it seems to me that if the Higgs is offended about being observed and quantum-destroys the universes it is produced in
You seem to be conflating the "Higgs production does something funny to suppress the amplitude of possible histories leading to the event" hypothesis with the "LHC destroys the world with black holes/stable strangelets/vacuum energy state change/etc., so only futures with a broken LHC have humans around to observe" hypothesis. The net result is similar but the mechanism is different (unless it's equivalent in some weird quantum way--I don't pretend to be an expert on this stuff).
In either case, if each event has a chance of occurring during a particle collision, there must be some finite number of collisions beyond which some event such as "a fragment of Boltzmann Bread[0] appears in a critical component of the LHC" becomes more likely than "no world-ending event occurring in X collisions".
In particular, there's not a need for the universe to anthropomorphically "reach across" an extended range of time; under MWI assumptions, all the hypothetical LHC-stopping coincidences will happen anyway in some world of arbitrarily small likelihood, but if the LHC is unexpectedly omnicidal, only those unlikely universes will have future observers to look back and marvel at how strange it is.
"You seem to be conflating the "Higgs production does something funny to suppress the amplitude of possible histories leading to the event" hypothesis with the "LHC destroys the world with black holes/stable strangelets/vacuum energy state change/etc., so only futures with a broken LHC have humans around to observe" hypothesis."
No, I'm not. I'm really only talking about the suppressed amplitude case, as it is the only interesting thing. "Quantum suicide" is a direct, uninteresting, and hopefully untrue trivial consequence of MWI, and generalizing to civilizations is also trivial.
""a fragment of Boltzmann Bread[0] appears in a critical component of the LHC" becomes more likely than "no world-ending event occurring in X collisions"."
Ah, but you're getting caught up in the wrong formulation of the problem. The question is not "Is sticking a bit of bread in the wrong place going to prevent the Higgs from forming?", it is, "Given that the universe will attempt to prevent the Higgs from forming for the sake of argument, what is the most likely way in which it will manifest?"
Does sticking a bit of bread in the wrong place do the job? Yes. But it's a fantastically improbable manifestation of a quantum effect where merely "not forming the Higgs boson at time of collision" is way, way less improbable, what with the odds against a Higgs in the first place.
The probability of "bit o' quirky event plus a bit of bad design" is well in the domain of human experience. The probability of "this is the lowest-effort way the Universe could find to shut down the production of a Higgs" is absurdly small, when I can easily lay out "easier" things, including simply not making Higgs when the opportunity arises, which requires merely a couple of quantum zigs instead of zags within very small fractions of a second of the potential collision, rather than an enormously long chain of causation spanning human-perceivable time frames, all of which are unspeakably huge from the quantum world's viewpoint.
You seem to be conflating the "Higgs production does something funny to suppress the amplitude of possible histories leading to the event" hypothesis with the "LHC destroys the world with black holes/stable strangelets/vacuum energy state change/etc., so only futures with a broken LHC have humans around to observe" hypothesis. The net result is similar but the mechanism is different (unless it's equivalent in some weird quantum way--I don't pretend to be an expert on this stuff).
In either case, if each event has a chance of occurring during a particle collision, there must be some finite number of collisions beyond which some event such as "a fragment of Boltzmann Bread[0] appears in a critical component of the LHC" becomes more likely than "no world-ending event occurring in X collisions".
In particular, there's not a need for the universe to anthropomorphically "reach across" an extended range of time; under MWI assumptions, all the hypothetical LHC-stopping coincidences will happen anyway in some world of arbitrarily small likelihood, but if the LHC is unexpectedly omnicidal, only those unlikely universes will have future observers to look back and marvel at how strange it is.
[0] Cf. http://en.wikipedia.org/wiki/Boltzmann_brain