because time passes but because risky decisions are made and risky actions are taken, over and over and over and over and over and over, year in, year out.

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

Ah so that Lemmino video is just a rundown of this wikipedia page hah.

I am not sure if I understand what the point is you’re trying to make

I took it to be: the original author's point was that nuclear attacks are so rare and our sample size so small that we can't use history to predict their occurrence in the future. This commenter is saying the typical way to handle situations like this is to look for more-common events that are proxies for the less-common ones in some predictable way.

To give a different example of my own: on an intersection-by-intersection basis, pedestrians dying from getting hit by a car is relatively rare, so if it's perceived that a particular intersection is unsafe and the design of the intersection is modified with a goal of reducing the chance of death, it can be difficult to know if that intervention worked. Supposing last year there was one death there and this year there weren't any; was that because of the design change, or just chance? But we do know that, in aggregate, some relatively constant percentage of collisions will result in deaths, and collisions overall are much more common than deadly ones, so even if the policy goal is to reduce deaths, measuring all collisions is a very useful, less-noisy proxy for determining if an intersection has been made less deadly.

So: original author is arguing that we can't meaningfully reason about these things because they're rare, and this commenter is saying "yes we can, if we find a more-common, less-noisy thing that's a good proxy for the rarer thing," and proposes that nuclear close calls are such a proxy.

Thank you, yes. As my buddy Jeeves would say “rem acu tetigisti”

Thank you, that was helpful

The author was saying we don’t have enough data to make a prediction. OC is saying we do have enough data to make reasonable assumptions if you look at high risk situations (which can reasonably be assumed to be a prerequisite to an actual disaster like death or nuclear war).

For example, take two people who have never been in a car accident. Driver A gets a speeding ticket a year and a DUI every 3 years. Driver B has never had a violation. It’s reasonable to assume that Driver A is at higher risk of a future accident. We don’t have to only look at the 0 accidents data point, we can look at factors that are typically common in accidents like speed and alcohol.

Amazing way to think about HPIs, thanks for the introduction of this concept.

Thanks randomjoe2. I do want to be very clear that it’s not a concept I’ve created, and also I have probably mis-stated parts of it - it’s something that I’ve heard about many many times from working for years in the oil and gas industry.

This runs counter to the black swan concept popularized by Taleb. In that they are by nature unpredictable, what one can predict is that there will be a next one.

I like this framework but I'd appreciate hearing more about it.

Ok I was waiting for someone to say this.

Taleb’s ideas on this would be very interesting. But is a nuclear war a black swan? Think about it. (Hint: have we ever observed a swan of this color before? Have we ever observed countries engaged in total war? Have we ever observed countries deploying nuclear weapons? Yes and yes)

Now a stronger form of what I think you’re saying is that, even if modelled all of the ways we think a nuclear war could break out — there are ways we would never think of. And the consequences of failing to account for those would be catastrophic, unlimited downside.

I agree with that whole heartedly. To that I would say — having nuclear weapons, allowing them to exist, exposes us to an unlimited downside. No matter how remote the chance of nuclear war seems to be, the only course we can take is complete nuclear disarmament. Global moratorium on nuclear weapons. End of story. We can’t foresee why/how a massive war would unfold. But we can say that it has unlimited downside. Immediate international cooperation on total banning of nuclear weapons is the only way forward.

Do you think it’s possible to model a mad man’s mind with this?

Another issue with this logic is that the probabilities are not independent.

Saying that you have probability x of nuclear war every year and so probability (1-x)^n of no nuclear war in n years assumes independence. In fact, if you don't have nuclear war on year 1, I would assume that the reasons for that lead to inferences and correlations that would affect the probability of nuclear war on year 2, possibly making it lower.

No, that’s not quite right. Measuring “did we have a nuclear war?” is not the input you’d ever want to use there, it’s only one binary digit of information.

Instead take info like “did we have events/incidents with a high potential of nuclear war?”

If three “Cuban missile crisis” level events happened in year X, then the probability of nuclear war in year x+1 is higher than it was in other years — even though nuclear war didn’t occur in either year. If national leaders make no statements threatening war in year A, then it’s probably safer than year B where there were 100 such threats.

I might not have said this clearly enough - I'm not saying that conditioning on "did we have a nuclear war" will change the probability to be less. I'm saying that the event "did we have a nuclear war in year n" is not independent from "did we have a nuclear war in year n+1" "n+2", ... because there are shared factors influencing them. Therefore the model of independent probabilities multiplying year over year and seeming to imply that nuclear war is inevitable because (1-p)^n goes to 0 as n grows doesn't make sense mathematically speaking.

Ok. I think the question that we want to ask from the model is more like:

“What is (one minus (What is the chance we have had 0 nuclear wars by year “now plus 70”?)”

So if there is a nuclear war in year n+3 (for example) it’s effect on year n+4 is irrelevant as the answer to the question is already “0% chance of no nuclear wars by year 70”.

So the angle you’re initially coming from is not quite relevant.

We can then turn that a bit though and continue your point — we can rephrase your claim to be like this, for example:

“if we have had 0 nuclear wars in the next 69 years, then surely that would lower the probability of a war in year 70.”

And to that I’m saying — actually no. Or rather: not necessarily. Or more accurately: not much!

The info of “we’ve had no nuclear wars in 69 years” is not what the model would care about. More likely it would care about, running the model a bunch of times, in the cases where there no wars in the first 69 years, how many times is there a war in the 70th year? And “brink of war” or “HPI” scenarios would be a much better indicator than the simple absense of nuclear war in those first 69 years. Maybe better to said “there is a lot more information in the question of ‘how many times have we been on the brink of war’ than there is in the answer ‘did we have a nuclear war yet? No’. It’s more about that “how much info is there here?”

I'll say it again: I'm not saying that we can use past lack of nuclear war to predict future lack of nuclear war. All I'm saying is that each year's probability is not independent, meaning the model of a probability of no nuclear war which geometrically goes to 0 is not accurate.

While it is difficult/impossible to accurately quantify the probability of war, and that the ground truth is far more complex than these simple models, there still has been a long list of historical incidents in which the outcome was decided by sometimes random and even independent factors outside of anyone's control and makes the gambler's ruin argument compelling:

https://en.wikipedia.org/wiki/List_of_nuclear_close_calls

https://en.wikipedia.org/wiki/List_of_military_nuclear_accid...

There's two particular incidents that are great examples of this. The '83 false alarm [1], and the 1980 missile silo explosion [2].

In both cases, they were razor-thin near misses. If either accident happened even once more, we absolutely would have either a radiological disaster or accidental nuclear war.

It's this sort of thing that makes me think maybe we shouldn't have any nuclear technology at all, the risks are just too high. Climate change is bad, but not as bad as burning our entire planet to the ground in under an hour, then poisoning it forever.

[1] https://en.wikipedia.org/wiki/1983_Soviet_nuclear_false_alar...

[2] https://en.wikipedia.org/wiki/1980_Damascus_Titan_missile_ex...

> If either accident happened even once more, we absolutely would have either a radiological disaster or accidental nuclear war.

The Soviet incident sure, but the missile explosion was pretty much as bad as it could have been. You could explode the rocket fuel in every nuclear missile on the planet and you wouldn't get a single nuclear explosion, just scattered chunks of fresh nuclear fuel over a small area, which isn't really a huge deal compared to, for example, an oil pipeline leak.

I think nuclear energy is mostly orthogonal to nuclear weapons. We should pursue more nuclear energy and seek to eliminate nuclear weapons.

I wonder what if the Chelyabinsk meteor had fallen now, instead of some years ago. Could it have been mistaken for a nuke and triggered a "counterstrike"?

Given the universe's tasteless sense of comedic timing, I wouldn't be too surprised if a Tunguska-scale meteor lands right on DC, Moscow or Beijing at exactly the wrong time during an international crisis.

Still, in principle, the Russians and Americans, and I believe the Chinese, have satellites that detect the distinctive radiation burst from a nuclear explosion. (There's a detector on each GPS satellite.) So it /should/ fail to pass some of the automatic and manual thresholds before we're off to press the button. But in a tense situation, ambiguous data... one does worry a bit. There's been ambiguity with the detectors before. (Was there, or wasn't there, a nuclear explosion in the south Indian ocean in 1979? https://en.m.wikipedia.org/wiki/Vela_incident )

Have a read of Eric Schlosser's book, Command and Control

https://www.amazon.com/Command-Control-Damascus-Accident-Ill...

The ways we could end up in a nuclear war are frightening to the point of madness.

> The ways we could end up in a nuclear war are frightening to the point of madness.

Haven't read the book, but I watched Dr. Strangelove when I was young(er).

What made the movie great to me was that it made me realize that, although there is a nuclear "safety net", it was designed and controlled by humans — flawed humans who don't always behave rationally and who may not act in the best interests of humanity.

It is not only possible for humans not behaving rationally, it is essential!

... and yet we haven't.

Every day since WW2, things could have gone wrong.

Every day, for 76.56 years. 918.79 months. 27965 days.

I think we'll be fine. And if I'm wrong, you probably won't be able to tell me! :)

Edit: not sure (and don't really care) why this got downvoted, but I meant it as a morale boosting thing, not as a probability calculation. Things could have gone horribly wrong so many times, but they haven't so far. That personally gives me hope for the future.

I hope you're right, but this feels like an instance of the hot hand fallacy. In a historical sense, nuclear weapons haven't existed for very long. 76.56 years without a nuclear attack is less than the life expectancy of one US adult. It's too soon to tell.

I would say that probability is 100%, having a corrupted system, with corrupted people and madmen, on all sides, it is not question of "will" but rather "when"...

The system being corrupted makes it less likely to have nuclear wars, not more? Idealists are willing to blow up everything for their ideals, people who are only in it for themselves have nothing to gain by destroying the world.

Idealism, nationalism, fascism... are forms of corruption, but in this case there is inherited problem with aristocratic corruption, the one that allows people with lot of money to command most deadliest weapons on the planet. When you think about nowadays leaders and there structures are not very different from mafia bosses behaviors, and I am not talking about Russia only, each super power has the same problem.

John Kay wrote a good book about this called Radical Uncertainty. It's annoying that there is an increasing tendency to frame what is a purely narrative statement in quantitative terms.

As the author points out war is not historically predetermined, at least in any way we could convincingly figure out, and pulling numbers out of thin air only serves the purpose of trying to talk about the future with less uncertainty than is appropriate.

Talking about events in terms of probabilities should be reserved for events that are not dominated by 'unknown unknowns'. You can attach a probability to a volcano outbreak, a transaction in an ordinary functioning market, a game of blackjack. That is to say 'small worlds' where outcomes are known, rules well defined and inputs quantified. You can't predict the overthrow a government in 5 years in numerical terms. That doesn't mean you can't talk about it. It's okay to say something is unlikely or very likely, but it makes more sense to explicitly talk about these things in qualitative terms rather than trying to be 'mathy' to make a prediction appear more objective than it is.

A useful metric - how often does a national leader willing to start a nuclear war come to power in a state with nuclear weapons? That's measureable from historical data.

Which data would that be? We don't know if anyone is willing to use nuclear weapons against a nuclear power because no one has ever done it.

How? We had no nuclear war yet. How can you know that leader is willing to?

It's not measureable. First of you need to account for many having the same weapons (MAD) and you also need to account for the knowledge of the impact of these weapons. The knowledge does change the probability for someone willing and able to use these weapons to come into power (as well as the whole structure around the usage).

You're missing half the equation. 'Every'[0] national leader is willing to start a nuclear war given some particular set of circumstances. There's always some line that they will not allow their enemies to cross.

The difference between leaders is what those circumstances are. It's in where that line is. [1]

Public communications on this subject have to be taken with a pound of salt. It's in every nuclear nation's interests to claim that line is further out than it really is... As long as they believe that broadcasting the claim itself won't start nuclear war.

It's also in every nuclear nation's interests to be very mindful about where their adversary's line is.

[0] Okay, there might be a closet pacifist somewhere that won't use them in any situation, but if there is, they certainly aren't broadcasting that.

[1] So far, we're batting a thousand for 'No national leader whose last name was not Truman started a nuclear war in a situation where his state was not fighting an existential war'.

This statistic, unfortunately, is not particularly helpful.

If they're not disarming they're either willing to use them or bluffing.

Being willing to use them is not the same as being willing to use them first.

Only one national leader -- Harry Truman -- has ever allowed nuclear weapons to be used in combat, in 1945, the same year the weapon became available. This was 77 years ago. He was the 33rd president, and we are on the 46th. By this measure, based only on the US, we get such a leader once every 77 years or once every 13 leaders. These numbers will only get lower once you start to add in other nuclear weapons states.

That said, I don't think any of this math is pointing us toward any kind of answer about how likely any person alive right now is to experience nuclear war, which is really what we're after. At the very most, every day that nobody uses one extends the lower bound of how often they are likely to be used (77 years and counting), but this is almost tautological. Also, maybe it follows some kind of power law and the farther away from their invention, the less likely we are to use them? Or maybe the opposite? Or maybe we oscillate between the two in some sinusoidal pattern? In any case, we can't know. History is not an experimental science.

This is true, but it also ignores that nuclear weapons are used every day in diplomacy. The best use of a gun is to point it at someone, get them to do what you want, and not ever fire it. Sometimes the nukes are just background information, sometimes it is a more direct threat.

If we have N nuclear states, the possibilities of conflicts between them must scale at least as N^2/2, counting the hypothesis of a country using nukes against itself (civil war/secession).

> Put differently, Putin does not roll a dice every morning and launches a nuclear missile if a six shows up. Even he himself cannot meaningfully answer the question “What’s the probability that Russia will launch nuclear missiles 6 months from now” unless his answer is 100%. It depends on many factors, like how NATO reacts and how strong the Ukraine military is, none of which can be captured by probabilities.

There's a guy[0] who spent a career analyzing and writing about these things.

His most recent take is that by refusing to get directly involved in Ukraine, NATO creates an understanding in Putin's mind that it wouldn't get involved in, say, Latvia either. According to Piontkovsky, (to Putin,) there's no essential difference, and the (lack of) NATO membership reason provided by NATO is a lame excuse indicative of a lack of resolve to commit, NATO member or not. So Putin is willing to keep upping the ante until NATO folds and becomes irrelevant.

[0]https://www.researchgate.net/publication/313258664_Putin's_R...

I don't buy this. The NATO message was clear & consistent that they will intervene in case of an attack on a member - there's 0 reason to believe otherwise. In fact, I'd say that current response in Ukraine enforces this (though it's not a NATO member it got plenty of offensive equipment, intelligence, money, support, etc. - basically everything possible short of direct declaration of war / direct military involvement from all NATO members).

Keep in mind that Putin's mythos was blown to pieces, there's no way he can (politically, if not physically) survive this. And he probably knows it. This is the weakest position he can be in - if the west attacked him, he could at least mobilize the Russians to "defend against western aggression" and could even justify the nuclear bombs to his generals; As things stand, his rhetoric and position are very shaky (he's not being defeated by mighty USA, but by the "weak" Ukrainians who "aren't even a nation")

> The NATO message was clear & consistent that they will intervene in case of an attack on a member - there's 0 reason to believe otherwise.

How did Putin take that message? I think that's the only thing that matters at this point.

> Keep in mind that Putin's mythos was blown to pieces, there's no way he can (politically, if not physically) survive this.

_If_ that mythos was blown to pieces, in the eyes of the Russian public. That remains to be seen.

And _if_ the general public of Russia still matters...

...and _if_ a viable alternative leader exists...

> The NATO message was clear & consistent that they will intervene in case of an attack on a member - there's 0 reason to believe otherwise.

Unfortunately that's not true.

The previous US president (who still has significant backing with US Congress and Senate) said this when asked about Article 5:

> I want to keep NATO but I want them to pay. I don’t want to be taken advantage of… We’re protecting countries that most of the people in this room have never even heard of and we end up in world war three… Give me a break[1]

And he later refused speeches that included a commitment to article 5:

> The crucial nuance is that Trump did not say that NATO’s original mission of countering Russian power in Europe is no longer obsolete. Indeed, he has never acknowledged this. The closest he has come is a general statement of strong support with caveats attached for burden sharing and counterterrorism. Trump’s failure to endorse Article 5 is not an oversight. Members of his cabinet have unsuccessfully tried to insert this language into his remarks, including at his meeting with Stoltenberg.[1]

Now eventually, Trump did endorse Article 5[2], but that kind of reluctance would have been very well noted in Moscow.

> Keep in mind that Putin's mythos was blown to pieces, there's no way he can (politically, if not physically) survive this.

After 9 days of war you think there is no way he can survive? Are you not familiar with the origins of Putin's Chechen War?

Claiming that Putin is done because a war he is winning is unpopular in the West is naive. It's certainly possible he is in trouble, but there is a long, long way to go before he falls.

[1] https://www.brookings.edu/blog/order-from-chaos/2017/05/17/t...

[2] https://edition.cnn.com/2017/06/09/politics/trump-commits-to...

[3] https://en.wikipedia.org/wiki/Second_Chechen_War#Bombings_in...

I don’t get the direct connection between response from a non-NATO country being attacked and a NATO member.

If Putin tries to invade Latvia, the default assumption in this analysis is that NATO will just crumble rather than respond?

> I don’t get the direct connection between response from a non-NATO country being attacked and a NATO member.

The idea is that the lack of NATO membership is just a formality. In any other aspect, Ukraine is a much more important geopolitical asset than said Latvia (with all due respect to the Latvians - I'm just imagining the way it plays out in Putin's head. Or at least the way I understand Piontkovsy here).

> If Putin tries to invade Latvia, the default assumption in this analysis is that NATO will just crumble rather than respond?

Yep. Piontkovsky even describes a potential limited nuclear strike scenario as part of the raising of stakes by Putin. AKA "are you willing to die for Narva[0]?"

[0]https://en.wikipedia.org/wiki/Narva

I mean I find these analysis absurd, they seem based on the ideas not just that Putin is irrational but also that he knows nothing about the very things we would assume he knows about, stuff like how treaties and military alliances work.

I could totally understand some guy on the street thinking hey Latvia is nothing, if they won't defend the Ukraine then they won't defend Latvia but why would a guy with supposed knowledge of military history and how alliances can often cause nations who are allied to come in to wars they would otherwise rather keep out of think such a thing.

In fact I would assume his kneejerk opinion would be to believe that the only way to get all these weak nations to actually go to war would be if an attack happened on a NATO nation.

> why would a guy with supposed knowledge of military history and how alliances can often cause nations who are allied to come in to wars

There's no military history of nuclear wars. This could be the difference in Putin's head.

> His most recent take is that by refusing to get directly involved in Ukraine, NATO creates an understanding in Putin's mind that it wouldn't get involved in, say, Latvia either.

That’s a weird take. Ukraine is not a NATO member, NATO can choose to fight for them or can choose to not fight for them. Latvia is a NATO member. If Latvia is attacked and the NATO doesn’t fight for them NATO stops being a thing immediately. The 5th article is a cornerstone of the whole aliance. It is not an optional “and one more thing” but the reason to join the aliance at all, if it becomes shaky that undermines the whole thing.

I wouldn’t dare to guess what goes on in Putin’s mind, but I also don’t see any reason why he would think that.

I think this take is exactly what needs to be considered.

We draw a line, it's NATO. Russia steps right up to the line by attacking Ukraine, and NATO does nothing. Is he going to believe that the line is still there? Why not test it?

In my mind it's perfectly plausible that NATO squirms out if Latvia is attacked. Then the line is "core NATO" or something like that.

At the very least if he doesn't test article 5, he can take every country that isn't in NATO, and each time the question grows about whether NATO is actually a thing.

> Why not test it?

Because it's not a thing where you say "let's test this out today." It takes months or years to get to this point. I'm hoping NATO would respond to a build-up on the Latvian border with a build-up of their own. Making it clear that there is no easy win.

There's no reason to just sit there and wait for the tanks to roll across the line.

> Why not test it?

Be our guest. Test it.

It doesn’t start with “and then Russia attacked Latvia”. They are not the Fire Nation. They need time to line up their logistics which will give time for NATO to detect their moves and amass troops in Latvia. Heck they were only lining up to attack a non-NATO country (Ukraine) and NATO sent extra troops to Latvia.

https://abcnews.go.com/Politics/us-nato-military-moves-send-...

I think, also, that faced with a similar situation with a NATO member as the target, the whole process would be different. NATO isn't going to wait for an invasion before responding, IMO. Pretty quickly after there is an obvious build-up of military equipment at the border with Latvia, I expect NATO would start matching it tank-for-tank. Might as well make it clear that even without nuclear weapons it's going to be a bloody fight.

> I expect NATO would start matching it tank-for-tank

I think you're being very optimistic here. Before the invasion, there were so many people in saying Putin was just bluffing. Certain US media, US political party. Ukraine tried downplay the risk. Even Germany had to get their arm twisted to do sanctions after it happened.

NATO activated the NATO Response Force for the first time in Alliance history. That's around 40,000 troops moved into forward deployment positions.

https://www.nato.int/cps/en/natolive/topics_49755.htm

> If Latvia is attacked and the NATO doesn’t fight for them NATO stops being a thing immediately.

That's what he might be betting on. To be fair, a series of decisions[0] by the West over the last decades might have led him to believe the West may not have the stomach for war with Russia, especially when it comes to defending new NATO members.

[0]

- Invasion of Georgia in 2008: no meaningful response to Russia

- Obama's "red lines" in Syria

- Annexation of Crimea in 2014: no meaningful response to Russia

- Invasion of Eastern Ukraine in 2014: ditto

- Invasion of the rest of Ukraine (under way): response limited to sanctions for Russia and supplies for the Ukrainians

> His most recent take is that by refusing to get directly involved in Ukraine, NATO creates an understanding in Putin's mind that it wouldn't get involved in, say, Latvia either.

The link posted was from 2015.

I'm unable to find any recent writing in English that indicates that he thinks that is a belief in Moscow.

Instead, on Twitter he shared this Feb 27 story (not by him) that indicated Putin has a great fear of NATO forces:

But this underestimates Russian respect for NATO’s conventional military power, which is far greater than its own. The only language Putin understands is force, and Moscow did not retaliate after Turkey shot down a Russian military plane in 2015 or when the United States struck and killed Russian mercenaries in 2018

https://www.foreignaffairs.com/articles/ukraine/2022-02-26/m...

> I'm unable to find any recent writing in English that indicates that he thinks that is a belief in Moscow.

For now, I can only find sources in Russian, e.g. this:

https://www.youtube.com/watch?v=Zv-WB8AXLKs

https://echo.msk.ru/blog/piontkovsky_a/2988037-echo/

Translated version: https://echo-msk-ru.translate.goog/blog/piontkovsky_a/298803...

He makes an interesting argument:

For years now, Russia has been training the United States like Pavlov's dog, delicately preparing the Americans for their most important capitulation - the refusal to fulfill their allied obligations under Article 5 of the NATO Charter.

I agree there is a risk Putin thinks he has achieved this.

Traditionally, during the Cold War, the US would station US troops in pretty-much indefensible forward locations (eg West Berlin).

The theory was that these troops would be killed by any attack, and US allies could be confident that the loss of US lives would be enough to make sure the US responded.

Interestingly, NATO has prominently been pointing out the existence of "four multinational battalion-size battlegroups in Estonia, Latvia, Lithuania and Poland... led by the United Kingdom, Canada, Germany and the United States respectively"

https://www.nato.int/cps/en/natohq/topics_136388.htm

And repeating their commitment to article 5:

"Putin’s war affects us all and NATO Allies will always stand together to defend and protect each other," he said. The Secretary General also noted that NATO is deploying elements of the NATO Response Force for the first time in its history. "Today, French troops are arriving in Romania as the lead element of this force. Our commitment to Article 5, our collective defence clause, is iron-clad,"

https://www.nato.int/cps/en/natohq/news_192548.htm

The reluctance of Trump to back article 5[1] will cost lots of lives.

[1] https://www.brookings.edu/blog/order-from-chaos/2017/05/17/t...

I would prefer being wiped out of a giant meteorite than all nukes going off because at least I know it wasn't humanity that did us in.

This makes very little sense to me. Nuclear war is not a random event that might happen at any given point with some probability P. It is a decision that is made based on an evolving situation, there are not independent random event you can apply high school math to.

Also an event being rare does not mean that it is unlikely. For example a 29th of February is a rare event, if you were to think about it in terms of probabilities, there is about 0.07% of chance for it to occur on any given day. So you would conclude that there is about 1-(1-.0007)^(2*365)=40% chance to occur within 2 years. But it is obviously absurd to think this way, since the next 29th of February will happen in 2024, in less than 2 years, the probability is 100%.

The whole article is basically about different interpretations of probabilities -- probabilities as reflecting the degree of belief, vs pure frequentist ones. Both make sense, they just need to be used in different situations and need to be interpreted differently.

May be missing some subtlety, but at a quick read I would respond to this as:

- The author has not deeply understood the subjective interpretation of probability, as widely used by bayesians, and applicable to this sort of reasoning

- the author, and the sources the author quotes based on a quick read, seem unfamiliar with quanifying uncertainties not just as point estimates, but instead by using a probability distribution, which you can use to easily tell the difference between a mean probability of .5 based on say 5 heads and 5 tails, vs one based on 500 heads and 500 tails.

The point that it's often difficult to tell whether a particular probability should be 0.0001 or 0.00001 is a fair one.

But overall I don't take much reassurance from this article.

the gambler's ruin fallacy is a typical example of cumulative probability. cumulative probability measures the odds of two, three, or more events happening. there's just one catch involved: each event needs to be independent of the others. you can't have the outcome of a first event influence the probability of the next. which we know for nuclear war it be dependent upon the probability of previous events. ie. if a nuclear war starts, either everyone gets annihilated, which in turn influences the probability of the next nuclear war to be zero, or one or both of nuclear powers get severely damaged with catastrophic consequences, therefore influencing probability of future similar events

In a sense the dice are rolled constantly. Every time someone monitors a nuclear power’s early warning system there is a risk that system will return a false positive. There is a chance the operator won’t recognize it. There is a chance the geopolitical situation will cause leadership to believe it is a true positive. There is a chance they will choose to launch on warning rather than risk loosing the opportunity at a meaningful counter attack. I agree that it’s probably impossible to estimate the probability, but it is definitely greater than 0. Eventually the dice will come up snake eyes. Hopefully not before we figure out how to stop rolling them.

The original article does this straw man trick of pretending that the models act like every day is independent. That’s not how it works at all. They’re not modelling the chance of a spontaneous war erupting without cause on day X. Motivated authors just pretend that’s what the models mean because it makes it easy to ridicule.

Regarding this — “it’s probably impossible to estimate the probability” - ok - well. I think we have to unpack that a lot!

I would say: “we can model the event.” That’s an absolute. (Anyone who disagrees with that statement just does not understand the statement)

It does not mean the models are accurate. The model will provide probabilities. How much confidence can we have in the probabilities? Now this is where the difficulty lies. The probability itself will be a low number. And our confidence won’t be very high. That’s automatically creates an opening for any bad faith commentator to say launch an attack. And yet… what’s the alternative? Pretend it cannot be modelled as various commentators in this page (but not you) have done? Pretend that a model with low probabilities or low confidence can be dismissed? The best response (if we agree that nuclear war is a Bad Thing and worth avoiding) is to improve the models, not ignore them - and of course in so doing —- to try and avoid nuclear war.

> Putin does not roll a dice every morning and launches a nuclear missile if a six shows up

Very true. Using, essentially, the infinite typing monkeys idea to say that a nuclear war is bound to happen eventually is just a pseudo-scientific rationalization of war. Put this way, almost anything becomes certain if you completely ignore the fact that you don’t know P(nukes|war) since we have no further data to base this on. Even a coin flipped once or twice might not be heads and tails. The idea of P(heads) = 0.5 is that eventually, as the flips extend to infinity, it equals 0.5.

This would be similar to someone living in the Black Plague saying “since 10% of Europe is dying per year, in 10 years we will most certainly be extinct.” I don’t know the technical or formal name for this fallacy or line of thinking, so I’ll just call this gross oversimplification.

That’s not how the modelling works. That’s a mischaracterisation, a straw man argument.

What about asking what's the probability that nuclear-powered nation gets an autocratic leader who can decide on his own whether to launch the first strike?

I think the current situation most strikingly points to the dangers autocratic governments pose to the whole planet. If people of a nation could vote on whether to start nuclear war I think they would vote no. But a crazy dictator? Not so sure. It should be the law everywhere: No nuclear first strikes. But that can only work if rule of law is the rule.

The other great concern is misinformation and ability to keep most of the nation misguided enough to support a war. There should be international law to prevent spread of misinformation that can cause or lengthen wars.

So this is basically saying that it's extremely hard to estimate these tail-event probabilities reliably, and therefore conclusions should not be taken as gospel. But we mostly knew that already?

Good piece. Typo in the last line: It's "stave off", not "stage off".

On the other hand we have a lot of data points of humans and nations using every available edge to expand their power throughout history. Also we have made social systems where narcissistic greed and lust for power seems to get pretty much rewarded with more money and power, and misinformation runs rampant. More powerful weapons systems will get developed until MAD might not be certainty for the aggressor, if bombs can be delivered stealthily in minutes. Or at least MAD might get tested with one warning shot, delivering a message of "what is your actual defense going to be, I don't really care about the world anyways"

I don't really see it as much as daily roll of a dice but more of a system that might guarantee to force us down that road some day within a few generations.

People think MAD is 'we launch everything at our enemy if we see one missile' -- this hasn't been the case for decades, since all sides can see launches almost as soon as they happen. So yes, 'warning shots' are now effectively the start of such a conflict. Where MAD comes in is it's generally accepted that if you do that your opponent is likely to lob one back. An eye for an eye. I suspect capitals of nuclear powers are special cases though, and would generally be avoided unless absolutely necessary.

I don't think Putin needs to actually start nuking European countries to get his way though. He knows the Americans simply do not have the resolve to escalate a nuclear confrontation unless its own domestic security is threatened. I don't think the UK or France would either. NATO is kind of a sham.

But Putin isn't worried about nukes in Ukraine as much as ground troops. The fact that he is suggests he's not worried about getting nuked himself, which is actually a positive sign, for everyone except Ukraine.

ETA: I just realized in Biden's SOTU he said the US would defend every inch of NATO territory and make the Russians pay, which was clever, since neither of those statements suggests either an invasion of Russia nor a nuclear response, only that Russia wouldn't be able to hold the territory and that there would be repercussions of some kind. Not a dumb guy.

Every day I wonder if what is happening in Ukraine right now - will turn out to be another invasion of Czechoslovakia 2022 edition written in history books as a missed opportunity to turn the tide as the world watched paralyzed by Putin’s nuclear rattle.

Legitimate concern, yes, but here’s the food for thought and mark my words - it is inevitable that very soon Russia’s rhetoric will turn to “economic sanctions are intended to destroy Russia and are an act of war”, along with “supplying Ukrainian resistance is an act of war”, while the nuclear rattle comes out once again in one form or another, like “Russia is forced to respond by arming with nuclear weapons all allied states, such as Iran, Talibistan, North Korea, Venezuela” etc…

what is Talibistan?

I believe the author refers to Afghanistan under the rule of Taliban...

Yeah, I started typing it out, but then took the shortcut figuring people will get that from the context.

but Afghanistan under the Taliban is certainly not an allied state of Russia...

I don’t think I would be saying something scandalous and shocking by pointing out that for Russia - enemy of my enemy is… if not a friend, then an associate at the least, and theres nothing that’s seems outside of ethics for Putin’s regime.

> Putin does not roll a dice every morning and launches a nuclear missile if a six shows up.

You wish! The problem here is that Putin is not rational.

This makes me chuckle. The Bad Orange Man played the "insane king" role really, really well. Everyone was on their best behaviour until he left office. Peace talks in the Middle East progressed. Russia stayed in their borders. North Korea huffed and puffed and ultimately did nothing. Trade with China became more favorable to the US. It's almost like his foreign policy was effective.

To your point: Putin isn't nuts. He's just a really good politician. He's pragmatic and knows the value of being seen as "crazy" and "he might actually launch a nuke".

Syria

Trump also pushed the EU and the rest of Asia towards China

https://www.americanprogress.org/article/trump-withdraws-ame...

The same was said about Nixon during the Vietnam war (which he inherited, but would never have started) but it is a really risky stratergy once you start ramping up the mania levels and passing them on to your nuclear forces, which Nixon (a good president, all in all) nor the Big Orange ever did.

Trump wanted to destroy NATO and European unity, he was doing Putin's job for him.

Destroy it by pressuring its members to pay 2% GDP? Or what are you referencing?

And how does sending Javelins to Ukraine or blocking Nordstream 2 help Putin?

Trump tried to prevent military aid to Ukraine; dunno how you give him credit for "sending Javelins..."

First off, I assume you're talking about the phone call that the Democrats impeached him for (leading to acquittal) to cover for the Burisma scandal. Accuse your enemy of what you're doing I suppose: https://www.youtube.com/watch?v=UXA--dj2-CY

Second, that aid was never held up and is separate from the Javelins that were sent anyway.

What other President sent Ukraine weapons (not blankets) before this fiasco?

What did the USA have to do with blocking Nordstream 2? Germany stopped certification, no the USA.

Germany just stopped it after Russia attacked Ukraine. Trump held it off with sanctions during his Presidency.

https://www.atlanticcouncil.org/blogs/ukrainealert/us-sancti...

He wanted them to pay their fair share. They wanted the US to foot their security bill which he balked at.

Now we see Germany at least say they will up their contribution and rebuild their military in order to be better prepared.

That’s pretty much in line with what he asked from them.

Allies and Former U.S. Officials Fear Trump Could Seek NATO Exit in a Second Term https://www.nytimes.com/2020/09/03/us/politics/trump-nato-wi...

“Yeah, the second term,” Trump had said. “We’ll do it in the second term.” https://www.washingtonpost.com/politics/2021/07/13/book-exce...

Those articles are lying by omission, and so are you.

Trump was certainly looking at getting out of NATO, which he saw as an arrangement where the US increasingly foots the bill for protecting other members. He openly criticised European dependence on Russian oil and gas, and Germany in particular for not paying their fair share towards defence.

There's a fantastically awkward video [1] of him explaining all of this directly to NATO General Secretary Jens Stoltenberg. Germany announced this week that it will do exactly what Trump said it should: massively increase military spending [2] and decrease its dependence on Russia for energy [3].

Who is more useful to Putin? Trump, who told Europe to get its ass in gear and treat Russia as a threat, or the stooges you support, whose selfishness and complacency emboldened Russia to carry out this invasion?

[1] https://youtube.com/watch?v=M5MJEMHZZsI

[2] https://www.spiegel.de/international/germany/budgetary-about...

[3] https://dw.com/en/ukraine-crisis-forces-germany-to-change-co...

Could you do me a little favour? I'd like you to watch a short video[1] from CBS News and explain what exactly Trump was getting wrong about NATO/Russia/Germany.

[1] https://youtube.com/watch?v=M5MJEMHZZsI

(Disclaimer: I agree that orange man bad)

You will get a lot farther if you are an insane man with a gun, than if you’re just a regular man with a gun.

Alternate take: all those participants played nice in the hopes that it'd lead to an autocratic take-over of the US, neatly getting rid of their major enemy. In that vein, Trump was just the "best moron" they'd seen in a long while.

The problem is when you consider the game-theoretic incentives of "Greater Madman" posturing, through the len of Trivers' "Folly of Fools" [0]: that the most effective way to fool others is often to first fool yourself.

It is entirely plausible that Darwinian selection pressures rewarded not just those who cynically postured as madmen, but those who actually were madmen (within some bound of Minimum Viable Rationality). Such madness only has to beat the average; unlike the MAD dynamics of nuclear weapons, during most of evolutionary history, a catastrophic error of genuine madness didn't result in destruction of the entire species.

[0] https://en.wikipedia.org/wiki/The_Folly_of_Fools

He might not be rational, but surely nobody can argue that he is insane to that level – at least, I haven’t seen anything that suggests that.

He's just a little bitch screaming to everyone's face "I'm gonna use nukes!" hoping we give him what he wants.

What does he have to lose by alluding to using nukes right now?

Nukes are the only thing stopping the West from rolling into Moscow, the magic wears off if the clock strikes 12. Nukes only work for the 15 minutes before they are used, Putin and Lavrov keep crying "Nuklear wolf!".

> Nukes are the only thing stopping the West from rolling into Moscow

Yeah, tell Napoleon and Hitler how easy that is.

Napoleon and Hitler were not defeated by Russia. Napoleon/Hitler could have raised 3 more armies if they were fighting just Russia.

On the contrary, Putin is the new insane and isolated Hitler.

The Soviets destroyed 3 times as many German divisions as the western front allies combined.

Because the german forces were split and the allies supplied USSR, otherwise the USSR would have collapsed. Without the allied material support and the german forces being split, the USSR would have been deffeated.

It's not something that I would have liked, just the truth.

People claiming Russia to be invulnerable miss the real lesson of those conflicts -- the ballance of power.

The ballance of power is set against Russia right now, if the West remains strong in its unity and resolve, Russia will collapse.

So?

Whoever is insaner wins?

Nah, remind Putin that we all saw how much the russian army sucks and in case of nuclear war the US still wins after they get out of their bunkers.

with the greatest respect, this is the greatest evidence of insanity I’ve seen in this thread…

This is the only approach that works and won the US the Cold War. A credible threat that if the USSR dared to use nukes, NATO would use nukes against the USSR.

Do you think Putin is a genious and it didn't cross other USSR leaders minds the same idea -- "surely, the USA would not use nukes again us if we used just a few nukes against a weaker NATO member"?

The only sane answer is to absolutely make sure the other guy knows you're going to use nukes the second they use them.

Macron's retort to Putin needs to be shouted from the rooftops -- we also have nukes, you're not that special.

Also, the question is: if Putin orders it, would it happen?

Nobody wins in a nuclear war.

This is actually exactly what Putin said - that NATO is much stronger, but if there is a nuclear war, it's going to be bad for everybody.

People look like they make irrational decisions when viewed from the outside with little inside knowledge, there is definitely more to this than we are aware of.

Please don't post unsubstantive and/or flamebait comments to HN. The issue is not that you're wrong about you-know-who, the issue is that it has seriously degrading effects on internet threads, and we're trying to avoid that here. To the extent possible.

If an article is substantive, it deserves better than this—much better, actually. And if it isn't substantive, it shouldn't be on HN's front page.

https://news.ycombinator.com/newsguidelines.html

>You wish! The problem here is that Putin is not rational.

"There is no greater danger than underestimating your opponent." - Lao Tzu

What is an ensemble probability of something that has never occurred and is essentially impossible to model? Have you seen all infinite alternate universes? Probability does not make sense.

Compare the following: What is the probability that you will die today?

What is the probability that you will die today given that you belong to the category of people with A_i characteristics and whose ensemble probability of dying on a given day has been measured?

The insurance industry does a pretty good job of “what is the probability that you will die in the next billable period” - but they would never reduce the bet to “what is the probability that you will die today?”

You say it’s “essentially impossible to model” something that hasn’t occurred before.

Nations have engaged in a lot of wars.

Nations have engaged in a lot of wars they could not possibly win.

People have committed suicide to harm other people. Many many times.

Nations have developed weapons and then used those weapons on other people.

Countries that have engaged in arms races and stock piled weapons have then gone to war. Many times.

In some decades we do more of these things and in others we do less.

To say it’s hard to model, or the models are imperfect (as all models are) - fine. To say it is impossible to model - that’s very naive.

If you are lost in a city would you rather have the wrong map or no map at all?

The person that has the wrong map and thinks that it’s the right map is the one that’s naive.

Regarding insurance companies, you are conflating ensemble probabilities with time probabilities. Insurance companies can estimate the ensemble probability of an event in a group of agents which belong to a certain category and not the time probability of said event on a single agent.

I’m pretty sure we’re in agreement in the insurance policy part.

If I buy a policy to cover my risk of death for the next X years, the insurance company uses the ensemble probability of the group to price that policy (and believes that the risk of me knowing more than them is covered by the fact they’ve factored that in, plus that they’ve insured multiple people in the group and the aggregate will be good even if 1 policy isn’t.)

If I said to an insurance company — okay I want to buy a policy that only covers me for one day - they would not do it. Firstly of course the transaction costs would swamp the cost of the policy, but let’s ignore that by pretending they use some magical technological solution (digital contracts… hardy har har) The second reason they wouldn’t do it is because that’s not how their model works. It has to smooth over time. Over a year, over a lifetime - not over a day. The third reason they wouldn’t do it is because the information asymmetry now dominates. My knowledge of whether I’ll die tomorrow in particular is better than their model. I’m able to price the policy much more accurately than them. The fact I want the policy is a big red flag. (They do offer some very short term contracts such as holiday insurance — but those disclaim every imaginable pre-existing… which is out of scope for this discussion as it’s not something we can do with nuclear risk)

I think we more or less agree on that stuff. Don’t we?

What I’m saying about it is that just as we all know that insurance policies don’t make sense when you reduce it down to thinking about an individual person on an individual day — the nuclear war modelling doesn’t work when (the author of the article) reduces it down to saying it’s like suggesting that Putin flips a coin each morning.

The most likely reason an author would pretend that the models claim such a plainly ridiculous thing is in order to ridicule the models, about a claim they did not make. In this way it is a straw man argument. They are recasting the original assertion into something that is easy to attack. It’s not a well intentioned thing to do.

That’s what I’m saying about that.

Regarding the question:

“If you are lost in a city would you rather have the wrong map or no map at all?”

I’d talk to a local, because they have knowledge on the topic. They’re a bit like the experts who have spent years in the field here, whose opinions are being dismissed with straw man arguments.

Generally when a non-expert on a topic says “here is a glaringly obvious flaw with an entire field that has been overlooked by everyone working in that field, that I spotted after two minutes of deliberation” - it’s quite a big call. When they then support their argument with primarily straw man arguments and other bad faith tactics, it’s very weak and okay to just ignore it.

Reply on news.ycombinator.com