Actually, it's more complicated than that, and that's the reason that the defense system was considered 'provocative'. It's also the reason the US and USSR arsenals were so 'over the top'.
(I read a book by someone involved with the so called 'nuclear calculus' of MAD a few years ago and assuming he wasn't lying through his teeth, it's interesting)
Let's say you want to nuke, say, Perth in Australia and remove it from the map. Without using the really really big ones, which were never deployed much really, you are talking about 6-10 mid 80s grade warheads. Let's say 10.
Now if you want to land 10 warheads on Perth, in the mid 80s, you need to plan to launch 18-25 or so at it.
The book went into the details of why.
Now because of some of those details, let's say that Australia deployed an ABM system that can stop 33% of the warheads that complete their ascent stage and separate from their missiles. We're not talking about shooting down the missiles themselves, just the warheads after they separate. (Interesting note, as of 80s grade tech, boosted fission weapons were fully 'fail deadly' and could detonate at full yield when struck by an interceptor weapon, before that weapon could destroy the hardware. Full Fusion weapons would probably 'fizzle' producing a much lower yield explosion than they were rated for.)
Based on his math, which was complex but did follow, assuming the underlying assumptions were correct, in order to turn Perth into a crater you now need to launch 60-80 warheads at it.
To get a 'for sure' 10 warhead kill.
Now when MIRVs were in style that doesn't seem like so much with a dozen warheads on each missile except that an iron clad rule was that those warheads each had to come off a separate missile. Because a lot of the reason for needing so many warheads was the assumption that a good percentage of those missiles carrying them would never make it to separation stage.
Add to this the fratacide problem of warheads. Any warhead hitting Perth within 'a short time (which he couldn't give exacts of because it was classified, but indicated it was longer than 3-4 minutes)' of any particular detonation would be killed by it's own brother explosion before it detonated. (And nuclear detonation waves were one of the few things fast enough to kill, for example, a boosted fission weapon before it could set itself off). So if you launched 60 warheads at Perth, not only do they all have to come from different missiles, but you have to plan for them to land over a at least a 4 hour period. Which allows the ABM system to be more effective because you can't swamp it with everything you have in one big go and, assuming Australia has deployed it's own nuclear weapons, also allows them to strike back at your missile launching fields and command and control facilities. Which means you need to target even MORE warheads at Perth if you want to evaporate it.
The Big Deal here is not that 'oh heck we may only lose half our country in a nuclear war woopie!'
The Big Provocative Deal here is that once you have that 33% kill shield in place it requires a massive expenditure of warheads on the enemies part to really for sure kill you completely. Suddenly things are not MAD and now you have to worry about 33% shield country launching conventional ground invasions of parts of your territories or spheres of influence, feeling more sure that you won't escalate the conflict to the nuclear stage because suddenly you can't ensure the destruction of the other side, when they still have the ability to annihilate you.
Now you may ask who would be insane enough to risk that nuclear war that wipes out only half their own country, given the rest of the situation, and my answer would Godwin the thread. Also, the USSR thought Reagan was that far off the rails as well. Who's to say who else would have risked such a level of brinksmanship.
90% would be enough for a country to act pretty much with impunity against anyone except the really big nuclear players, without fear of major nuclear damage. The thought of 5 or so nukes hitting the US is horrifying, but politicians and the military talk in phrases like 'acceptable losses' and 'can't make an omelet without breaking a few eggs'.
Which is why, in the end, the only solution to the nuclear problem is political.
(Mind you I'm summarizing a 300 page book here so give me a little bit of a break here).