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

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.

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...

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.

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’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.