Is our continued existence evidence that Mutually Assured Destruction worked?

The stan­dard view of Mu­tu­ally As­sured Distruc­tion (MAD) is some­thing like:

Dur­ing the cold war the US and USSR had weapons ca­pa­ble of im­mense de­struc­tion, but no mat­ter how tense things got they never used them be­cause they knew how bad that would be. While MAD is a ter­rify­ing thing, it did work, this time.

Oc­ca­sion­ally peo­ple will re­ply with an ar­gu­ment like:

If any of sev­eral near-miss in­ci­dents had gone even slightly differ­ently, both sides would have launched their mis­siles and we wouldn’t be here to­day look­ing back. In a sense this was an ex­per­i­ment where the only out­come we could ob­serve was suc­cess: nukes would have meant no ob­servers, no nukes and we’re still here. So we don’t ac­tu­ally know how use­ful MAD was.

This is an an­thropic ar­gu­ment, an at­tempt to han­dle the bias that comes from a link be­tween out­comes and the num­ber of peo­ple who can ob­serve them. Imag­ine we were try­ing to figure out whether flip­ping “heads” was more likely than flip­ping “tails”, but there was a coin de­mon that kil­led ev­ery­one if “tails” came up. Either we would see “heads” flipped, or we would see noth­ing at all. We’re not able to sam­ple from the “tails: ev­ery­one-dies” wor­lds. Even if the de­mon re­sponds to tails by kil­ling ev­ery­one only 40% of the time, we’re still go­ing to over-sam­ple the happy-heads out­come.

Ap­ply­ing the an­thropic prin­ci­ple here, how­ever, re­quires that a failure of MAD re­ally would have kil­led ev­ery­one. While it would have kil­led billions, and made ma­jor parts of the world un­in­hab­it­able, still many peo­ple would have sur­vived. [1] How much would we have re­built? What would be the pop­u­la­tion now? If the cold war had gone hot and the US and USSR had fallen into wiping each other out, what would 2013 be like? Roughly, we’re over­sam­pling the no-nukes out­come by the ra­tio of our cur­rent pop­u­la­tion to the pop­u­la­tion there would have been in a yes-nukes out­come, and the less lop­sided that ra­tio is the more ev­i­dence that MAD did work af­ter all.

[1] For this wikipe­dia cites: The global health effects of nu­clear war (1982), Long-term wor­ld­wide effects of mul­ti­ple nu­clear-weapons deto­na­tions (1975). Some look­ing on­line also turns up an Ac­cel­er­at­ing Fu­ture blog post. I haven’t read them thor­oughly, and I don’t know much about the re­search here.

I also posted this on my blog