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
Ultimately, I think what this question boils down to is whether to expect "a sample" or "a sample within which we live" (i.e. whether or not the anthropic argument applies). Under MWI, anthropics would be quite likely to hold. On the other hand, if there is only a single world, it would be quite unlikely to hold (as you not living is a possible outcome, whether you could observe it or not). In the former case, we've received no evidence that MAD works. In the latter, however, we have received such evidence.
I don't see what your reasoning is (and I find "anthropics would hold" to be ambiguous). Can you explain?
Suppose half the worlds adopt a strategy that is certain to avoid war, and half adopt one the has a 50% chance. Of the worlds without war, 2/3 have adopted a strategy that is certain to avoid war. Therefore, anyone in a world without war should have their confidence that they are in a world that has adopted a strategy that is certain to avoid war go from ½ to 2/3 upon seeing war fail to develop.