Some modest corrections to information provided above:
At its peak in 1964 to 1967, the US arsenal had about 30,000 warheads deployable by various means with a peak ICBM force of ~1100 missiles (LGM-30 F/G ‘Minuteman II/III’ and LGM-118A ‘Peacekeeper’) delivering a total of ~2500 warheads, and a peak fleet ballistic missile (FBM) force (Trident I/II) of ~6000 warheads. US ICBMs were all stored in underground silos. While they were “hardened” (i.e. equipped with systems to survive a seismic event such resulting from a near miss and protected against air burst effects) they were also dispersed and located away from urban areas. SLBMs were deployed on submarines (James Madison, Benjamin Franklin, and Ohio class ballistic missile submarines) which trolled large areas in the Atlantic and Pacific oceans on “deterrence patrols”.
At its peak between 1986 and 1988 the Soviet arsenal had about 35,000 warheads deployable by various means with a peak ICBM force delivering ~7000 warheads and SLBM force delivering ~3000 warheads. The Soviet ICBM arsenal was deployed in a combination of hardened silos, semi-hardened shelters, rail mobile, and road mobile, and was largely dispersed throughout the central and eastern Soviet Union. Soviet SLBMs were deployed on various submarines (NATO designations Yankee I/II/alt, Delta I/II/III/IV, and Typhoon).
The largest ICBM delivered warhead was the 20 MT warhead on the Soviet R-36M (NATO designation SS-18 ‘Satan’ Mod 3), of which the Soviet Union deployed 30. The largest ICBM delivered warhead in the American arsenal was the 9 MT W-53 warhead delivered by the LGM-25C ‘Titan II’, of which a maximum of 58 were deployed in silos.
The vast majority of warheads were sub-megaton weapons, and in general a disperse of smaller weapons in MRV or MIRV deployment scenarios is more effective than single large devices. Soviet missiles were not nearly as unreliable or inaccurate as portrayed in various fictions (re: Tom Clancy). While they did not demonstrate the standard of accuracy of American weapons, by the late 'Seventies they were sufficiently accurate to target individual installations, which is about as accurate as any nuclear device needs to be. Despite the accuracy, the vaunted disabling first strike which would render an opponent incapable of effective response was impractical from about the late 'Sixties onward; the ability to detect launches from orbit and incoming attack over the horizon, combined with the almost instant readiness of the ICBM fleets gave the expectation of complete parity in a nuclear exchange.
The deterrence theory of Assured Destruction (the “Mutual” was added later by critics like Herman Kahn to make an amusing acronym) was a game theory approach to deterrence that informed US, and to a much lesser degree, Soviet planning but the necessary prerequisites were never satisfied in practice (and arguably never could be in a real world scenario).
The reality of any limited nuclear exchange is that it would likely escalate to a strategic-level warfare with the destruction of both, and other, nations. Immediate deaths would likely be in the tens or hundreds of millions, with upwards of a billion people dying of disease and famine, and likely greater worldwide repercussions as the industrial base and transportation infrastructure upon which the majority of the world’s population depends collapses. Ultimately it would not be unreasonable to anticipate a loss of 2-4 billion people, and a return to pre- or early Industrial era population levels and technology in the near term (20 to 50 years). The hypothesis of atmospheric occlusion (‘nuclear winter’) by dust and soot particles as advanced in the TTAPS Report is somewhat overstated as it makes many simplifying assumptions, but would certainly cause near term impacts on the climate of the Northern Hemisphere. ‘Salting’ of arable land in the continental United States, Europe, and the Ukraine (the breadbaskets of the industrial nations) would likely cause longer term effects on the health and sustenance of surviving populations.
In a conventional European land war it is clear that the US and NATO powers have both technical and numerical superiority over Russian forces. Even back in the 'Seventies and ‘Eighties when the Soviet Union and Warsaw Pact forces had a clear numerical advantage of greater than 2:1 (more in some areas) simulations showed that NATO forces would prevail. However, this makes the assumption that Soviet forces would aggressively move into Central and Western Europe in expeditionary fashion. In truth, the Soviet and East Bloc forces never had the logistics to support any extended thrust and it is doubtful they ever would have attempted such. The function of the buildup of Warsaw Pact forces was to buffer any presumed invasion into Soviet territory (of which the Russian have a cultural fear, Russia and the Kievan Rus’ having been invaded repeatedly from time immemorial). Their strategy would almost have certainly been to draw in NATO forces into a long logistical chain and then attack the chain via naval power while waiting for time and weather to isolate forces, as they have also done again and again throughout history, most recently with Napoleon and again Germany in WWII. The NATO logistical chain, heavily dependent upon the industrial capability of the United States in that era, would have been very difficult to maintain over an extended period.
Had the Soviets truly felt in peril of their homeland they would likely have responded with tactical nuclear weapons using SRBM and IRBMs (which they had many) and a resulting exchange that very likely would have escalated as discussed above. While the whole world would not be made uninhabitable it would be a bad day for everyone involved. It is generally not a good idea to prod bears even if they appear slothful and diseased.
Stranger