click on the image for a larger version
The trajectories for the bus/warhead/decoy complex (in red) and a cloud of chuff released just before burnou (in yellow)t.
(With apologies to Langston Hughes.) The Missile Defense Agency (MDA) might paraphrase Langston Hughes to say to potential adversaries “Chuff your rocket like you ought to do but [please] don’t let your rocket chuff [us].” It turns out that the MDA requires any potential incoming missile to behave extraordinarily well. Any deviation from absolutely perfect behavior could throw the missile defense into a dizzy. The latest example of this is the 31 January 2010 failure of the Ground-Based NMD interceptor to hit its target. One of two contributing reasons for the failure was a “chuffing” of the LV-2’s solid propellant stages, which apparently changed what the Sea-Base X-band radar sees enough to prevent it from adequately assisting the interceptor close with the target. (Or, perhaps, help it discriminate between the simulated warhead and the officially sanctioned decoys.)
That, of course, needs some explanation. Chuffing is pretty much what it sounds like; a type of puffing. Solid propellant motors (like liquid propellant engines) can suffer from instabilities in the combustion of their fuel and oxidizer. Chuffing involves a sudden extinguishing of the flame burning inside the rocket with a consequent loss of pressure, perhaps down to a vacuum if this happens at high altitudes. Gases released from the heated inner surfaces then spontaneously catch fire and the process of burning and snuffing out starts all over again. This can happen as much as one hundred times a second and it doesn’t take much imagination to believe that there is a lot of variation the amount of aluminum particles coming out the end during this process (aluminum is added to many types of solid-propellant rocket motors to increase the burn temperature and hence thrust).
Here is my guess on how chuffing completely changes what the SBX sees: The LV-2 target vehicles consists of the first two stages of a Trident C-4 missile a bus with the simulated warhead and the officially approved decoys. In my hypothesis, the second stage experiences chuffing near the end of its burn time, releasing excessive amounts of aluminum (in unburned chunks of propellant that broke off during the trauma of cyclic pressure drops). This cloud propellant chunks containing aluminum is, naturally, released with a lower velocity than the warhead and decoy bus because the bus/warhead/decoy complex continues to be accelerated. This difference in velocity, however, does not need to be very large if the chuffing lasted all the way up to burnout.
It turns out, however, that the cloud, traveling on its lower trajectory, actually appears to move ahead of the warhead because it is also on a lower trajectory. I have simulated this as shown in the diagram at the top of this post and, assuming that the SBX is some 1000 km West of Vandenberg, it could actually shield the warhead/decoy complex during the important period when the bus released the various bits and pieces.
click on the image for a larger version
You would think that an X-band radar could discriminate between chunks of aluminum and a warhead. That is certainly what Lincoln Labs thought in 1999 when they gave a briefing on just this issue. ( Here is their graph showing all potential sources of confusion, both “natural”, meaning stuff unavoidably produced during target vehicle flight, intentional countermeasures. Note II.b, the right diagram in the middle row, which shows fuel debris cloud as an important factor.) A little later in the briefing, they show a graph of measured “length” of the objects and their mean radar cross section (the right most diagram). The box, almost totally obscured by the “deployment hardware” square, shows that they expect the debris from chuffing to have very low length while the RV (and empty stages trailing along behind) have very large lengths. The battle management computers and filters on the individual sensors—including the SBX—would use these differences to discriminate between targets and decoys (and chunks of unburnt fuel). In this flight test, however, they weren’t able to.
Whether or not this is going to be a fundamental flaw in the whole missile defense development program remains to be seen. But remember that this is just a result of a fully functional target vehicle behaving in an “acceptable” manner. Based on MDA’s history, we might expect them to put increased attention to making sure the target booster does not chuff. And if they do attempt to solve the more general problem of discriminating between the nuclear tipped warhead and small bits of unburnt rocket fuel, the chances are they will come up with a new algorithm that is good only for chuffing from Trident C-4 stages.
We haven’t even started to think about what an aggressor might do to make life difficult for a missile defense system. Ted Postol has been working on similar issues and will be discussing his results in a forthcoming article. From what I’ve seen of his article, it’s going to prove very interesting. As for today, I think we have a pretty good idea of why chuffing contributed to the failure of the January 2010 NMD flight test.
Amy Butler’s article at Aviation Week (http://bit.ly/mda-test-fail) seems to indicate that SBX didn’t have its anti-chuffing filter active. That’s a pretty big screw-up (what happens if they forget to turn on the filter during real ops? the same thing!) and it was compounded by a failed connector on the kinetic kill vehicle.
I don’t think it’s anywhere near a proven fact that the BMD can’t see through chuffing, or that chuffing represents a show-stopper.
Also, your emphasis on the fact that the decoys are government-approved is disingenuous – the government is paying for the development and testing of an extraordinarily complex weapon system. Of course they’re going to vet the targets – otherwise their testing program would lack consistency and rigor. You imply that the government signing off on the targets makes them easier; I would suggest that without government approval, the government would have no assurance that their test proved anything. Aegis BMD has made fantastic strides biting off a little bit of the threat space with each additional test, and they do it with government-approved targets.
There are lots of problems with BMD (see above) but running expensive flight tests against ‘gotcha’ adversary targets designed to make the system fail isn’t going to fix any of them.
JR,
You are reading too much into my comment on “government approved” decoys. It’s simply a fact that they haven’t even started to think about adversary’s decoys. When they started out building this system, they planned on doing such tests solely on the computer. I do think that this unexpected occurrence raises serious questions about the fidelity of such Monte Carlo investigations.
Where did the KV end up in relation to the cuff clouds and the target plus decoy cloud?
How do I know the SBX wasn’t confused by the decoys and was confused by the chuff?
We need our own chuff-filter before trusting what MDA tells us about this test. Would they readily admit that even their known decoys fooled SBX?
I think your parenthesis is correct: “One of two contributing reasons for the failure was a “chuffing” of the LV-2’s solid propellant stages, which apparently changed what the Sea-Base X-band radar enough to prevent it from adequately assisting the interceptor close with the target. (Or, perhaps, help it discriminate between the simulated warhead and the officially sanctioned decoys.)”
Let’s hope the NK rockets are chuff-suppressed.
A revived CHOP program would go a long way toward establishing what needs to be done to address the countermeasure problem. It wouldn’t be entirely cheap, and it almost certainly wouldn’t be pleasant, but it would help a lot.
Not to rain on anyone’s parade, but the last paragraph of the article suggests that the EKV picked out the warhead from the accompanying debris, etc. If that is true, the target would have been destroyed even if the SBX wasn’t working. However, as the EKV suffered from a mechanical failure (a big problem in and of itself) there was no intercept.
I wouldn’t be so quick to use this as cast iron example of the GBI’s uselessness. The results from this test are decidedly more mixed. The sensor’s on the ground failed, but the sensor’s on the EKV appeared to have worked. But the EKV was broken.
Actually, after re-reading this blog posting, it appears that the author did not understand at all what was said in the Aviation week article. The SBX’s failed because its ‘anti-chuffing’ algorithm was turned off. Unless that algorithm was running, its not a legitimate test of the SBX’s ability to discriminate the target from the debris. You’re article was useful to explain what chuffing is, but in this test failure it is somewhat beside the point.
ArkadyRenko,
You are not raining on my parade. I would like to see a missile defense system that actually works . Unfortunately, too many supporters of missile defense don’t seem to actually care if it works or not. (Something I simply do not understand.)
Also unfortunately, the system will definitely need the SBX in the future even if it might have hit the target in January. It’s not me that is saying that, it is MDA. Natural debris from the upper stage just makes the problem that much harder.
And most unfortunately, it appears that the system—KKV, radars, etc.—really needs to know exactly what the target looks like. Every time we see a new example of this in the testing there are people who say “its a development program! They don’t have to get the environment right. Give them a break!” But I cannot help feeling that these people are not seeing the big picture that this “development” program is starting to show.
The basic assumptions, made many years ago as the Lincoln Labs briefing shows, is that a wide band radar can use a calculated length vs the radar cross section to discriminate between a warhead and the natural debris (not to mention boosters etc.):
I suspect–though I don’t have access to the test data–that the box representing natural debris is bigger than what is shown here and actually overlaps the square representing the “ideal” aggressor warhead. If true, that would mean that the basic assumptions NMD is based on are wrong. MDA needs to demonstrate that is not true and Congress should hold up funding until they do.
> the article suggests that the EKV picked out the warhead from the accompanying debris, etc.
Which, in the depressing spirit of half-empty, I’d take as an indication that MDA has yet to start using realistically challenging countermeasures.
Thank you ArkadyRenko — the article says: “Officials found that while the EKV was operating on incomplete data from SBX (which was confused by the threat scene), its onboard sensor managed to pick out the warhead—the actual target—from the threat scene.”
This could mean many things — what I believe it means is that there is something very fishy with this story.
If the SBX is directing the EKV to the threat scene and we are told by MDA that the chuff was the SBX’s issue, and Geoff’s analysis shows that the chuff and the real threat scene are spatially distinct, then why was the EKV viewing — with its limited FoV sensor — the threat scene and not the chuff region?
If the EKV was viewing the threat scene then the SBX “problem” was not a major contributer to the lack of intercept — as MDA claimed it was.
Why the brouhaha by MDA that SBX was at fault, when the EKV was correctly directed to the threat scence and apparently the miss was caused by the EKV failure?
From the start, I found the official story somewhat incredible, in the bad sense of the word.
The raytheon EKV has likely a 1-2 degree CCD for the endgame. That is the same solid angle as if you put a drinking straw to your eye.
If MDA is telling us that the EKV was guided well enough by the SBX to the threat scene to place the warhead within its small sensor field, then the SBX, apparently, did its job like FSB says.
If the chuff-filter was off that is no reason to implicate the SBX — some person should be fired but the SBX experienced no anomaly, and apparently guided the EKV just fine.
Where does the chuff fit in?
Will the MDA release details of what they say went wrong with the SBX?
What was the trajectory of the EKV through the final threat region?
Curious taxpayers await.
Yousaf,
Good questions but I have a different take on it. I would hypothesis that when the SBX failed to cue the KKV, MDA used its knowledge of the warhead’s trajectory to cue the KKV. (After all, as somebody above pointed out, these are development tests and you don’t want to waste taxpayer money simply because a vital component failed. You want to try to get as much information as possible out.)
So that separates out the SBX from the equation. As for a chuffing-filter “failing,” I think rather that the filter worked as planned but the background—ie the debris cloud formed by the chuffing— was much, much different from what they had expected. That is what I meant by the solid fuel “box” expanding out to include the expected warhead area on the graph above. Let me change that graph explicitly:
Right now, I am working on understanding how a debris cloud formed by the chuffing could fool the filter program into thinking there were multiple warheads or completely masking the warhead. I hope to have some more to say about that soon. But its not a question of the filter failing, the background simply looked very different than they expected. It could very easily be an insurmountable problem. If so, the whole concept of separating warheads from debris/countermeasures is called into question.
Geoff,
Thanks, that is helpful. I agree that the SBX is not needed to cue the EKV, and that this has been a recent addition to the tests.
But this is what the AWST story says:
“While the chuffing did not affect target performance, a problem arose when the primary sensor for the test, the massive Sea-Based X-band radar (SBX) managed by Boeing, had difficulty reading what MDA officials call the “threat scene.” This includes the rocket body, nose cone, countermeasures, warhead and any other objects that may be within view of the SBX during the test, according to the officials. Algorithms designed to help SBX “filter out” chuffing were not engaged during the test, one official says.”
In your hypothesis you say “I think rather that the filter worked as planned but the background—ie the debris cloud formed by the chuffing— was much, much different from what they had expected.”
So it is important that MDA clarify what the issue with SBX and chuffing was: was it really the case that a viable filter exists, was actually on, but did not work, (your view) or is it the case, as the one official quoted above says, that the filter was not even engaged.
Was it really the chuff that caused the SBX to hiccup, or would it have done so even in the presence of the countermeasures alone?
Was it the case that the SBX did not work properly, or that the countermeasures (plus or minus chuff) worked in fooling it, as many MDA critics have argued they could?
On the graph, I am not clear on what you mean by the solid fuel box expanding out to include the expected warhead area?
The chunks of chuff are in reality small and would occupy the region the Lincoln Labs people said they should. Are you saying that there are large pieces of chuff that people did not suspect, or that the fuel cloud masquerades as a large object?
If you have not already, it may be worth getting in touch w/ George Lewis about artifacts possibly showing up in reconstruction of the phased-array data, when the size of the chuff particles is on the order of the radar wavelength. X-band is 8-13 GHz or ~3cm. So 1/2 wavelength metallic dipoles (~1.5cm) would give a strong return signal. I am guessing that size is consistent with the bulk of the chuff. If a phased array radar is peering into a cloud of such particles it may well be that artifacts are realized in the reconstruction.
Yousaf,
You’re right that I have totally discounted the “turned off” theory.Its just too convenient and does not have the “fatal flaw in the system” connotation to have the ring of truth. Besides, its not really a filter; its the only way they can identify the warhead from the second stage or bus..
Geoff,
perhaps the issue is simpler than I thought: it may be that resolution of the SBX is high enough that the chuff particles simply overwhelm the tracking software — or, at least, slow it down.
If the SBX has to track ~100 “useless” chuff particles in addition to the warhead and decoys and rocket body etc. this may slow the processing and/or make for telemetry issues back and forth to the EKV.
So it may well be that anti-chuff software was off and the SBX essentially “hung-up” with too many particles to track.
However, on the downside, having such anti-chuff software could allow a smart enemy to use the filter to its advantage.
Yousaf,
At this point, its going to take my next post on this issue to fully explain my thinking. Without meaning to avoid your comments, Im going to ask you to wait until then.
To those more knowledgeable in RF physics than myself:
How would a cloud of tiny conducting particles (aluminium chuff) affect the radar signal if the average distance between particles in the cloud was on the order of the wavelength (or half-wavelength) of the signal?
(My basis for this question is the explanation once given for why microwaves don’t escape through the door window on microwave ovens: because electromagnetic radiation cannot pass through holes (in conductors) smaller than half their wavelength.)
Thanks,
Just a little addition from my side:
1.) Chuffing won’t be an issue with liquid-fueled missiles. Most north-korean and iranian long-range missile systems don’t use solid fuel (with the notable exception of the iranian Sejil – but that doesn’t have intercontinental range, most likely only about 2000km).
2.) According to my interpretation, an anti-missile shield is mostly a political weapon (for ‘bluffing’, not actual wartime use), adding another layer of ‘what if the other’s weapons work while mine don’t’-uncertainty to the MAD-game. So, while the one with the missile shield won’t launch a first-strike for fear of a) his own attack-weapons failing, b) the other one’s attack-weapons working perfectly and c) the own missile defenses failing, the other side won’t attack either since a) they have to realistically assume that their attack-weapons do not work 100% while b) the opponent’s missile shield eliminates the rest…
3.) It’s irrational to assume that any weapons system, be that offensive or defensive in nature, can be designed to work in 100% of the cases.
That’s exactly why we have denominations like C.E.P. (Circular Error Probability).
I must say that i’d definitely prefer a missile shield that works only with 50% probability (meaning that i’d have to launch at least two interceptors for every incoming warhead to be on the safe side) to no shield at all!
Just imagine what would happen in case of an accidental launch of just one single ICBM by somebody – a not completely improbable scenario, if you ask me (we’ve been disconcertingly close to this several times in the past!).
Of course, the higher the intercept probability, the more cost-effective such a system would be in case of use, but that is not what should be used as the primary criterion in this context in my eyes.
4.) I think that even with shrinking nuclear arsenals, it will be highly unlikely that any nation can (financially) afford to install a missile shield capable of reliably warding off an all-out attack by one of the larger nuclear-armed nations. On the other hand, most nations on this planet won’t be able to afford a nuclear arsenal large enough to credibly threaten a nation with a missile shield (even if that doesn’t work with 100% probability) – so ‘nuclear blackmailing’ by small players (like North Korea or Iran) can be elegantly invalidated (without the need to conquer that particular country…).
Ano,
Your question raises a dilemma for me. This is actually the subject of my next post on chuffing, which will discuss more about the interaction of radar with the debris from solid propellant missiles. Let me give a very brief overview here:
X-band radars are capable of actually imaging objects using what is called synthetic appature radar (or SAR). However, that takes time. In the stressed environment of a defensive missile engagement, the SBX probably does not image the warhead. Instead, it relies on the enhanced radar cross section of the warhead tip and its rear to determine the length of the warhead.
It turns out the debris from chuffing produces waves of particle density that can have a wavelength on the order of a meter (or several times more or less); about the size of warhead. That can (and I would argue does) cause considerable problems of identification and discrimination. Again, more about this later.
Jochen,
Patriot shot down zero SCUDs during the first Gulf War. That was because there was a conceptual flaw in its design: they didn’t imagine a warhead accelerating transversely by several tens of G’s. It looks like the current NMD system is suffering from a similar lack of imagination (perhaps purposefully willed). Chuffing doesn’t not prove that but it does prove that the system is extremely sensitive to changes in the threat scene. Other counter measures, which any country capable of building an ICBM should be able to employ, can stand that threat scene completely on its head. I realize that your country does not contribute anything toward the cost of NMD (why should it?) but mine does, and I prefer to pay for weapon systems that work, or certainly have a greater than zero chance of destroying the warhead.
Proponents of missile defense seem to always be so quick to dismiss any problems with the system because they say that any protection is better than none. This excuses the system’s developers from even trying. That is why the Bush administration moved the system out from under the Pentagon’s Test and Evaluation office, an office dedicated to making sure that weapon systems actually worked to some level, not to prevent its construction all togehter.
Jochen,
There is a large difference between missile defenses aimed at conventional weapons — which are sensible — and those aimed at nuclear tipped missiles, which are not.
The reason the latter are not sensible is that they do not alter the strategic equation between nuclear armed states, and may actually instill a false sense of security leading to conflict.
Further, they are expensive and wasteful, are ineffective and destabilizing.
For further details, please read my piece in Foreign Policy.
The issue is not just technical , but conceptual.
I remember watching a documentary about the performance of patriots against the SCUDs, where an MIT professor (his specialty was in control systems if I remember it correctly) who was closely related to the program was very much against the practicality of the very idea of the missile defense. Does anyone know the name of the gentleman?
No_name,
I would guess that you were thinking of Ted Postol but he was never closely related to the Patriot program.
Geoff,
I should have known you were already way ahead of me! I’ll wait patiently while looking forward to your next post.
Some more on missile defense’s technical issues.
“If decoys are clustered within a
kilometer or less of the warhead, the wide beam of the radar will not be able to
measure the angular position of the different objects with enough accuracy for the
kill vehicle to select the warhead using data from the radar – even if the EMR has
correctly identified the warhead. The kill vehicle will be able to observe the
angular position of decoys and warheads with very much greater precision than the
radar, but it will not have sufficiently accurate radar data to tell which of the
objects the radar has identified are warheads and which are decoys.2 In other
words, the kill vehicle does not have sufficiently accurate information from the
radar about the actual location of each object in the field of view of its infrared
sensor to associate, on a one-to-one basis, each object seen by the radar with each
object seen by the kill vehicle. Because of this “association problem,” the kill
vehicle must be able to discriminate between warheads and decoys using only its
own infrared sensor….The implications of this are profound…”
btw, Geoff, Ted did a lot of critical analysis of the Patriots actual performance during the 1st Gulf war.
Geoff,
It was quite a few years ago when I watched the program so I maybe mistaken. Thanks a lot for the info.
Geoff:
The main problem the PAC-2 apparently had with intercepting the S80/Al-Hussain (which is not to be mistaken for either the Scud-B, Scud-C or Scud-D!) was that the reentry speed of that missile was ~5-10% higher than the design specifications of the Patriot (which had the anti-missile capability added to protect West-Germany from Warsaw-Pact SS-21/Scarab, SS-1c/Scud-B and SS-23/Spider up to a maximum range of ~500km – it’s no coincidence that the particular Patriot-batteries ‘protecting’ Israel were fromerly stationed at western Germany). The iraqi S80/Al-Hussain was launched at Israel at a range of ~600km.
Additionally, as you correctly mentioned, the iraqi missiles were so badly designed that they spiraled wildly on reentry (which grossly degraded the accuracy of that missile – in reality, the Iraqis simply ‘sprayed and prayed’, since aimed shooting actually was impossible…) or even broke up. To counter this erratic behaviour (and to generally make the missile more maneuverable during intercept), the PAC-2 later received the GEM-upgrade.
As i wrote somewhere before, to put it bluntly, the Iraqis pretended to be able to hit Tel-Aviv, while the Americans pretended to be able to intercept those missiles (Actually, the three Patriots launched at each incoming Al-Hussain more than tripled the destructive effect on that city – what goes up must come down…at least if it stays sub-orbital like Spaceship-1…) – population-lulling propaganda at it’s best, and rather expensive and militarily ineffective, too.
I think it’s somewhat unfair to compare an ABM-system designed from the start to intercept ICBMs with an ad-hoc improvisation using anti-aircraft-missiles belatedly upgraded to TBM-interceptors out of their intended design-specifications…
And i fully agree with you, a hit-probability of more than zero is indeed desirable (otherwise, it’s only a ‘bluffing weapon’ for politicos…somehow useful nonetheless, but best not to be used or even relied on), but if i look at the testing results so far, NMD seems to already fulfill that criterion. It’s just not 100%, and it never will be. And we all should take great care to avoid the general impression that such a system will ever be able to protect the U.S. against an all-out first strike by the Russians (and possibly also the Chinese, or the French, or the Brits…) – such an illusion might indeed provocate the delusion of general invincibility leading to inconsiderated actions with terrible consequences for everybody on this planet.
Nonetheless, such a system will be worth it’s weight in diamonds (and i also take the silos and other infrastructure into consideration…) in case of a small-scale attack like an accidental or rogue launch, and will definitely invalidate any small-scale arsenal by second or third world countries (thus elegantly eliminating the threat of ‘nuclear blackmailing’ by such nations, since it’s obviously impossible to stop those from acquiring a small-scale nuclear arsenal – nobody wants to invade them after all the gulf-war Bush-bashing and North-Korea and Iran apparently won’t get stopped by pretty words or toothless UN-resolutions alone).
Just an example: a single, ‘accidentally’ launched SS-18/Satan with ten MIRV warheads in the 550kt-range will without doubt do tremendous damage to e.g. the US-east-coast. And if it’s neccessary to use 100 interceptors to stop those ten warheads, i’d say it would be fully worth the money. The alternatives to this are absolutely unacceptable in my eyes: simply taking the hit and then automatically destroy the rest of the world in revenge or simply accepting a lame excuse by the Russians for an ‘unintentionally’ devastated east coast (or whichever region would be hit) with millions of dead bodies.
yousaf:
In my eyes, it’s exactly the other way around: conventional missile intercept is expensive, wasteful and ineffective, while i see no alternative to at least trying to intercept a nuclear-tipped one.
Just look at the effect the conventional missile bombardment of cities like London (~1300 A4/V2 with 975kg-warheads), Antwerp (~1600 A4/V2) or Tel-Aviv (~24 S80/Al-Hussain with 450kg-warheads) actually had.
Now compare that to what happened to Hiroshima and Nagasaki. And those were relatively small bombs at only ~13 and ~20kt, respectively.
A destroyed city-block with hundreds of dead (and the Iraqis never even managed to do something like that!) may be a terrible tragedy, but is absolutely incomparable to the thermonuclear devastation of a whole multi-million population-center!
Also keep in mind that in case of a successful intercept debris will rain on the targeted city, doing some damage, too (all in all comparable to what an actual conventional warhead impact might have caused, only in a more concentrated form; case-shot instead of solid-slug, so to speak) – again, totally incomparable if a nuclear explosion over the targeted area has been avoided instead.
On the ‘destabilizing’ and ‘instill(ing) a false sense of security leading to conflict’ part:
May i remind you that the Russians already have their (ABM-treaty-compliant) A-135-system online since 1995? Following your logic, we should all be dead by now.
It could be said that the U.S. is at a disadvantage and that the strategic equation already has tipped in favor of the Russians (if i’d believe the ‘destabilizing’ part or that their ABM-system would actually work 100%…or even with only 50% probability…) – an ‘anti-missile-gap’, if you like.
With this in mind, i don’t see the value of the U.S. voluntarily staying completely vulnerable to attack by even the smallest nuclear nation.
Nonetheless, i think we can agree in one point: it’s absolutely vital to avoid the general impression in the american population (and political and military leadership) that NMD would render the U.S. totally unassailable. This would indeed be a rather stupid thing to assume, and may lead to rash actions and grave mistakes retroactively bemoaned by all humankind.
But NMD may (hopefully, or rather eventually?) be able to ward off a small-scale attack, like in an accidental or rogue launch scenario (how would something like that be handled under the current doctrine?), or by one of those 2nd or 3rd world countries like Iran, North Korea or Pakistan that can only afford a rather limited nuclear arsenal. Also, please consider that with the (obviously unstoppable, at least by non-military means) proliferation of nuclear and missile technology (most of that of russian/ex-soviet or chinese origin), it may sooner or later be impossible to please all of those nascent nuclear states (some, if not all, of them unfortunately governed by not overly rational or responsible individuals!) at the same time – not a very comforting thought, if there are nuclear weapons involved and there’s no option to stop those once they are launched (i hope i don’t have to fight the Hollywood-delusion of ‘recalling’ ICBMs after launch…).
On the other hand, if a (small-scale) nuclear attack, be that accidental, rogue or small-player, can be successfully parried, this opens up the opportunity for responses other than automatic retaliatory nuclear counter-strike.
I think this last, in my eyes very important point gets way too little consideration in the general discussion on those ‘evil’ and ‘destabilizing’ ABM-systems.