What is the Probability of a Nuclear War?

I agree with Tyler who wrote recently that “the risk of nuclear war remains the world’s No. 1 problem, even if that risk does not seem so pressing on any particular day.”

The probability of a nuclear war is inherently difficult to predict but what strikes me in this careful survey by Luisa Rodriguez for the Effective Altruism Forum is how much higher all the expert predictions and model forecasts are compared to what we would like them to be. Keep in mind that the following are annualized probabilities. For a child born today (say 75 year life expectancy) these probabilities (.0117) suggest that the chance of a nuclear war in their lifetime is nearly 60%, (1-(1-.0117)^75). At an annualized probability of .009 which is the probability from accident analysis it’s approximately 50%. See Rodriguez and also Shlosser’s Command and Control on the frightening number of near misses including one nuclear weapon dropped on North Carolina.

These lifetime numbers don’t strike me as crazy, just crazy high. Here is Rodriguez summarizing:

If we aggregate historical evidence, the views of experts and predictions made by forecasters, we can start to get a rough picture of how probable a nuclear war might be.[8] We shouldn’t put too much weight on these estimates, as each of the data points feeding into those estimates come with serious limitations. But based on the evidence presented above, we might think that there’s about a 1.17% chance of nuclear war each year and that the chances of a US-Russia nuclear war may be in the ballpark of 0.39% per year.

Addendum: A number of people in the comments mention that the probabilities are not independent. Of course, but that doesn’t make the total probability calculation smaller, it could be larger.


Good points. But, if America is to fight a nuclear war with totalitarian Red China, we need to leverage our resources in an adequate way, right? Allies, for example. Who can we trust?

England is collapsing. Poland is ridiculous. France and Germany don't like us. Canada is France-light. Mexico is a mess. While we pamper a bunch of ungrateful, our most loyal and stable ally, Brazil, has fair demands we overlook to our own risk. Brazil's bases are near the Equator and can be used to launch nuclear weapons. Imagine what it would be if Red China get hold of them and used them to attack America. Wouldn't spirning our Brazilian friends a penny-wise, pound-foolish measure. I think so.

England isn’t collapsing we are having a robust debate over the EU and it’s not England it’s Great Britain&Northern Ireland. Your ignorance is embarrassing

1) England is the problem. It is the Englishmen who want to leave, but have no idea how to do it.

2) "we are having a robust debate over the EU".
This debate should have been finished before the referendum. The current debate, a short one, should be about how to leave the EU or shut up already.

3) It is not clear for how much time English colonial control over Ireland and Scotland will last. As the English regime collapses into anarchy, its control over the British nationalities will melt as Russian control over Georgia did.

"Your ignorance is embarrassing" Indeed. It is happening right before your eyes and if anyone points it out to you then you become offended.

You are Thiago Ribeiro and I claim my 5 reals.

They are not reals, they are reais. There are nomreais involved and you are mistaking me for another person.

Why do you persist? Your audience only laughs at you, Thiago.

No, it doesn't because I knkw no Thiago. You must be mistaking me for another person.

Ha ha. Keep dancing, clown.

I don't dance.

Brazilians do not like the US. I can't tell you how many rants I have had to endure. I think they are mad they were asked to pay back world Bank loans stolen by their countrymen, I dunno.

In contrast, Brazilians are warm and open to Americans as individuals.

IAC, Thiago's post is a troll.

Why do I even bite? 😂🤣

My point is, we should act soon.

Allies? You are forgetting Australia which already hosts US forces and is not merely a loyal and reliable ally but a positively sycophantic ally.

True, but note that the superforecaster probability is about an order magnitude below the rest. The betting market (ie, the entire economy of the northern hemisphere) appears to place the odds much lower than that.

Also, there is just no way that the next nuclear exchange has a 1/3 probability of being between US and Russia. I’d put the odds of US/Iran, Israel/Iran, US/China, US/DPRK, India/Pakistan, and other/other individually higher than US/Russia.

+1,"Also, there is just no way that the next nuclear exchange has a 1/3 probability of being between US and Russia."

Someone is still stuck in the Cold War.

A nuclear exchange implies two parties using nuclear weapons, one against the other. There's also the possibility of one party pre-emptively using a nuclear weapon and the other being unable to respond effectively. Or, the object of a nuclear attack being unable to know with absolute certainty who the aggressor might be.

The precision of 1.17% implies an accuracy that doesn’t exist. It would be much more credible to say something like “around 1%”.

Putting the “historical” number into an average seems bogus.

For whatever reason, economists are bad with this. My guess is because most aren't exposed to much science and those who are are too few, percentage wise, to change the rest.

Obviously, the probability of using a nuclear weapon prior to 1945 was nil. And the atomic bombs were purpose built TO be used. We weren't building them merely for capabilities. This distinguishes those two bombs from every nuclear weapon built after 1945. So figuring in Little Boy and Fat Man into the calculation grossly overestimates the probability of using every bomb thereafter.

Three has been more than one dropped nuclear weapon.

'The 1966 Palomares B-52 crash, or the Palomares incident, occurred on 17 January 1966, when a B-52G bomber of the United States Air Force's Strategic Air Command collided with a KC-135 tanker during mid-air refueling at 31,000 feet (9,450 m) over the Mediterranean Sea, off the coast of Spain. The KC-135 was completely destroyed when its fuel load ignited, killing all four crew members. The B-52G broke apart, killing three of the seven crew members aboard.

Of the four Mk28-type hydrogen bombs the B-52G carried,[2] three were found on land near the small fishing village of Palomares in the municipality of Cuevas del Almanzora, Almería, Spain. The non-nuclear explosives in two of the weapons detonated upon impact with the ground, resulting in the contamination of a 0.77-square-mile (2 km2) area by plutonium. The fourth, which fell into the Mediterranean Sea, was recovered intact after a ​2 1⁄2-month-long search.' https://en.wikipedia.org/wiki/1966_Palomares_B-52_crash

Nuclear weapons simply are not that easy to detonate by such types of accident.

+1, "Nuclear weapons simply are not that easy to detonate by such types of accident." Coincidentally, I just read the Soon-ish chapter on nuclear energy this morning, which also emphasizes this:

"It’s reallllllly hard to make a nuclear bomb. In order to get a big boom, you need to split as many atoms as possible before it all blows apart. This is hard because every time an atom splits, it pushes other atoms away. Making a bomb requires extreme precision just to get 1% or 2% of the nuclear fuel to fission. It can’t happen by accident any more than shaking a bunch of metal parts can accidentally produce a functional watch." https://smbc-comics.com/soonish/lostchapter/index.html

These probabilities presumably aren’t independent so a 1% annual chance does not equate to a 60% lifetime chance.

“Nuclear war” can also mean a lot of different things, from the US dropping a tactical nuclear weapon on some fortified military position in third-world country, to a limited nuclear exchange that takes out a city or two, to the end of civilization. I’d grant that there might be a 10%+ lifetime chance of the first happening, and maybe even a 1%+ chance of the second, but the third seems even more unlikely than the destruction of civilization from climate change. People have way too much to lose from a nuclear exchange. These days, all the big nuclear-armed countries aren’t even willing to fight conventional wars except against much weaker foes. Most nuclear countries have already disclaimed nuclear first use and the existence of submarine, mobile, and air based delivery systems means that countries won’t face a use it or lose it dilemma for a second strike.

Even more assuming with this assumption is that there has already been one nuclear war in the last 75 years.

Though for whatever reason, most people do not consider WWII an example of a nuclear war.

There was no MAD since only the U.S. had the bomb.

As nukes migrate down to smaller, crazier, countries, the rules change.

Does Kim have a red button that launches without human mediation? Does he get drunk on smuggled whiskey, and press it?

I tend not to worry about such things, because they are beyond our control .. and they will remain so until some figures out a better answer than sanctions.

All the little countries will endure sanctions to keep their nukes.

Revealed preference.

"All the little countries will endure sanctions to keep their nukes."
Well after Libya they will. Given just that one f* up, anyone who says the Obama administration had a good foreign policy needs to be kicked in the balls good and hard.

Really, just learn one true thing today.

The Libyan Civil War of 2011

United Nations Security Council Resolution 1970 was a measure adopted unanimously by the UN Security Council on 26 February 2011. It condemned the use of lethal force by the government of Muammar Gaddafi against protesters participating in the Libyan Civil War, and imposed a series of international sanctions in response.

The Security Council resolution marked the first time a country was unanimously referred to the International Criminal Court by the council.

Some people said "hey, we had an agreement with Gaddafi," but who was "we" and does a full blown civil war change that?

Should Kim expect that in a popular uprising against him, the US can or should stick with him?

Why, exactly?


Learn something else today.

Ironic isn't it that Obama's legacy is open slave markets in North Africa.

North Korea is a catastrophe no matter what happens. Mad men who kill each other, with nukes. Not able to feed themselves. The possibility of a refugee crisis that would make the current European situation pale into insignificance. Imagine hundreds of thousands hitting the Chinese border.

Snipe from the sidelines all you like but just remember those Obama slave markets before you start pontificating.

lol, this has to be an impersonation.

Nope, just the usual partisan hypocrisy. Had to get rid of the tyrant Saddam, the mess afterward was not Bush's fault. But don't support the tyrant Qaddafi as his own people rise up against him? Oh man everything after is all on Obama.

Politics makes people stupid.

"All the little countries will endure sanctions to keep their nukes."
Well after Libya they will. "


Iran was being good until Trump put the sanctions back on. Now they are hellbent on getting those nukes again.

No, they weren't being good, which is why we put sanctions back on.


"A Defiant Iran Defies the UN and International Laws Again - 02/20/2016 10:50 pm ET Updated Dec 06, 2017:

For many, it is baffling that Iran is now capable of getting away with breaking international laws, particularly in the last few months after the nuclear deal was reached between the six world powers (known as P5+1; China, France, Russia, the United Kingdom, and the United States, plus Germany) and the Islamic Republic.

Lets take a look at some of the latest violations which are linked to Iran’s military institutions. In clear violation of the United Nations Security Council (UNSC) resolution, Iranian leaders test-fired, long-range ballistic missiles and laser-guided surface-to-surface missiles several times. Additionally, in October and November, Iran tested a new ballistic missile capable of carrying multiple warheads.

Obama and everyone else knew they were cheating. At least now we’re not footing the bill for it.



When you can't actually stop them short of war, maybe a blind eye really is better than brinkmanship.

A week or two ago we did the full cyberwar press for an attack "in the works."

It was close, but you know, convince yourself that "Sherman Tank" Trump is on top of things.

Except we now pay in downed drones, tanker attacks, lost oil, and cyber attacks. Might have been cheaper to just keep the old Obama agreement.

Why pay them anything if they're not holding up their part?

Please tell me that you don't think we paid them any of our money.

-1 for misuse of the term 'revealed preference.' You are not good at this whole writing thing.

"Revealed preference models assume that the preferences of consumers can be revealed by their purchasing habits."

It's not a huge leap, from there to:

"Revealed preference models assume that the preferences of [actors] can be revealed by their [actions]"

It's a simple generalization of the concept.

I'd worry about anyone who can't see it.

I'd reckon that the choices for political actions is a tad larger and more dimensional, budgets less constraining, markets less complete, and the choice set less convex. Objectives of the political actor need not be utility optimizing.

And if you are suspicious of utility in the beginning, the ground looks even less soli

I can see what you were going for; that doesn't make it correct. Terms have definitions. Don't try to repurpose them to make yourself sound smarter. It doesn't work. Invoking a few academic sounding terms doesn't camouflage stupid.

I wasn't trying to sound smart.

I didn't pick this petty fight.

But it is clear to (simple) me that purchasing decisions cannot magically be an insight into underlying emotions, and all other decisions not.

Economics is a subset of human nature. It is not removed from human emotion nor cognition.

The use of very small nukes to take out a critical fortified position seems like the only realistic Western use. Perhaps against a smaller nations' leadership after they nuke us (or a significant asset like a carrier group).

Nukes may, unfortunate, be the only option since we (the west) gave up an effective chemical response. Certain chemical agents (HCN) are very useful for killing effectively, dissipating quickly, and leaving structures intact. Shortly after an HCN strike, a well trained (and protected) assault force could then swoop in and gather intel.

It's all seems horrible to contemplate. But think: are deadly (but quickly dissapating) chemicals worse than a nuke blast (with decades of irradiation)?

As I noted elsewhere fallout is not inevitable with a nuclear weapon burst. It depends very much on how close to the ground the nuke explodes.

The calculation of probability over 75 years ignores that the probability of a nuclear war in a given year is not random, but a mix of effects from actions of previous years, current political and social trends, and so on.

The "60%" or "50%" number makes for a scary post or article, but is meaningless in the real world.

Does this mean that odds makers at racetracks and sports books are wasting their time?

- The probability of nuclear war is not independent from one year to the next.
- The default assumption is that superforecasters are taking into account all non-garbage info in expert surveys and the residual difference in expert surveys is because the expert surveys are garbage about big-sounding rare events.

These points are right on. I believe that the GJP algorithm also "extremizes" superforecasters' estimates, which is one reason you'd naturally expect those probabilities to be lower than in other, expert surveys.

In principle, one could try to deal with the non-independence problem by varying prediction windows: e.g., ask experts to predict the probability of war in X years, 2X years, 3X years, and so on. But most people really struggle with this kind of scope sensitivity.

The extremizing algorithm is applied to the overall consensus. No single superforecaster estimate was changed.

While I agree that it's a rather important risk to keep in mind I don't think the probabilistic approach is correct. We're not in any form of stable population we can sample to calculate yearly probability and then expected probabilities over time.

That is simple wrong thinking.

... casually assigning the statistical term "Probability" to what is essentially mer subjective "Opinion" is deceptive nonsense. GIGO

What exactly is a "Nuclear War" for such "probability" calculations?
Is it an all out nuclear war between Russia and America ... or some terrorist group detonating a crude nuclear device against their nation-state foe? Definitions matter a lot.

Are you unfamiliar with subjective probability?

The problem is you people write as if nuclear war is a bad thing.

The gravest threat the World faces is liberalism.

username checks out

People are pretty bad at time sensitivity though, so this is going to produce an overestimate. For example, currently PredictIt has an 18% chance for "will Trump be impeached by the end of 2019." If you were to do the same analysis- "well, that's six months, and there are three six month periods left in his term in total"- you'd get roughly a 45% chance of impeachment during his first term, but that exact scenario is only priced at 28% by the exact same prediction market.

Someone should give an award for the comment that best combines smug assertions with head slapping missing point mental feebleness on this craphole site. You’d win for today. The closer you get to the election the faster the possibility of impeachment erodes. After 2019 there is zero chance of impeachment.

There has always been zero chance of impeachment, because the Republican Senate will never convict.

Absent a significant event or revelation. A recording of him admitting to a crime, perhaps one already alleged, would cinch it.

But I dont think those chances are any greater with him than any other president. They certainly differ in character, but we are all guilty of at least one felony.

I doubt even then.

Holy shit you just swooped in and snatched the award. There was no chance the senate would convict Clinton either. Nonetheless.

Impeachment is the charge of high crimes and misdemeanors brought by the House.

The Senate then conducts the trial to judge whether the POTUS should be removed from office.

No shit. Holy crap this award is going to need a gold silver and bronze medal. Read the comments in order moron. You might actually get the gold because you were so anxious to try and correct someone you didn’t even grasp the context. This place is shittier than Central America.

The fact that the annual probabilities of a nuclear war are not independent does not mean that the total probability is lower.

Sentences like"See Rodriguez and also Shlosser’s Command and Control on the frightening number of near misses including one nuclear weapon dropped on North Carolina." do not inspire confidence in the overall argument: If the device had detonated (highly unlikely) that would have been very bad, but it would not have led to nuclear war.

So it wasn't a near miss at all and was not really related to the subject, which was the chances of nuclear war breaking out.

Shlosser's Command and Control has been criticized for various reasons (see the negative Amazon reviews) but his technical knowledge seemed weak in his dramatic opening where he explains that had the nuke gone off, history would have been altered because Al Gore was in Little Rock that day talking to the Clintons. But Damascus, AR is 45 miles away from Little Rock, AR and much too far for Gore and the Clintons to have gotten radiation sickness, let alone died.

The nuke that fell out of a bomber would have plastered a pretty small, sparsely inhabited area so hundreds or possibly thousands would have died had it gone off, but it wouldn't have sparked a nuclear war and death or cancer from radiation would have been small.

The counterfactual must be what would have happened without nuclear weapons. I would suggest that the likelihood of another world war should be 300%. In other words the would have been three more since 1945. Nuclear weapons have been responsible for decades of unprecedented peace and prosperity.

And the possibility of use with the horrific consequences for in mind is why.

Peace Is Our Profession.

You're Welcome.

Even ignoring the possibility of ICBMs filled with nerve gas and anthrax, a conventional fuel air bomb can destroy a city center. So even if nukes weren't a thing, the ability of (large) ICBMs, cruise missiles, and bombers to mass murder families at reasonably low cost would still be there, just not as cheap as nukes.

Of course, how that would actually affect outcomes is not easy to predict.

1) Is there even a perfunctory wink in the direction of falsifiability in this?
2) Who is an "expert" in this? Have any of these experts started nuclear wars, or at least advised people who did? Does running war games make you an expert in the same way as playing Halo makes you an expert on alien invasions?

It reminds me of attempts to quantify the Drake equation as to the probability of contact with aliens.

Aside from your point, estimates ignore how incredibly improbable that sentient life can exist. Our Earth is a cosmic accident resulting not merely from a rocky planet in the habitable zone, but also the presence of Jupiter and Saturn to sweep the solar system of debris, a relatively large moon to stabilize our orbit, and a core of molten iron to produce a strong magnetic field to protect the atmosphere from solar wind. I think being on the outer edge of the galaxy also protects us from cosmic radiation and more frequent collisions. It is very much a Goldilocks planet, far more rare than people imagine.

The Rare Earth Hypothesis seems to be going out of favor as a Fermi paradox explanation, in part because you don't need many civilizations to arise to have the paradox, and because the latest science seems to show earth-like planets aren't rare.

Earth-like planets might not be rare, but how rare are earth-like planets in solar systems on the galaxy's edge with large gas giants shielding them and a large moon and a molten iron core?

What can I say? The professionals who think about this stuff no longer believe it's a viable explanation. It's anthropocentric and it posits as essential a bunch of stuff that doesn't seem like it should be essential.

That and the galaxy seems stuffed to the gills with earth-like planets.

There is a good deal of anthropocentrism in amateur estimates, but the smart folks know better.

On the other hand, no one gets media attention by saying that the Delta Sigma 9 solar system can't support sentient life. There is an enormous financial and prestige incentive to identify potential life. It is an almost rabid pursuit.

The point of my comment is that there is far more to being "earth like" than similar size, composition, and location. Earth probably has countless idiosyncratic factors that made it capable of supporting sentient life. It had to be completely free of cataclysmic disasters for hundreds of millions of years, among other lucky breaks.

Indeed, our molten iron core and large moon were the result of an off center collision with another rocky body. This is not merely as unlikely as hitting a bullet with another bullet, but hitting the bullet in just the right manner to A) create a large moon, B) leave an iron core, C) not smash both worlds to smithereens.

The earth also required regular deposits of water from the rocks of millions of asteroids after this event, but not ones so large that it wrecked the earth again.

To overcome this presumption of rareness, one would have to show with a simulation that this confluence of events is actually rather common.

There might also be sentient life on another world that is incapable of movement. Human hands are at least as important as our brains in our development.

It had to be completely free of cataclysmic disasters for hundreds of millions of years

You're looking at time in earth human terms, not in galactic terms, which is probably much different.

Most of the specifics of the Rare Earth Hypothesis are confusing, like the importance of the Moon which just doesn't seem important. You look at the solar system and you see Earth a hit, Mars a possible hit, Venus and all the icy moons with liquid oceans as near-misses, and it's hard to claim other star systems won't be okay. Ours has, what, a dozen reasonable possibilities?

The principle benefit of our moon is that it stabilizes our rotation. The Earth wobbles a bit as it spins on its axis, but we are not at risk of the planet flipping over on its side which be catastrophic for our climate.

There were a bunch of hypothesized benefits of big moons in the book, none of which have survived scrutiny to be considered all that important. Tides and active geology, were prominent among them.

I don't want to beat down on it too hard. It was an interesting idea in a popular book that hasn't held up to scrutiny.

My money is on something between amino acids and multicellular life being astronomically hard, but hey, maybe it's berserkers.

Metallic cores aren't dependent on collisions with other bodies. When a gestating planet is still in molten state (and kept that way due to frequent bombardment of infaling matter as it clears its orbit) haavier Matter will tend to sink to the center.

End of civilization by a nuclear war = 0.005% in 100 years, i.e., slightly higher than by climate change. Risk to get close to that is 5% in 100 years by nuke wars, 5% in 200 years by climate change.
by Global Challenges Foundation, 2015, page 19. https://api.globalchallenges.org/static/wp-content/uploads/12-Risks-with-infinite-impact-Executive-Summary.pdf

It's pretty hard to see how a nuclear war could end civilization. The world would have to ramp up the number of nuclear weapons to the level that the US and the USSR had during the Cold War. And even then it's doubtful that such a war would end civilization.

The chance of climate change ending civilization just seems absurd. I've never seen even a realistic worst case scenario that would come close to ending civilization.

There is enough methane locked up in Siberian permafrost clathrates (35,000 Gigatons CO2 equivalent) to create a possible existential risk for civilization.


Current C02 levels are 3 x 10^12, if a "clathrate bomb" went off it could add 11.6 x 10^12 (1/3 of total) to the atmosphere bringing levels to approximately 14.6 x 10^12. CO2 in PPM is 415, after the clathrate bomb levels would increase to approximately 2,200 PPM (14.6/3 * 415PPM). If roughly 300 PPM cause a 1C increase in a century, we can expect (2200 - 415 = 1785/3) is 5.95C for a total of 7C increase. The

IPCC middle forecast for the middle path adds 3 more C for a total of 10C increase with the clathrate bomb hypothesis.

Existential Risk:


A 10C increase in temperatures by 2100 would increase global average temperatures from 14C to 24C. The exercise for the reader is to determine if our modern standard of living can be maintained at such a temperature.

If we can maintain a civilization (that continues modern standards of living in the northern latitudes) that would be a good outcome in my opinion.

What is the criteria for "expert"? What is the evidence that these people have any ability to predict geopolitical events

I've read the geopol literature for 30 years. I doubt any of these people have predictive skill.

One-third of Americans support a first-strike with nuclear weapons against North Korea even if it would kill one million civilians. https://www.tandfonline.com/doi/full/10.1080/00963402.2019.1629576 And get this: It turns out Americans don’t care about the human toll: respondents in the poll were about as likely to support a first strike that killed a million civilians as a first strike that killed 5,000. Moreover, they were as likely to support a nuclear attack as an attack with conventional weapons. Politicians are likely listening: "If politicians don’t expect voters to care about the difference between 5,000 and 1 million civilian casualties overseas, we probably can’t expect that they’ll care either."

Of course, the only nation to have actually used nuclear weapons is the U.S. , which dropped two atomic bombs, one on Hiroshima followed by another on Nagasaki. Support in the U.S. for the bombings has declined since 1945 (when 85% of Americans supported the bombings as compared to roughly one-half of Americans today). Still, only about 20% of Americans have enough regrets to warrant an apology to Japan.

More Japanese were killed in the conventional incendiary attacks on Tokyo in March 1945 than in either of the nuclear attacks. Should we apologize for the March attacks?

Preparations were underway for an invasion in Fall 1945, which would have resulted (estimate vary considerably) in substantially more American dead than in either nuclear attack, and many times more Japanese.

Many of the 85% who supported the bombing in 1945 were on troopships headed to invasion staging areas, or their families. After Iwo Jima and Okinawa, casualties were expected to be extremely high.

The other main option on the table was continued naval blockade. The population was beginning to starve.

Had the war been prolonged, with either invasion or blockade, many, many more Japanese would have died. Should we apologize for avoiding that?

I knew one of those soldiers. He was a lifelong Democrat who campaigned for Howard Dean back in the day, but he said nuking Japan was necessary, to save American and Japanese lives.

I'm pretty sure I owe my life to the atom bomb. My old man was only 350 miles off the coast of mainland Japan when the bomb went off in 1945. Even though he was a devout Catholic and went on to be a lifelong peace activist, he always spoke in gratitude to the atom bomb -- which he wholeheartedly believed saved his life and the lives of his buddies, who were probably slated to go in on one of the first waves.

Was my miserable life purchased at the cost of the lives of tens of thousands of Japanese civilians? Perhaps it's a good thing that I'll never know.

It was not only your life purchased for that, but the lives of many more thousands of others, both American and Japanese.

There is no peace in the middle east: I drove along a straight stretch as empty construction vehicles and deserted oil factories and cars left to disintegrate passed by, then I weaved in and out of a series of white trucks, some tankers and other grifters, mostly plumbers and electricians, in a Chinese checkers run. I forgot about my turn signal and the break pedal and with a smile I ran a cigarette over the yellow line and held it in my right hand, my thumb and palm on the wheel while my left wrist wavered over the top of the crest of the window. I painted stripe after stripe, and upon breezes, I colored them in, and against the gusts, I outlined stars and along the gales, I drew breaking waves, and when it was tranquil, I broke off crooked r’s.

Everyone is conservative about what they know best.

Downfall was originally to have used large chemical weapons attacks, but once the atomic bombs were disclosed, 10 bombs were added for the southern islands attack and 16 for the area near Tokyo. If I recall correctly, the southern attack (Olympic?) was scheduled for late 45 and the main islands (Coronet?) for 46 and both dates would likely slip. They intended to use all the nukes they could build in the meantime.

There was concern about maintaining Allied commitment during a long blockade, to say nothing of the moral question of starving them out.

Still, only about 20% of Americans have enough regrets to warrant an apology to Japan.

Ray, if you would like to apologize for something Harry Truman decided to do 75 years ago, no one is stopping you.

The atomic bombings saved at least 100s of thousands, if not millions of Japanese lives, as well as countless American, Korean, Chinese and even Russian lives.

Why apologize?

The Japanese should apologize for starting the war, committing war crimes and non surrendering.

I'm more worried about a serious electromagnetic pulse attack than a good old-fashioned bombing. Does that count as "nuclear war?"

Anyone who EMPs a nuclear power had better follow it up with a nuclear attack on that country's nukes, else their own country will be turned onto a parking lot by way of retaliation.
A natural EMP event (via a massive solar storm) is a very real danger.

What probability does or would or might any citizen have of SURVIVING any nuclear war likely to occur given the planet's arsenals?

that's somewhat higher, than probability of nuclear war. consider scenarios:
1. North Korea - US - while millions people will be dead - most will survive
2. India - Pakistan - hundred millions will probably perish and possibly all people will be affected - but still many will survive
3 China - US the same as above
4 Russia US - good probability that few people will survive (total power of all active nuclear weapons is about the same a the eruption of Mount Tambora in 1815 and all nuclear weapons are about the same as Toba eruption 75 thousand years ago- still most of people in countries at war will die and many across the world)

"Millions of people"? North Korea's nukes aren't that powerful and have limited range. I'd put the death toll in the high hundred thousands, or very low one million range.

Millions of North Koreans live in Pyongyang alone, and most would die.

Who would be nuking Pyongyang?

Or are you suggesting civil war or something else?

Sergey's first scenario is a North Korea-US nuclear war. I presume the US would retaliate pretty harshly if NK nuked her.

You were suggesting the US would deliberately target the North Korean civilian population in a nuclear exchange! While we're estimating probabilities, I'd put that at near-zero. Wow.

The US went to great pains over the course of the Cold War to minimize the potential harm to the Soviet population in a large exchange, moving to counterforce, shrinking warheads, and downsizing the arsenal. The idea that they'd deliberately target NK civilians is preposterous, even assuming a successful attack on the US. There'd be no point.

So Kim nukes the US, a population center surely like LA or SF or Seattle....and we don't hit their only city back, where Kim and his regime are located? M'kay...

We didn't nuke Tokyo in WWII. You want the power structure to survive in order to perform an orderly surrender. Otherwise you're facing chaotic conditions that would make securing the defeated country- and its remaining nukes- very problematic.

Nor did we execute the German population, or even the army, after the fact. So even in WWII we attacked things when there was a good military justification to do so. We didn't punish or harm civilian populations without darned good reasons for doing so. There's no analogous reason in the hypothetical. It's just "revenge."

"4 Russia US - good probability that few people will survive"

Way off. Half the world, almost 4 billion people, live in rural areas that would hardly be affected. The most recent nuclear winter show a mild cooling effect and radiation isn't an issue away from the city.

"2. India - Pakistan - hundred millions will probably perish and possibly all people will be affected - but still many will survive"

This is also way off. India has around 100 warheads and Pakistan has maybe 100 as well. If each hit a city, around 25 million would die. Maybe more, but not more than 50 million. Not even close to 1.5 billion would be physically affected.

Civilization has a very high change of surviving a nuclear. At least 90%+, probably more like 99%+.

But the odds for "any citizen" are of course highly dependent. It depends on where you live, is your country involved, how big a war, etc.

A decent estimate is that if all the nuclear weapons were used on cities, then maybe 1 or 2 billion would die. But hey, that means 6 or 7 billion would survive! Even if 3 billion would die 5 billion would be still around to rebuild.

The U.S., Russia, and China are looked into MAD: mutual assured destruction. Those countries have too much to lose to risk nuclear war with each other.

The risk is small countries run by reckless madmen. I could see a nuclear strike by Iran against Israel, or a radical Pakistan against India, or North Korea against South Korea. Or a terrorist bomb ignited by extremist groups (see Tom Clancy's Sum of All Fears). Such exchanges would be awful, but limited. Global cancer rates would rise from fallout, but humanity would continue.

I suspect Israel would be more likely to attack another (non-nuclear) country with its nukes, than be the victim of such an attack. The fact that Israel, despite playing coy, is well armed with nukes itself operates as a deterent except against stateless terrorists.

There have already been 450 above ground nuclear tests. albeit speed over a couple of decades. The global cancer loading from fallout from a single warhead terrorist attack, or even a Pakistan vs. India 50 warhead war, seems fairly modest.

If the warheds are airburst fallout will be very minimal. Fallout is created when a nuke is exploded at or near ground level, as would be done if the target is hardened and especially if underground. Airbursts are preferred otherwise since they are more destructive. The two WWII nukes were both airburst and there was almost no fallout in either Hiroshima or Nagasaki, or downwind from them. (Radiation sickness in both cities being due to exposure to the prompt radiation flux of the detonations themselves)

This is just as silly exercise in pretending to be scientific about a singular and entirely human-behavior-dependent event, which has exactly zero to do with the past or even with current alignments.

These guys should stick to guessing when the Yellowstone Caldera is gong to erupt.

But but did you see the number of decimal points of precision? If that's not science, I don't know what is.

Given the known issues with experts and the over-weighting of unlikely but top-of-mind outcomes, what's the justification for utilizing those as opposed to just the GJP/superforecaster numbers?

Ostensibly this is where Wisdom of the Crowds comes in. While any particular estimate is likely to be wrong (perhaps greatly wrong), idiosyncratic errors average to zero. I'm not sure how this works when an estimate has a lower bound, but I think the central limit theorem applies.

Of course, if the forecasts have systematic error, the estimator will be biased.

The fact no one has used a nuclear weapon in war since 1945 (74 years) should indicate the probability is much less than 60% over an average lifetime. That hypothesis is soundly rejected. Note that since Nagasaki, more nations have acquired these weapons (as opposed to just one) and those nations have built many more than just two weapons. It appears that having used nuclear weapons in 1945 has made their use far less palatable and arguably far less necessary.

When else in history have you ever heard of a combatant seeking alternatives to using their most powerful weapon?

How often have their been international agreements limiting the use of a weapon? (Several, but not many in relative terms)

Fun with imaginary numbers

The highest probability of anything nuclear happening is that of a nuclear attack, if the above survey can be taken seriously.

Given the nature of the present leaderships in both the U.S. and Russia or N.Korea; in India and Pakistan, this is not possible at least among state actors, except perhaps Israel and Iran.

Trump though outstandingly nutty is still not nutty enough to order an attack. But the case with the Nut-in-Yahoo is quite different. Netanyahu a perfectly rational guy can miscalculate to pre-emptively attack.

Further a rogue attack by non-state actors using radiological weapons is also possible. Also with powers in the Nuclear Club routinely exercising with nukes, even a Broken Arrow episode is possible !

So long as America doesn't launch a false flag attack under that pretext, the world should remain safe I guess, despite the hightened probabilities !

The wild card in this assessment is that for as long as nuclear weapons exist, there remains a non-negligible risk that they will be used at some point in time.

Certainly, the principle of MAD (Mutually Assured Destruction) and various non-proliferation treaties have reduced the risk substantially. However, the recent introduction of nuclear weapons that have scalable yields from 0.01 kilotons to one kiloton and beyond, plus delivery systems that are pinpoint-accurate actually increase the chances that they will be used.

Variable yield nuclear weapons were developed during the 1950's and are common. For example, all UK nuclear weapons are variable yield.


Yet people continue to support China, Cuba, North Korea, Iran and other terrorists in the Middle East. Maybe we should figure out how to use our comparative advantage here.

I would think that world experience with poison gas would be instructive.

We saw widespread use in WWI. We saw it used a couple of times in the 30s (Ethiopia, The Rif War, and internally in the USSR). In WWII the Nazis made minimal use of gas weaponry on the Eastern Front to clear out a few bunker complexes. Japan made extensive use of chemical weaponry against China. Following WWII we saw most sporadic and frankly amateur use of chemical weapons in North Yemen, Rhodesia, and possibly Vietnam and Angola.

The only systemic use of chemical warfare after WWII was the Iran-Iraq war.

What we saw with poison gas was that an extremely effective weapon of mass destruction was basically shelved by all concerned after one brutal war where it is was used. After that, its use was sporadic and virtually always against opponents who could not retaliate in kind. The one time it was used systematically was one of those weird cases where the country in question had the backing of both major superpowers to one degree or another, was lead by a dictator with a God-complex, and was facing a technically deficient opponent who was using human wave tactics.

Syria rounds out the list.

This suggests to me that there is something which limits mass deployment of truly devastating weapons. The rate of chemical weapons use is falling over time and I see no reason why the rate of nuclear weapons use should not also fall in the long run.

Using the chemical weapons as a proxy suggests that every decade or so drops the rate of use by half. This suggest a much lower risk of nuclear warfare.

Late further query:

being no Bayesian, no statistician, and no mathematician, I nonetheless wonder how probabilities (can) accumulate over time.

Case in point: humanity would be remiss to permit this 21st century CE to elapse without throwing at least one world war, to judge by last century's enthusiasms. Now that we approach being a fifth or a quarter of the way through with no such outbreak, odds would seem to be increasing that one will in at least several likelihoods break out by mid-century, if not sooner.

Now that numerous nations, states, and countries possess atomic and nuclear arsenals, this century would seem poised for their use in the event of any global war (which unlike past events could be a small handful of nuclear exchanges among two or three pairs of opponents more or less simultaneously) which might suggest at least one further likelihood of a nuclear war (an exchange of nuclear weaponry, between at least two states) breaking out sometime within the next century.

To say nothing of accidents and opportunistic miscalculations.

Global recovery from such a conflict might pose far greater challenges than many may think or prefer to believe.

What justifies the adjective 'careful' before 'survey'?

I recently read an article where they looked at the accuracy of experts in predicting rare events. This is, of course, difficult. What they found is that if you look at experts in aggregate over multiple predictions, they are actually pretty accurate. That's pretty scary when the estimate is 1.17% likelihood of nuclear war.

I wish I could give you the citation, but it's lost.

Comments for this post are closed