A nuclear war involving hundreds of detonations would almost certainly not by the result of a terrorist attack, as any terrorist group would find it difficult to acquire and deploy so many warheads and any country that suffered such an attack may not be able to pin point a physical target so vast as to justify deploying hundreds (or at least several dozens) of detonations.

As such, if total nuclear war were to materialize, this would likely be between states, including any current or future rouge state that has access to nuclear weapons. At present there are perhaps 6 or 7 states with a suspected arsenal of at least 100 warheads, and perhaps 2 or 3 states that might have the ability to ramp up is current arsenal to these numbers. Over the course of the next 3 decades, perhaps a few more states com also achieve sufficient nuclear strengths as to be considered in this analysis.

The sheer number of states that have, or might have, such nuclear capability represents a significant risk of the likelihood that a nuclear confrontation could happen in the next 30 years, and once a nuclear was has started, there might not be sufficient clarity and incentives for the parties to restrain and scale back from full blow nuclear war (in particular, for the country first hit a decision not to strike back with full force would imply accepting all the destruction inflicted by the attacker without a commitment to inflict at least, but more likely a greater - amount of pain on the adversary).

Current and future not foreseen geopolitical events and conflicts would increase the likelihood of any confrontation and, by extension, of nuclear confrontation.

On the other hand, one would hope and expect that (with the possible exception of rouge states), awareness of retaliation would keep any would-be nuclear attacker restrained for fear of its own annihilation. I suspect that this fear has probably played a significant role in some of the most tense geopolitical moments of the last 60 or 70 years.

Also, current and continued future economic integration (if materialized) means that the economic well being of any state is increasingly linked and dependent on multiple foreign factors that are not under the control of one particular state. It also means that disruption of supply chains, limited access to raw materials and hindered industrial and scientific capacity would imply that very few states could be in a position to sustain a prolonged war without allied support. Of course if one considers war between allied blocks, the likelihood of a full blown nuclear war increases, as the internal dynamics of each block render its actions more unpredictable and unstable. 

All considered, while there serious and significant risks that a nuclear war involving hundereds of detonations and devastating consequences might happen, I feel that the deterrent effect and the economic integration of the world might suffice in preventing such an outcome in the foreseeable future.

Notwithstanding the above considerations, below are three predictions on Metaculus regarding questions that if they were to happen would likely push humanity closer to the doomsday scenario of total nuclear war.

A deadly clash between the US and Russia would certainly imply a great risk of escalation, possibly even to nuclear escalation. The military rivalry between the US and Russia is long standing ans has recently found new hot spots in which tension is ripe. In particular, I suspect that the most likely triggers for such a confrontations in the short term would be consequence of a Russian attack on a NATO country (the 3 Baltic states come to mind) or a attack on a former soviet country, Ukraine being the case in point.

Here I would distinguish two scenarios. 

First I think that the likelihood of US involvement in a confrontation with Russia in defense of a NATO state should be quite high, although fore that same reason, I would consider the likelihood of a Russian attack on a NATO Country to be quite low.

On the other hand, the recent past has shown that Russia has been able to attack and seize territory in the former soviet states without much of a response from the US (even if the official stance is suggests otherwise). As such, I would suspect that the likelihood of a Russian attack of the former soviet states to be significant in the medium term, but the US likelihood of a US response to be small.

All considered, I would put the probability of a deadly clash involving gunfire or explosives between the US and Russia before 2024 at between 10% and 15%.

  


In the event of an offensive nuclear detonation by 2024, which I feel is unlikely, I would imagine that there would be more significant military targets than the battlefield (which might also inflict losses on the nuclear attackers forces). I suspect that targets such as the enemies own nuclear arsenal in order to try to limit the intensity of a nuclear counter stike would have top priority.

As such, I would put the probability of this question in the range o 2% to 10%.

  


If NATO fails to fulfill its perceived guarantee of mutual defense, or if one or more countries drift politically away from the majority of NATO members, then I would suspect that the likelihood of a NATO departure to be significant.

This being said, for NATO to be perceived in the short term as to not assure mutual defense, an extreme event for have to take place, such as failing to respond to an attack on a member state (an attack on one of the Baltic states by Russia come to mind as one possibility). Since I consider the likelihood of an attack on a NATO member to be very low, the likelihood of a non response by NATO in the fore mentioned event is even lower.

On the other hand, the possibility of some countries drifting away (politically) from core NATO member is a greater concern. Currently Turkey, Hungary and Poland seem to be the likeliest candidates. However, at least for now, the two European countries seem to be sufficiently anchored into NATO (not least for fear of a hypothetical Russian aggression) and while Turkey may yet abandon NATO, I would suspect the while it maintains expectations of someday joining the EU, this would not happen.

All considered, I would put the likelihood of any state leaving NATO in the next 3 years at a very low value, about 2%.