Agreed. The most cogent justification for nuclear use I’ve seen involved the Russians suffering a massive battlefield defeat (ie another Kharkiv offensive) followed by the desire to “freeze” current battlefield lines.
The possibility of that kind of Ukrainian breakthrough seems extremely unlikely at this point. Yes, the Russians have pretty much used up most of their Soviet stock of APCs, tanks, artillery, etc. but Ukraine doesn’t have massive reserves of people and materiel either.
The main reason not to use nukes is simply that the Russian trading partners like China are strongly opposed. Never mind how quickly the Russian army, Air Force, etc. would cease to exist if NATO took a more active interest to eliminate same.
The overconfidence with which these models are used by some is striking indeed. The small number of observations problem is even worse in a way, since, thinking of WW2, we have 1 observation of use without having seen some of the actual effects, 1 observation of use having seen some of the effects, and 0 observations having seen most of the effects. Which seems psychologically relevant. Speaking of which, type of weapon and target should be relevant psychologically, too. Not to speak of how fundamentally different models of human decision-making predict different things, as do RAM vs. prospect theory.
Yeah absolutely, so many factors that are relevant. This of course theoretically introduces another issue in that your model loses any type of parsimony if you try to capture it all. But again, in practice this is not an issue because we are just unable to build this kind of model for lack of data.
Interesting article. However, I'm puzzled as to why the most commonly referenced earlier nuclear crises typically don't include the 1969 Sino-Soviet border war. True, it was an actual territorial dispute, unlike today's naked aggression, and involved two nuclear adversaries instead of today's one. But it resembles today's situation much more than Cuba or a misinterpreted NATO exercise. One side seemed prepared to use or at least threaten to use tactical nukes to both prevail on the battlefield and signal that an even more serious escalation was possible.
There are plenty of crises to choose from and the one you cite would be an entirely valid example as well. Personally, I know the literature on US-Soviet nuclear crises better, so I cited those.
I agree with almost all in the article, especially in responding with sceptisim about even ballpark estimates of risk... but nuclear risk is very real a
nd increases with escalation in the level of confrontation. Actions that increase confrontation raise the risk of nuclear use. Therefore, whilst there may indeed be policy influence from over confident numerical estimates of risk, it is critically valid to point to the nuclear risks as reasons to be very hesitant about escalation and to fsvour action that de-escalates.
Nuclear risks are absolutely real, that's why study them. I also disagree with the "Russia will never use nuclear weapons" crowd. There are scenarios were nuclear use absolutely is a possibility. Luckily, we can influence that with our own actions.
Agreed. The most cogent justification for nuclear use I’ve seen involved the Russians suffering a massive battlefield defeat (ie another Kharkiv offensive) followed by the desire to “freeze” current battlefield lines.
The possibility of that kind of Ukrainian breakthrough seems extremely unlikely at this point. Yes, the Russians have pretty much used up most of their Soviet stock of APCs, tanks, artillery, etc. but Ukraine doesn’t have massive reserves of people and materiel either.
The main reason not to use nukes is simply that the Russian trading partners like China are strongly opposed. Never mind how quickly the Russian army, Air Force, etc. would cease to exist if NATO took a more active interest to eliminate same.
The overconfidence with which these models are used by some is striking indeed. The small number of observations problem is even worse in a way, since, thinking of WW2, we have 1 observation of use without having seen some of the actual effects, 1 observation of use having seen some of the effects, and 0 observations having seen most of the effects. Which seems psychologically relevant. Speaking of which, type of weapon and target should be relevant psychologically, too. Not to speak of how fundamentally different models of human decision-making predict different things, as do RAM vs. prospect theory.
Yeah absolutely, so many factors that are relevant. This of course theoretically introduces another issue in that your model loses any type of parsimony if you try to capture it all. But again, in practice this is not an issue because we are just unable to build this kind of model for lack of data.
Interesting article. However, I'm puzzled as to why the most commonly referenced earlier nuclear crises typically don't include the 1969 Sino-Soviet border war. True, it was an actual territorial dispute, unlike today's naked aggression, and involved two nuclear adversaries instead of today's one. But it resembles today's situation much more than Cuba or a misinterpreted NATO exercise. One side seemed prepared to use or at least threaten to use tactical nukes to both prevail on the battlefield and signal that an even more serious escalation was possible.
There are plenty of crises to choose from and the one you cite would be an entirely valid example as well. Personally, I know the literature on US-Soviet nuclear crises better, so I cited those.
Thank you for the article. I have to say that also enjoy reading the comments too. 👍
Great to hear, many thanks!
I agree with almost all in the article, especially in responding with sceptisim about even ballpark estimates of risk... but nuclear risk is very real a
nd increases with escalation in the level of confrontation. Actions that increase confrontation raise the risk of nuclear use. Therefore, whilst there may indeed be policy influence from over confident numerical estimates of risk, it is critically valid to point to the nuclear risks as reasons to be very hesitant about escalation and to fsvour action that de-escalates.
Nuclear risks are absolutely real, that's why study them. I also disagree with the "Russia will never use nuclear weapons" crowd. There are scenarios were nuclear use absolutely is a possibility. Luckily, we can influence that with our own actions.