93.16% Likelihood Of Nuclear War
You can’t calculate the odds. I did it anyway.
We were so close so many times, it seems statistically odd that it hasn’t happened yet.
How To Calculate It
Method 1 — Extrapolate Estimates From Surveys.
I found 5 (+historical data):
- Nuclear war 30% Nuclear terrorism 15% before 2100
Bostrom and Sandberg, Oxford — informal survey at the Global Catastrophic Risk Conference 2008 - Any nuclear attack 10% and 29.2% in 5 or 10 years
The Lugar survey on proliferation threats and responses 2005 - Nuclear war 4% in 50 years
Routledge Analyzing and Reducing the Risks of Inadvertent Nuclear War 2013 - Nuclear conflict 8.53% in 25 years
PS21 Great Power Conflict Report 2015 - Nuclear Attack in 2 years 1%
Good Judgment Inc. Superforecasts - Historical frequency of 1.4% per year
Historical data
Putting In The Numbers
You can’t just extrapolate “In the next 5 years” to 100 and then combine the results. Not to mention just how much needs to be omitted to even get there.
The Max Planck Society for the Advancement of Science estimates one nuclear disaster every 20 years with the same historic frequency.
However, it does not make sense to calculate this way. E.g. an interest rate of 50 percent can both be high and low. Per year it may be high, but per century it would be low.
Irony
The more near-accidents, the more it could be argued that everything is safe since no incidents of any kind ended in a nuclear war. Sadly, statistics and a catch-22 don’t go well together.
Method 2 — Probability tree
This is what happens when you try to simplify the wrong things. The probability of a nuclear war is not a coin-flip tree and if the nukes do go off, there won’t be a secondary outcome to work with either.
These [military] brass hats have one great advantage. If we … do what they want us to do, none of us will be alive later to tell them that they were wrong. -John F. Kennedy to his longtime aide Kenny O’Donnell.
Simultaneous Probabilities
Suppose a tea kettle had dual breaking circuits: each circuit can turn off the kettle and both circuits operate independently so that should one malfunction it does not impair the other.
If the probability of either one failing is 10%, the probability of both failing simultaneously is (0.1)² or 1/100. A third independent circuit would raise the probability of all circuits failing at the same time to 1/1000 and so on.
This could be applied to the various safety measures in nuclear power plants, political safety-net strategies, and so forth — except the formula will be near infinite.
There are simply way too many unknowns. Consider just the more known instances and you are left with hundreds of highly complex data points each of which brings their own lengthy story
Any attempt to draw this tree would leave you with either the words largest whiteboard or something like this:
So What Are The Actual Numbers?
Turns out we got none. It’s all assumptions.
As I dig through the footnotes and references, you notice a common trend: the numbers are all made up. A couple of dozen experts, an event crowd, book authors, weapons specialists, and your friendly neighbor — everybody takes their best guess. Given, some are more educated than others.
Any actual math happens on top of references to either the surveys or probabilities that don’t add up. The same holds true for the 100 Seconds To Midnight analogy. To give an example, scrapping the internet I came across the following table numerous times.
We can establish a base rate for the probability of nuclear war by looking at the number of times nuclear weapons have been used during a war: one time since they were developed 74 years ago. This could be interpreted to mean that the likelihood of nuclear war is around 1.4% per year.
Look each source up and what you will find is lengthy documents on consideration recommendations and X people think/assume/suggest/guess Y.
When you do find any math, it’s formulas where you are simply unable to fill in the unknowns.
John F. Kennedy & Cuba: Just don’t.
Kennedy estimated the odds to be one out of three… or even. He stubbornly only took actions he believed lessened the chances of nuclear war and went against all of his advisors and chiefs in doing so. Neither he nor his staff had any way of knowing the actual odds. Just that they were…bad.
Martin E. Hellmann, Stanford, describes the odds as inevitable. He argued that no matter how high or low they are, it would be but a matter of time. The logic is sound, and yet it isn’t.
Schrödinger’s cat.
The Night Club Analogy
It’s 1 am. You talked to 10 people since an hour ago and got 3 phone numbers. How many more will you get until 2 am? None, the club is closed. Say it wasn’t — the people you were interested in might not be around anymore. In addition, someone pulled the firealarm at 01:30.
Do we now calculate the number of fire alarms pulled per 90 minutes? That’s the problem with outliner data points and everything about nuclear accidents seems to be an outliner.
Wisdom Of The Crowd
Not all crowds are wise.
A way of calculating how new information about an event changes preexisting expectations.
I.e. asking groups of people if event A goes X, how does event B go?
These are the conditions versus nuclear:
- The crowd has a diversity of opinions.
-> Our crowds have similar opinions - One person’s opinion should remain independent of those around them and should not be under any influence.
-> Crowds and experts were asked as groups, unanimously and there is bias. - Anyone taking part in the crowd should be able to make their own opinion based on their individual knowledge.
-> This one checks out - The crowd should be able to aggregate individual opinions into one collective decision.
-> This one doesn’t
Conclusion
It’s all fun and games to crunch the numbers — but the reality is grim.
In favor of shortsighted goals and personal interests, politicians have been undermining denuclearisation efforts for years.
The population has but forgotten the dangers. Under the guise of modernization, new bombs are being built every day.
They are more precise and safe but don’t be fooled. It takes but one accident, one little misunderstanding for a literal extinction event to happen.
I’m but a humble statistician, so I will leave you with the clearest formula I can and the perhaps most interesting argument I’ve found.
X nukes = X chance, 0 nukes = 0 chance
If you want to but can’t beat the US in a conventional war — use nukes.
Except you can’t beat the US with nukes either, so you would need to go asymmetrical.
