Banner
    Official Quantum Randi Challenge
    By Sascha Vongehr | June 21st 2011 12:10 AM | 45 comments | Print | E-mail | Track Comments
    About Sascha

    Dr. Sascha Vongehr [风洒沙] studied phil/math/chem/phys in Germany, obtained a BSc in theoretical physics (electro-mag) & MSc (stringtheory)...

    View Sascha's Profile

    The Quantum Randi Challenge (QRC), first introduced here, exists in order to stop the spread of pseudo-science by simply teaching quantum mechanics. Here is the official version of the challenge (also published here and partially in Annals of Physics 339: 81-88). [We are still looking for people who can help to turn it into a multiplayer internet app.*]


    What is a “Randi-type” challenge?

    A Randi-type challenge is one modeled after the James Randi Challenge. James offers USD 1000000 to anyone that can demonstrate paranormal abilities under laboratory conditions. A Randi-type challenge has the following characteristics:

    1) It cannot be overcome according to the laws of nature as observed with the scientific method.

    2) If the pseudo-scientists’ claims were correct, it could be easily overcome.

    3) Overcoming the challenge would give the pseudo-scientist great rewards fast.

    4) Judging whether the challenge has been met or not must not depend on peer-review or anything else that could be discredited as establishment conspiracy against the truth.


    Why is it Important?

    Educators can point to the bare existence of the challenge to immunize the public against pseudo-science.

    I) The challenge having not been overcome in spite of the large rewards (and the fact that it should be easily overcome if the pseudo-scientific claims were true) proves the claims wrong.

    II) Very important: The aim of pseudo-scientists (like “intelligent design” proponents for example) is mainly to let it appear as if there are doubts about an issue, as if there is a serious controversy among experts. Trying a serious communication with them never works and only backfires by supporting their deception, painting the picture of established science being in dispute. Whenever you can, you should refuse entering into rhetoric arguments that give pseudo-science a platform to promote itself, simply by pointing to a Randi-type challenge: “We do not discuss your claims until the challenge is overcome!”


    The Quantum Randi Challenge, hence forth QRC, challenges any pseudo-scientist who claims that quantum physics is not true and that quantum entanglement experiments can be explained by a classically realistic and locally causal model.


    Why is the QRC a Randi-type Challenge?

    It fulfills the four criteria 1-4 and has been shown effective regarding aspects I and II:

    1) John Bell has proven [1], via the Bell inequality, that by quantum physics predicted and also by experimental observation confirmed measurements cannot possibly arise within a locally real world. Here is a lay-person-accessible version of that proof.

    2) a) Any classically local model can on principle be simulated by today’s classical computers. In other words: That a computer can model a system is the very essence of that the system is classically real (everything depends on locally available, perhaps hidden data**).

    b) The experimentally confirmed observations that are to be reproduced are simple and well defined. The challenge is to reproduce only and nothing else but the behavior of this most simple setup which is well known to violate Bell’s inequality and thus shows nature to be fundamentally quantum physical.

    c) The program that needs to be written is already almost finished and published (see below). All that needs to be done is to modify the hidden variables and measurement codes in order to reflect the pseudo-scientist’s local model.

    3) Instant fame is certain (and likely a Nobel Prize in physics) for whoever modified the program so that the Bell equality is violated.

    4) The program, if written as a multi-player game and published on the net and if it violates the Bell inequality, would become famous via the internet without any chance for established physicists to prevent it.


    I) The QRC has already been successfully employed once to disprove a pseudo-scientific model (via this email exchange).

    II) Strictly scientific refutations of quantum pseudo-science indeed fueled a vicious cycle of ever more mumbo-jumbo that invites to be refuted again. This has indeed spread the doubt that pseudo-science always feeds on. The QRC has literally stopped such “discussions” here on Science2.0 and on some other internet discussion groups it has successfully avoided giving crackpots yet another stage.


    The Core QRC-Algorithm:

    The following picture depicts the single computer “getting started version” of the QRC, written using Mathematica5™. With this obviously very short program, the QRC is officially published! (Being published and widely known at all is the main power of a Randi-type challenge. Please help to make it known.)

    The unmodified program models the experiment described in Hidden Variable Madness. Running it seldom violates the Bell inequality. Anybody with a locally realistic model of quantum physics would be able to modify just a few lines in order to make the program violate the Bell inequality 99 times out of 100 (50% is not enough - I can do that myself - see below). The program above only uses the following subroutine for analysis:


    The following is an example for how anybody who claims to have a local model is supposed to modify the QRC-program. In this example, we assume that the person believes that every photon has a certain, hidden polarization vector, while each photon individually however behaves quantum mechanically when measured. As you can see, that the hidden variables are now a "real vector" does not make the program more complex, on the contrary.


    The following is an example for a modification that violates the Bell and the CHSH inequality 50% of the time:


    Lastly: If you think you can meet the challenge, do not contact me. The QRC is specifically designed in such a way that you can easily post it on your own blog for example, and that if it meets the challenge, it will become famous fast without established physicists being involved.

    ------------------------------

    [1] J.S. Bell, "On the Einstein-Podolsky-Rosen paradox," Physics, 1(3) 1964 pp. 195-200. Reprinted in J. S. Bell, Speakable and Unspeakable in Quantum Mechanics, 2nd ed., Cambridge: Cambridge University Press, 2004; S. M. Blinder, Introduction to Quantum Mechanics, Amsterdam: Elsevier, 2004 pp. 272-277.


    *If you want to help to turn the QRC into a multi-player game via Java-IDE/XML Sage or suchlike, you are very welcome to contact me, but please be sure you understand that it is not about challenging quantum physics or about making an awesome ‘app’. A “Randi type” challenge is an effective tool against pseudo-science. Please also read the linked articles first: "The programs are open source; everybody can check that the programs do not secretly establish an internet connection between Alice’s and Bob’s computers after the angles are chosen (in other words: there is no cheating via secret non-locality). If the Bell inequality is violated in this way, it will be a huge confirmation that is going to spread over the internet like a firestorm."

    **Finite memory is not an excuse, because all experimental observations have finite resolution, too! “Complex geometry” is not an excuse. All geometries and topologies (e.g. the SU(2) versus SO(3) double covering important to Fermions) can be simulated in virtual reality – there is no difference to a computer about whether it simulates our usual Euclidian three dimensional world or anything else. The words “simulation” or “virtual reality” or “model” are not an excuse, because the QRC, once it is a multi-player game as described in the remarks, already constitutes a real physical system itself (computers are physical systems(!)). If a modified program could violate Bell’s inequality, that would already be a classical physical system that violates Bell’s inequality and result in fame and a Nobel Prize!

    Update: Didactic part introduced and defended in:

    S. Vongehr: Exploring Inequality Violations By Classical Hidden Variables Numerically.” Annals of Physics 339: 81-88 (2013), www.sciencedirect.com/science/article/pii/S0003491613001863 Preprint version adds section on realisms and shows programs with output: arxiv.org/abs/1308.6752


    Sascha Vongehr's Science2.0 Articles by Topics

    Comments

    I like the idea. In fact, there was an exchange between Richard Gill and Luigi Accardi a while ago where Richard basically challebged Luigi to come up with a computer program similar to this (see http://arxiv.org/abs/quant-ph/0110137). The only suggestion I would make would be to switch to using the CHSH inequality rather than the original Bell setup. This is the one that is usually tested in experiment and it does not require perfect anticorrelation. Crackpots can always claim that perfect anticorrelation is never realized in experiments due to errors.

    To clarify a few things from your previous posts:
    - Joy Christian is not employed by Oxford university. He lives in Oxford and has some sort of college affiliation, but he isn't paid by them.
    - Joy Christian was invited to Perimeter Institute before he became cranky about Bell's theorem. His research was always unconventional, but not crackpot. No one had any idea that he intended to try to 'refute Bell's theorem' whilst he was there. In fact, since he was a Ph.D. student of Abner Shimony, the idea that he might have problems with Bell's theorem would have seemed laughable at that time. As far as I know, he has not been invited back to PI since.
    - I'm sure that your Quantum Crackpot Randi challenge would have a good chance of attracting funding from FQXi. It is exactly the sort of unconventional outreach project that they are looking for.

    vongehr
    Thank you very much for your suggestions Matt.

    Joy Christian still has his profile and the crank papers on the official Perimeter site (which states "... Wolfson College of the University of Oxford, where he has remained affiliated both with the College and the Physics Department of the University."), he uses his Oxford university email address, and his book that is about to come out is funded by FQXi.

    About CHSH: Randi challenges are more about the public understanding of science than about experts discussing the fine points of an issue. The CHSH is not so easily conveyed. Note that the fifth picture above shows the program that violates the CHSH as much as Bell.
    I asked about Bohmian Mechanics at "Quantum Entanglement Without Spooky Action At A Distance?" I am skeptical it can work, and it isn't local anyway, but maybe this is a sufficiently appropriate place to ask: how well can BM (in particular, the pilot wave) handle the challenges it must face?

    BTW, re FQXi: In order to find possible gems in unlikely places, FQXi will have to wade through all kinds of stuff. I was disappointed that my essay there, on how to test a key claim about decoherence (not usually considered testable as such) didn't score high in their contest. I think the judging process valued theoretical cleverness and "art" over empiricism.

    vongehr
    Never understood the appeal of BM. It is contrived, unnatural, worst kind of dualism possible with some particle that is connected to some "pilot" field without clarifying any natural reason why they should see each other at all, and gives nothing back in return for all this artificiality. Compared to Everett relativity, which is naturally expected from mature philosophy like modal realism, BM is a waste of time.

    About FQXi: They plainly select whatever is "nice" and appealing to the masses, with lots of distorted science history about the big males that we supposedly have to thank for all progress, and anything that would actually be cutting edge, let alone critical, does plainly not appeal, period. It is a bunch of insiders like in every other place. If you think to have actually a really pioneering idea that is natural only to people who can grasp the next paradigm, forget FQXi.
    han geurdes
    Dear Sacha, It is not quaternions I am talking about. It is not the famous computer program. Bell's theorem is no theorem at all. Its definition contains an ambiguity. That is the message. I may employ the definition of the Heaviside, I may define sgn(x)=2H(x)-1, I may use the equation sgn(x-y)=(x-y)/|x-y|, I may also use the equation |x-y|=(x-y)sgn(x-y) when differentiating |x-y|, I may differentiate the .left hand and right hand of sgn(x-y)=(x-y)/|x-y| to x, Hence the ambiguity. No need for quaternions, no need for difficult arguments. Bell was wrong from the very start. Good day
    Halliday
    Han:

    I just did a search for where you, supposedly, brought this up, that I must have missed.  Sure enough, rather than responding under my response to you, you wrote something at the top level of the discussion thread (just as you did here), so I didn't know it was directed at me.

    However, while you are correct that the Heaviside "function" (actually, distribution, just as the Dirac delta "function" is actually a distribution) has no definite value at zero (though there are versions that do choose definite values there), such is not the case with the signum function (sgn).

    The signum function is an odd function on the real line, and has the definite value of zero at x=0.  Furthermore, while the Heaviside function, in general, has no definite value at zero, one "popular" choice is the H1/2(x) function that has the value of 1/2 at x=0.  (So sgn(x) = 2H1/2(x)-1.)

    As for other "mystic math" aspects of your "argument" see my original response to you here.

    David

    P.S.  Unfortunately, Han, you are an excellent example of the sort of crackpot Sascha is trying to address in a far more productive way than exchanges on "mystic maths".
    Halliday
    Sascha:

    Your code is extremely fuzzy on what parts are to be computed independent of the particulars of the hidden variable scheme, and what are the particular portions that are to be hidden variable "aware".

    Of course I recognize that this particular hidden variable scheme is so utterly simple it seems truly ridiculous to concern oneself with such trivialities.  ;)

    However, perhaps it can help further discussion if we make such matters explicit, even in this utterly simple case.

    For instance, you generation code is
    n = 8000; Table[Hid[i, j] = If[Random[] < 0.5, 0, 1], {i, n}, {j, 3}];
    I can see at least two ways to factor out the internal hidden variable code:*
    1. The internal hidden variable generation code is allowed to generate all the"photon" pairs en masse:
    HiddenVariableGenerator[n_Integer] := Module[{i, j, iHidA}, Table[iHidA[i, j] = If[Random[] < 0.5, 0, 1], {i, n}, {j, 3}]; {iHidA, 1 - iHidA}];
      • So the outer code becomes:
    n = 8000; {HidA, HidB} = HiddenVariableGenerator[n];
    1. Or The internal hidden variable generation code is restricted to a pair-by-pair basis (note:  our code isn't actually doing anything with the "seed" like argument—it's just there for whatever use the hidden variable generation code may have for such):
    HiddenVariableGenerator[s_Symbol] := Module[{j, iHidA}, Table[iHidA[j] = If[Random[] < 0.5, 0, 1], {j, 3}]; s = s +1; {iHidA, 1 - iHidA}];
      • So the outer code then becomes:
    n = 8000; s = Random[]; {HidA, HidB} = Transpose[Table[HiddenVariableGenerator[s], {i, n}]];
    In both cases, HidA is to be sent to Alice's compute, while HidB is to be sent to Bob's.

    Similar choices can be made for the hidden variable detection code within the client code.

    David

    *  Please forgive me if I have made any syntactical or semantical errors in the Mathematica code.  I'm doing this from some rusty memory of the syntax and semantics.  Of course, feel free to apply corrections.
    vongehr
    I do not understand what you mean by outer and inner code. It must be as simple as possible and reflect the physics. The hidden physical world is supposed to know everything (1) and its only limitation is Einstein locality (2). That implies:
    (1) One simple code and all hidden variables are known to both photons at creation, so they are carried completely to both Alice and Bob.
    (2) Alice does not know Bob’s angles and vice versa. Everything else is explicitly allowed.
    Not sure why you are concerned with the triviality of the HD. They must be allowed to be as complex as desired.
    I start to suspect a misunderstanding. We publish the code and game in order to show that it is easily possible - that is all it needs to be a Randi-challenge. I am not willing (neither is Hank and his poor server capacity) to deal with hundreds of crackpots trying to put huge hidden variable codes onto our systems asking for technical support. They are supposed to modify what we did and publish the downloadable game on their own website. If it violates Bell, people will become interested, check the code for cheating, and the Nobel prize is theirs.
    Halliday
    Sascha:

    Have you never done modular programming?  I know you understand that there are aspects of this system of programs* that must not be allowed to be modified, while others are quite free to be modified as desired, since you have expressed some such limitations within the comments within your Mathematica code.  However, do you not recognize that enforcing and checking for allowed vs. disallowed modification is vastly facilitated by modularization?

    This is the principle reason for my specifying, in my original proposal, that the hidden variable aware portions of this system of programs be encapsulated within modules/subroutines/functions that are specified by (rigid) application programming interfaces (APIs):  The specification of the APIs determines the degrees of freedom within which any compliant code must work.

    As I tried to point out, within that proposal, there are two and only two modules/subroutines/functions that need to be hidden variable aware (given a sufficiently flexible encoding for any and all hidden variable schemes):  1) The hidden variable "generation" code that is responsible for generating the pairs of "photons", with all their (arbitrarily complex) hidden variables; and 2) the hidden variable "detection" code that takes in the hidden variables for the photons received by the particular client, along with that one client's chosen measurement angles, and is responsible for generating the resulting detection measurements.  All other aspects of the code are to be completely hidden variable "blind" (hence the reason they are referred to as hidden variables  ;)  ).

    So, what I refer to as the "outer" or "wrapper" code is all the machinery coded that is quite "blind" to the hidden variables, and should be completely unmodified for any given hidden variable implementation, while the "inner" code is the particular hidden variable aware "generation" or "detection" code that is the only code that is allowed to be modified for any particular hidden variable implementation.

    So, does that help clarify the situation?

    David

    *  We have the generation, two clients, and the correlation processes:  Four processes, three programs (since the two clients should be running copies of exactly the same program).
    vongehr
    enforcing and checking for allowed vs. disallowed modification is vastly facilitated by modularization?
    Yes, of course, that is why I wrote the different parts. But the crucial is keeping Bob and Alice apart, which means the important part is not the HV (sorry about the "HD" typo) or how they interact with the detection modules at each separate locality. Important is that one can easily check that there are never angle settings going between Alice and Bob (mediated secretly through the computer that creates the HV), except after all is over and done in order to present the results. Making sure there is no cheating communication should be easier if there are only a few communications, every time sending the whole 8000 HV or angles or results, not sending pair-by-pair having communication going on all the time.
    Halliday
    Sascha:

    You say:
    Making sure there is no cheating communication should be easier if there are only a few communications, every time sending the whole 8000 HV or angles or results, not sending pair-by-pair having communication going on all the time.
    I agree!  (If you are surprised, then you haven't "gotten it".  ;)  )

    If you really think I'm advocating any such "communication going on all the time", then please show me where I've got any such "communication going on all the time" within my rewrite #2, the one I say is an acceptable implementation, vs. rewrite #1.  (I knew that having concrete implementations would be helpful.)

    David
    vongehr
    You want the photon-by-photon enforced via the code rather than the communication, I know, but it needs to be as Randi-type as possible. People may understand that the system is real because it is done on real physical systems (computers) and the locality and all that is enforced via the computers not communicating at crucial times. If it is just slightly too much about whether something is coded right or wrong, it won't be Randi-type anymore. You having a little bit of the same problems as Chantal. This is not about physics (the physics has been settled decades ago), this is about public understanding of science.
    Halliday
    Sascha:

    I do understand the desire for "public understanding about science".  I also understand about the great benefits—even appeal—of this computer system being a physical realization of an EPR-like system, and the things that can imply.  However, I also understand the dangers if we allow ourselves to open a loophole (quite analogous to the communication loophole) that can allow a clever skeptic to "slip through the cracks".

    Now, I have provided an example Application Programming Interface (API:  in the code I sent you earlier today) that tries to close this loophole, and even shows how simple such truly is.  However, if the restrictions required within that API are not acceptable, then we will need to go to a more physical realization.  One more like the real EPR like experiments this is supposed to be a realization of.  Then the only remaining recourse will be to go to your "photon"-by-"photon" communication, where we also allow for the result of prior measurements to be fed back to the hidden variable code, also on a "photon"-by-"photon" basis.

    Lots of communication, but, for a sufficiently low frequency of entangled pair production (as is usually the case in present experiments), a quite realistic scenario.

    (By the way, provided one programs the network packet passing correctly, this does not require any more reliable of network connections than people have had under dial-up and/or WiFi.  I've only seen Microsoft's network protocols have any real problems with such a lack of network reliability.  In fact, under worst case senarios, we can allow packets to be dropped, and we will simply be dealing with less than 100% "photon" detection.  In any case, our network communication needs will still be small compared to something like SecondLife.)

    So, which will it be:  Door number 1?  Or door number 2?  (Or you can leave all of physics open to a potential loophole.  One that is no violation of the rules of the "game" because we didn't impose the proper restrictions from the beginning, yet one that left us unprotected because it is outside the criteria of Bell's Theorem [that is, of course, unless you know of a version of Bell's Theorem, with an accompanying proof, of course, that closes this potential loophole].)

    After all, if "this is about public understanding of science", then this code had better be about the science, and not some "fantasy" reality that allows communication back in time.

    David
    vongehr
    Of course, there should not be a too obvious loophole - though even if there are none, I promise you there will be crackpots insisting on conspiracy via the detection loophole and even some who will simply add the equivalent of Print:"Bell is violated!" into the coincidence counter code and post the game on their website. Such programs may get some attention if it is done carefully, but they will not survive the audit by skeptics and get a Nobel Prize.
    So is it door 1 or 2? Let's see whether I can convince you in a fresh new way: Put Alice on Venus and Bob on Jupiter and create all the photon pairs closely together in time here on earth and let them both determine their angles all at once before any of the photons arrive, and let the photons circle around the set of 8000 crystals twice inside a glass fiber before measuring them while only assuring that the measurements are over before Bob's angle choices' future light cone reaches Alice! Do you think that would change anything?
    Halliday
    Sascha:

    You ask:
    Put Alice on Venus and Bob on Jupiter and create all the photon pairs closely together in time here on earth and let them both determine their angles all at once before any of the photons arrive, and let the photons circle around the set of 8000 crystals twice inside a glass fiber before measuring them while only assuring that the measurements are over before Bob's angle choices' future light cone reaches Alice! Do you think that would change anything?
    First, you don't indicate whether the photon pairs are Quantum Entangled, or whether we are talking about some "classical" hidden variable case.  (For the Quantum Entangled case, I wouldn't expect that this scenario would change anything.)

    Second, for the "classical" hidden variable photon pair case, I cannot say one way or the other, since Bell's Theorem does not cover this possibility.  (See my discussion below.)

    Third, it is not just a matter of whether "Bob's angle choices' future light cone reaches Alice" (and vice versa), since even having the pair production be within the future light cones provides a sufficient loophole.  (Of course, I assume this is why you have all the photons in flight before angle selection.)

    (Technically, even the use of "glass fiber", in order to artificially lengthen the time delay between pair production and observation provides for a communication loophole.  However, we have usually made soma allowances for such, until such a loophole can also be closed.)

    Now, if you were to slightly reformulate this as:
    Put Alice on/near Saturn and Bod on/near Jupiter, when these two planets are at opposition, with respect to the Sun; create all the photon pairs closely together in time here on Earth; and let them both determine their angles all at once before any of the photons arrive (but after all the photons have left Earth); and have all the measurements complete before Bob's angle choices' future light cone reaches Alice (and vice versa).
    Then, at least, you would have a very good multi-photon-pair experiment.  Of course, for the "classical" hidden variable photon case, my answer cannot change:  Bell's Theorem doesn't fully apply.  However, my "gut" expectation is that an extended version of Bell's Theorem could be formulated that could cover this case.  (Actually, probably the simplest way to address this would be to extend Bell's Theorem to far more complex states than the two polarization states of the present theorem.  Then one would simply reduce the result based upon the nature of the successive polarization measurements.)

    David
    Halliday
    Sascha:

    I am fully aware of the freedoms allowed and limitations placed upon the hidden variables.

    For instance, it is the violation of Einstein locality that prompts my concern with permitting an application programming interface (API) that would provide the hidden variable code the freedom to process all "photon" pairs in an all-at-once manner, like my rewrite #1, above (and the corresponding all-at-once implementation for the "detection" code).

    It is certainly permitted, under Einstein locality to retain any and all information about past events within a given event's past light cone.  However, it is most certainly a violation of Einstein locality to have information about future events, such as future hidden variable "photons".  This is especially the case if one allows the knowledge of future angle choices, even of only the single user's choices.

    Does this help clarify things, somewhat?

    Unfortunately, I'm not at all certain what you mean by "HD" within your comment of "Not sure why you are concerned with the triviality of the HD."  So I'm afraid I cannot address that, unless you meant HV, for hidden variables.  I believe I have attempted to address the latter.

    David

    P.S.  I do realize that there is nothing, really, preventing the "photon" pair "generation" "device" (code, in our case) from predetermining the hidden variables of all "photon" pairs ahead of time (especially when we have a predetermined, fixed number of such).  However, the problem is when this is coupled with allowing, or, worse, even requiring the "detection" "device" to have the entirety of this information at it's disposal when determining the measurement results—especially if coupled with having access to the entirety of angle choices, on the client (to say nothing of having access to the other client's angle choices, which none of us are advocating).

    Actually, even requiring the "generation" code to comply with with a one-pair-at-a-time API/protocol does not disallow an implementation from performing such a "predetermination".  However, such will be more explicit, within the code.

    Furthermore, again even requiring the "detection" code to comply with a one-"photon"-and-one-angle-at-a-time API/protocol does not disallow an implementation from communicating the totality of all hidden variables within the hidden variables of the first "photon".  However, such will still be more explicit, within the code.

    The real issue that requiring the "generation" code to comply with a one-pair-at-a-time API/protocol, along with requiring the "detection" code to comply with a one-"photon"-and-one-angle-at-a-time API/protocol does address is prevention of the "detection" code from having knowledge of future choices in angle measurements when determining measurement results.

    This addresses the same kind of loophole that choosing angles "in flight" provides in real experiments.

    (Note:  If one does not require the "generation" code to comply with a one-pair-at-a-time API/protocol, one has a very difficult, if not impossible task if one wishes to require the "detection" code to comply with a one-"photon"-and-one-angle-at-a-time API/protocol.)

    So, does this help clarify things, somewhat?
    vongehr
    However, it is most certainly a violation of Einstein locality to have information about future events, such as future hidden variable "photons".  This is especially the case if one allows the knowledge of future angle choices, even of only the single user's choices. ... This addresses the same kind of loophole that choosing angles "in flight" provides in real experiments.
    You claim that violating Einstein locality at Bob's place (and Alice's) may violate Bell's inequality, which is based on enforcing that Alice cannot know Bob's angles. One of us has a blind spot, I hope it is not me. ;-)
    Halliday
    Sascha:

    What I have always been saying is that such violates the conditions of Bell's Theorem.  As I expect you are well aware, if one violates the conditions of a theorem, one has no result—the theorem and its proof do not apply.  This is not the same thing as trying to assert that such a violation necessarily leads to a violation of Bell's inequality.

    Since Bell's inequality is part of the "then ..." clause of Bell's Theorem, the violation of the "If ..." clause of the theorem simply means that one cannot determine whether the "then ..." clause applies, or is violated:  You have no guarantees one way or the other (since Bell's Theorem is not of the "If and only if ..." variety).

    Furthermore, both the conditions of Bell's Theorem and Einstein's concept of locality preclude more than just Alice and Bob communication during the measurements.  I thought you knew that.  They also preclude knowledge of future events.

    Now, I'm not absolutely certain, but I would expect that the proof(s) of Bell's Theorem depend upon this condition as well, even though it may not be explicit.  I certainly have seen no proof that permits a violation of this aspect of Einstein locality.  Besides, if there were such a proof, one could actually expand the theorem to allow for this aspect of Einstein non-locality, thus expanding the reach of the theorem, but I most certainly have never even heard rumors of such, except from your most recent assertions that it only precludes Alice and Bob communication during the measurements.

    If your assertion were the case, then those that argue that violations of Bell's inequality implies "faster than light communication" would certainly be strengthened.  However, I don't believe that is the case (and I wouldn't think you would want to be responsible for adding to such a misconception).

    David
    vongehr
    Violation of Bell needs Alice to know Bob's angles. That and only that is the Einstein locality of interest here. If you like to make that a new theoretical insight (I strongly doubt it is), that is OK, but a Randi challenge is about stuff that lay people can kind of follow and so it is just about whether Bell is or is not violated by any of the programs out there, which will not happen if there is no violation of Einstein locality that lets Alice know Bob's angles. I can live with Alice knowing her whole future light cone up until it starts intersecting with Bob's. That this somehow in logic would disqualify (If-then clauses) somebody's conclusions that are to be drawn from that specific model of HV is entirely beside the point. On that level, Bell has been proven, period, no Randi challenge necessary.
    If I am wrong, and I may well be, please tell me where in the proof is it necessary that Alice does not know the angles she selects in the future for example?
    Halliday
    Sasach:

    First, you assert:
    Violation of Bell needs Alice to know Bob's angles.
    Actually, I have absolutely never seen a formulation of Bell's Theorem that says anything like this.  All formulations of Bell's Theorem I have seen are of the "If the two (or three) conditions are met, then Bell's inequality (or something equivalent) is satisfied."  As I would expect you to be fully aware, if any of the conditions are violated, then one may or may not violate Bell's inequality.  All we can know for certain is if Bell's inequality is violated then one or more of the conditions of Bell's Theorem must have been violated.*

    Second, as for the violation of Einstein locality via knowledge of future angles (but not involving the other observer's angles, of course), and whether that actually constitutes a violation of Bell's Theorem (as in the conditions thereof):
    1. Have you ever seen a formulation of Bell's Theorem that explicitly permits this future information?
    2. Have you ever seen a proof of Bell's Theorem that explicitly permits this information to be "fed into" the determination of photon measurements?
    3. Are not the statistical computations that feed into Bell's inequality computed based upon an assumption of statistical independence of each photon (or photon pair) measurement?
    Unless the answer to #3 is No or Wrong or something similar, and the answer to at least one of #1 and #2 Yes,** then I rest my case.

    This is most certainly not a "new theoretical insight".  This "theoretical insight" is as old as Bell's Theorem itself.  It is your assertion that would be a "new theoretical insight" that would require a new theorem with an accompanying proof (unless, of course, one such already exists, and I just haven't seen it yet—hence the questions above).

    David

    *  Of course, if what you meant was that assuming realism (so we have hidden variables that determine the outcome of measurements, as we do within this challenge), then the only way to "Violate Bell" is to violate the other condition(s) of Bell's Theorem (such as Einstein locality), then you are correct.  Of course, that still rests upon how the other condition(s) of Bell's Theorem (such as Einstein locality) apply.  (The second part of this post.)

    **  Of course, one cannot truly get a Yes for #1 without having a corresponding Yes for #2—similarly, if one has a Yes for #2, then one can reformulate the corresponding Theorem so as to obtain a Yes for #1 as well.
    vongehr
    never seen a formulation of Bell's Theorem that says anything like this.  All formulations of Bell's Theorem I have seen are of the "If the two (or three) conditions are met, then Bell's inequality (or something equivalent) is satisfied."
    There is more to a proof than stating conditions and pulling a conclusion out of the hat. Einstein locality is applied at certain points inside the derivation. It is not applied to stop Alice knowing her future angles for example (which would not make much of a difference as she knows her past ones already - if their was a trick, future knowledge would at most double the impact of the trick). Einstein locality prevents Bob's part of the hidden variables to know Alice's angles and vice versa. That is what stops the correlations between their measurements from violating Bell-type inequalities.

    Explicitly allowing (points 1 and 2) is unnecessary. Why must a proof explicitly allow everything it does not explicitly exclude? That would render every proof impossibly large. The assumption of statistical independence I think is also not necessary. There is the fair sampling assumption in the analysis of experiments, but I refuse to go down the rabbit "detection loophole" argument (you will never shut up the crack pots - they will find reasons to blather on endlessly no matter what you do).
    if what you meant was that assuming realism (so we have hidden variables that determine the outcome of measurements, as we do within this challenge), then the only way to "Violate Bell" is to ...
    The Randi-type challenge aims to educate the public about that those who claim Local Realistic Models (thus publicly arguing that modern science is all wrong) are pseudo-scientists.
    Halliday
    Sascha:

    You correctly state:
    There is more to a proof than stating conditions and pulling a conclusion out of the hat.
    Well duh!  ;)

    Of course, I recognize that probably the majority of the readers, even here on Science 2.0, have probably had very little experience with Theorems and Proofs.  I can't even be certainly whether a majority of Physicists have much more than a passing familiarity with such matters (it doesn't seem to get much attention within Science).

    I would expect (though I most certainly cannot guarantee it) that you would know that while a theorem is typically expressed in an "If A Then B" manner, any proof of such a theorem must address all the conditions of "A", in its "derivation" of "B".  (Of course, not all proofs are of a "constructive" form, like a "derivation".  However, all of the proofs of Bell's Theorem I have seen, are, indeed, of the "constructive" variety.  This, of course, is more like what we Physicists expect or prefer to see.)

    Now, when I used the term "explicit", I was trying to emphasize the need for the conditions ("A") of both the Theorem and the Proof to address the weakening of Einstein locality you have advocated, above.  Perhaps using the term "explicit" is too strong.  However, just because some interpretation of some condition, like Einstein locality, allows for some such cases, does not, by itself, indicate that such cases are or were permitted within a given statement of the Theorem, especially if such were not addressed—unequivocally—within an accompanying Proof.

    This is one of the reasons I always prefer to see a theorem accompanied by a proof.  It is not that I am some pedantic Mathematician.  It's just that the accompanying proof often helps clarify the meaning of the conditions ("A") of the Theorem.

    You ask:
    Why must a proof explicitly allow everything it does not explicitly exclude?
    Now, I would hope that you recognize that unless the proof of a theorem, unequivocally addresses all the conditions ("A") of the theorem, including any edge and corner cases, then the theorem has not been proven.  Instead, only a more narrow form, with the conditions narrowed to what was actually addressed, has actually been proven.  So, unless "a proof explicitly allow[s] everything it [the theorem] does not explicitly exclude", then one has not addressed all the conditions.

    It's as simple and straightforward as that.

    While you may wish to believe that Einstein locality, at least as used within Bell's Theorem, and as embodied within any associated Proof thereof, does not preclude knowledge of all future angle choices (of the local observer), such belief, alone, does not make the Theorem apply in such cases.  This would most certainly qualify as an edge and/or corner case for Bell's Theorem.

    Have you ever seen an instance of Bell's Theorem that unequivocally includes such additional information?  Have you ever seen any Proof that unequivocally includes such?

    Have not all versions of the Proof you have ever seen treated each "photon" pair, along with the accompanying pair of local angle choices, as independent of all other pairs and angle choices?

    In truth, even allowing for the accumulation of hidden variables, and the accumulation of local angle choices, can violate statistical independence (thus leading to auto-correlations:  A problem that can plague even completely classical measurements of successive local events).  However, as I recall, such is usually addressed within the Proof by reformulating such as being simply a part of the hidden variables of the "photons" and/or the detection apparatus.

    I have never, ever seen even the allowance for future local angle choices to be included within such.  After all, if future local angle choices are known at the time of the individual measurements, then what is to prevent such being known back at the "photon" pair generation apparatus before some future set of "photon" pairs are sent off to "Alice" and "Bob", which we already know provides for a sufficient loophole.  (This would require your previously mentioned "all photons in flight before angles are chosen" scenario in order to plug this loophole.)

    Now, whether extending the conditions of Bell's Theorem to include such local future angle information will lead to a violation of Bell's inequality, I cannot determine.  Violating "A" does not tell you whether "B" will be violated.

    Instead, it would require an extended proof, if one can indeed formulate such, in order to have a Bell Theorem that can accommodate such an extension to the conditions while still maintaining the Bell inequality.

    David
    vongehr
    Einstein locality is applied inside the proof to preclude that Alice's and Bob's angle choices are known to the photon pair creation (or, which is kind of equivalent, to the other guy's measurement). It is not some adding of v<c in strange ways that may turn around and bite you. The Bell proof knows nothing about relativity, all it uses is that certain stuff cannot depend on the angles that some place else are chosen, the motivation being Einstein locality. That is it. So your "A" isn't even Einstein locality. Bell cannot be violated by having the Lorentz contraction at Bob's place being twice the one Einstein predicts.
    This is a stupid question, but how can a compter simulation of a real world experiment justify certainty that such experimental results couldn't happen in the real world?

    And if you aren't disagreeing with real experiments, but computer simulations then is Randi's Challenge applicable?

    vongehr
    **Finite memory is not an excuse, because all experimental observations have finite resolution, too! “Complex geometry” is not an excuse. All geometries and topologies (e.g. the SU(2) versus SO(3) double covering important to Fermions) can be simulated in virtual reality – there is no difference to a computer about whether it simulates our usual Euclidian three dimensional world or anything else. The words “simulation” or “virtual reality” or “model” are not an excuse, because the QRC, once it is a multi-player game as described in the remarks, already constitutes a real physical system itself (computers are physical systems(!)). If a modified program could violate Bell’s inequality, that would already be a classical physical system that violates Bell’s inequality and result in fame and a Nobel Prize!
    Halliday
    Benjammin:

    As Sascha has already addressed, a system of computers, as stipulated within the challenge, is also an actual physical system.  So, even such a computer simulation can constitute a real experiment.

    As to your first question:  "how can a compter simulation of a real world experiment justify certainty that such experimental results couldn't happen in the real world?"  Unfortunately, I'm not entirely certain what "experimental results" you are suggesting there is "certainty that such ... couldn't happen in the real world".  So I'm even less certain why you appear to think that such "certainty" is derived from "computer simulation of a real world experiment".

    Now, my guess is that the "computer simulation of a real world experiment" is referring to the computer simulation stipulated within the challenge.  Since the "computer simulation" pertains to a certain class of EPR like "real world experiment[s]" that are covered by Bell's Theorem, the "certainty" has nothing to do with any "computer simulations" but the pure mathematics of Bell's Theorem and its accompanying proof(s).

    The "computer simulation" is purely a challenge for those that make various claims of being able to violated Bell's inequality (the conclusion portion of Bell's Theorem) even with a system that adheres to the conditions of Bell's Theorem (such as the conditions of this challenge).  The idea is to provide a Randi-type challenge to allow such claims to either "put up" or "shut up".

    The "certainty" is all with Bell's Theorem, and its accompanying proof(s), that no compliant attempts can possibly succeed.

    Does that help clarify the situation?

    David

    P.S.  You do recognize that actual, "real world experiment[s]" violate Bell's inequality.  Correct?  So, therefore, you further recognize that such violations tell us that the "real world" must violate the conditions of Bell's Theorem.  Right?

    So the challenge is to create a "real world experiment", in the form of this "computer simulation", that adheres to the conditions of Bell's Theorem, but, nonetheless, violates Bell's inequality (a Mathematical impossibility, according to Bell's Theorem and its accompanying proof(s)).
    Splendid project! I will try to make a sage version of the challenge (http://sagemath.org). The open source computer algebra system.

    There was some discussion about Bell's original inequality versus CHSH. One could also use the situation in the Nick Herbert's simple proof of Bell's theorem. In one wing of the experiment we have angles 0 degrees and 30 degrees, in the other wing 0 degrees and -30 degrees. The idea is to simulate the probabilities cos^2(theta) that the binary outcomes in the two wings of the experiment are the same, where theta is the difference between the angles (settings) left and right. cos^2(0)=1, cos^2(30 degrees)=3/4, cos^2(60 degrees)=1/4. This corresponds to correlations 1, 1/2, 1/2, -1/2 and a CHSH quantity 1+1/2+1/2-(-1/2)=2.5 > 2.

    vongehr
    Thank you for your comments - sadly "self-destruct" does not mean to vanish from academia in case of people like Joy - in fact he is doing a lot better than me in terms of funding and positions and suchlike. I looked at your site and found you have some surprisingly rational input on issues like Monty Hall and the two envelopes (I have a many worlds version of that problem in my drawer that would likely drive you up the wall, ha ha, though perhaps you would even agree).
    Nice to still get some interest into this. If you are really interested in writing code, I certainly still am [and also try to properly publish this challenge, but reactions are very hostile (maybe you like to have a look)], but please consider that the whole challenge really makes only sense when aimed at the wider public (experts don't need convincing and crackpots refuse it). What I mean is, you are all correct about CSHS and many interesting modifications, but wider science outreach is cut further in half by every such slight complication. Bell's inequality can be explained to school kids I hope in some way similar to what I attempted in the posts on hidden variables, but CHSH sends them packing.
    Hello Sascha. I am not so much interested in writing code but I am interested in public education and the whole idea of a computer challenge. But I'll see what can be done with Sage, since I'm playing with it at the moment. Also I'm interested in the probability theory of your variant of the challenge, when all "particles" are generated in one go, and each measurement station gets all settings in one go. One cannot enforce a rule that the local hidden variable model must only process the settings (inputs) sequentially, committing to an output (measurement outcome) before looking at the next settings. So the job to prove probability bounds on some deviation above CHSH is non-trivial. In the sequential case I could use martingale theory.

    Han Geurdes, who is a friend if mine (despite his IMHO nutty ideas about Bell) is presently working on a computer program to implement his model. Maybe he is interested in informing journalists that he took up the challenge.

    vongehr
    I am interested in public education
    good -that is already something surprisingly many seem to miss and refuse to grasp.
    your variant of the challenge, when all "particles" are generated in one go, and each measurement station gets all settings in one go. One cannot enforce a rule that the local hidden variable model must only process the settings (inputs) sequentially, committing to an output (measurement outcome) before looking at the next settings. ... In the sequential case ...
    Notice that the point of my version being not sequential is merely to avoid consistently ongoing communications between the computers, every instance of which which makes cheating so much more easily hidden.
    Maybe he is interested in informing journalists that he took up the challenge.
    Well, he would not be the first - B. Sanctuary (among others I am sure) claims to have taken it up straight away and defeated (via straight away violating all the rules of the challenge of course). Again, this is not to bring more attention to crackpots. The point is that the challenge should be out there as a working program and it will, just like the Randi challenge, of course never be beaten. The point of a Randi challenge is that it is out there and stays unbeaten although the rewards for beating it are large and it should be easy if the crackpot's claims about the classicality of QM are true.
    I understand the advantage of doing all communication in one go! It leads to an interesting probability problem: can the local realist take advantage of the "bulk processing" of inputs and outputs in order to achieve larger chance fluctuations, or even a systematic bias?

    When fixing a sequential protocol with Han Geurdes this turned out to be a very sticky point, and it also requires fixing time intervals within which results must be delivered. Otherwise the measurement computers can use the fact that the other one is taking an extra long time getting its work done to communicate to one another.

    vongehr
    can the local realist take advantage of the "bulk processing"
    Since the entanglement makes all the photon pairs distinguishable, there is no difference between sequential or bulk. They stay to have their numbers, first pair, second pair, ... . As long as the angles are chosen first (first angle, second angle, ...), it does not matter whether afterwards we analyze the whole lot. If the code somehow applies the 77th angle to the 23rd pair and such shenanigans, well, I do not see how that is not making the correlation just worse, and anyway, it would be fast discovered in case such should lead to apparent violation of locality.
    Richard Gill
    The local realist's code is completely arbitrary. It inputs 800 angles and outputs 800 outcomes. Any shenanigans whatever have to be allowed. We can't rely on hiring an independent IT specialist who verifies that the code operates strictly sequentially (each outcome is in effect committed to before the next setting arrives). The crackpot will simply say that the IT specialist is part of our conspiracy.
    But don't worry, I'm sure probability theory has an answer.


    vongehr
    Yes, of course, all shenanigans are allowed at that point and won't help a bit. What I mean by that shenanigans that do introduce cheating type of "non-locality" will be discovered is not our problem. Our problem is merely to put the challenge out there in a form that is more accessible to a wide audience. A crucial aspect of the challenge is that the success is independent of us. If somebody indeed were to come up with variables that violate Bell, this would be so revolutionary that we would have to do nothing at all - no need to pay price money nor need to check. The interest for solutions that claim to work would be so enormous - all that will be done for us. Somebody who comes up with a solution will get famous even if I personally really really do not like it. That is the inherent guaranty if we do it the way I propose. This is what makes it a "Randi type" challenge in the first place.
    Halliday

    Richard:

    This is an argument I have already had with Sascha.

    At least he appears to be somewhat more willing, with you, to suppose that the batch processing may pose a problem.  However, there is a simple solution that accomplishes batch data transfers, but isolates the "completely arbitrary" "local realist's" code as subroutines that are called in a sequential manner.

    If you wish, you can read what I posted on such things, above, and in Sascha's previous article.  Also feel free to correspond with me on such matters.

    David

    P.S.  I had cobbled together a mockup version in Javascript.  Unfortunately, the framework I was using was not accessible by Sascha, so he couldn't take a look at it.  I then did some searching for other frameworks, but got too busy with "real" work (you know, the kind that pays the bills).

    Richard Gill
    Sorry I missed this exchange (been busy with other things). But I do understand enough now about the bulk processing problem to be able to make safe bets with a local realist about what his (or her) computer programs can do. 
    The problem with bulk processing is that the local realist can create arbitrary dependence between the outcomes of different measurements. This can be used to increase the variance of the outcome. He can't use it to improve the expectation value (CHSH applies) but he can get the variability of the result very big. So he can improve his chance of winning a one-bulk-run bet to close to, say, 50%. But not more. So we just have to repeat the experiment say 10 times, and talk him into agreeing that he has to win every time.
    vongehr
    Well, as you know, but let me repeat just for those readers that may not have followed our discussions that closely:

    1) My example number 5 in the article above already "violates" Bell/CHSH 50% of the time, thus I stay confirmed in my original claim, if I may quib, that we do not need to be worried about bulk processing.

    2) Making "local realists" (LR) agree to win whatever number of times makes no sense, because:

    2.1) We live in a quantum world, thus both sides win plenty of times in parallel worlds (QM is not claiming that a LR cannot win 10 times - that is just not true and should therefore not be taught) and

    2.2) with today's fast computers, a LR can test any complex model in minutes before agreeing to any bets (meaning: you won't find anybody agreeing to the bet anyway, as is your experience if I remember correctly(!)).

    So, lets focus on what can be done and what makes sense to teach and what would be a contribution to public understanding etc., i.e. let us forget the crackpots who never will agree to having really lost anyway and focus onto millions of young bright people who are kind of interested into science but not that mathematically adept: We can teach them the core of quantum mechanics and reject the pseudo science.
    Richard Gill
    I agree public outreach is a different matter from private challenges. 
    But private challenges are also useful. I've been talking one on one to several anti-Bellists. They have accepted my challenge. We are not betting money, but I've promissed to publicly admit I was wrong if they win the challenge. I give them a challenge which I know they will only win by chance, and that only with a small probability. The law of large numbers means that even if they happen to be lucky once, they won't be able to reproduce their success. So they'll learn at home, while testing their program, that it repeatedly fails.

    They have accepted the challenge and are now busy at home programming. I think that they will learn from their failure. At least it keeps them quiet for a while. And I've said: "I'm not going to study your theoretical paper till you've succeeded in the challenge". 

    Is there any tangible reward attached to the QRC?
    Particularly if the challenge is poorly formed?
    My simple model of EPR reproduces the known distribution and makes predictions that will need to be checked against experiment.
    It makes Bell's theorem not incorrect, but irrelevant: based on the usual misunderstandings that QM generates and the attendant pseudo-science that is continually foisted on the public.
    Best
    J

    Halliday
    Jay:

    You ask:

    Is there any tangible reward attached to the QRC?

    Have you read the article?  Read items labelled "3)".  Whether that is "tangible" enough is up to you.

    You claim:

    My simple model of EPR reproduces the known distribution and makes predictions that will need to be checked against experiment.

    Can your "simple model of EPR" be formulated according to the QRC?  If so, then go for it!

    David

    I am having trouble figuring out what the real problem is in having computers reproduce the correlations, because last time I actually read up in detail on the polarisation experiments I got the impression that what was spooky was simply that a cosine was involved. As I recall it was something like the chance the photon would pass the filter was given by a cosine and that a cosine is so structured that the chance of a photon passing plus the chance of an oppositely polarised phton also passing added up to more than 100% at some angles?

    If so, couldn't the results be reproduced simply by sending a random sequence of angles to Alice and the exactly opposite angles to Bob and having them both do some kind of cosine based calculation on them?

    Basically the impression I had was that what was spooky was the fact that a cosine determined the probability, since by nature cosines result in weird statistics.

    Thus although Joy's work seems to indicate that using quaternions should suffice for a simple case, and octonians for all cases, I am now wondering why not just use a cosine in the detector simulator and be done with it?

    -MarkM-

    Richard Gill
    MarkM: the problem is that it can't be done. The point of the challenge is that people figure that out for themselves. 

    This is how I explain the mathematical core of Bell's theorem to newcomers:

    Consider 4N runs of a Bell-Aspect-Weihs delayed choice CHSH type experiment. Suppose that Nature is such that in each run, binary outcomes A, A', B, B' (each +/-1) can be thought to all exist alongside one another, but that only one of A and A', and only one of B and B' are actually observed - the choices being made by independent fair coin tosses, independent of the physical processes generating the 4N realizations of the four binary variables A, A', B, B' 

    i.e. suppose we assume realism (aka counterfactual definiteness), locality (aka relativistic local causality), and freedom (from superdeterminism) (aka no conspiracy).

    It's easy to see that AB+AB'+A'B-A'B'=A(B+B')+A'(B-B')= +/-2 in each run. (B and B' are either different or they're equal ...)

    It follows from taking averages over the 4N runs, that ave(AB)+ave(AB')+ave(A'B)-ave(A'B') lies between -2 and +2.

    Finally: if N is very large, the average of AB over the runs where A and B are both observed (that's about N out of the 4N, and they're selected completely at random) will be very close to the average of AB over all 4N runs; and similarly for AB', A'B, A'B'.

    If this last point is doubted, one can put numbers to "how close, with what probability" using some probability theory (technical details: I like to use Hoeffding's inequality for tails of the binomial distribution and of the hypergeometric distribution. It turns out that the probability that CHSH is violated by more than some amount delta is less than C exp( - D N delta^2) for certain positive constants C and D. To be precise, C = 8 and D = 1/64 will do, if we restrict delta to the interval (0,2). )

    The point is, everything here is discrete, finite, including the probability, which is really a counting argument, going through the 2^8N equally likely sets of different outcomes of the 8N independent fair coin tosses.

    Now it maybe that you are not convinced by my argument. Go ahead then, try to program your computer to win the challenge. I predict that you'll find that it can't be done. More importantly, I predict that from trying it yourself, you'll begin to understand what the issue is.



















    That sounds a bit harsh so I shall try to clarify. It seems you have a preconceived model of your own in mind, your code implements your model, and you want someone to make your model behave as Christian's model does while leaving it as your model not as the actual model Christian pro;oses which you seem to have contructed your code in such a way as to preclude.

    -MarkM-

    vongehr
    you have a preconceived model of your own in mind,
    There are three different models, all are the simplest of a certain type in order to help challengers to modify according to any model they please.
    your code implements your model,
    The QRC is specifically about that challengers can put any local realistic model in whatsoever. I cannot explain yet more clearly or make it yet more clear in the code, which explicitly states that the whole model part is completely up to the challenger.
    you want someone to make your model behave as Christian's
    The QRC has nothing (NOTHING) whatsoever to do with anybody by the name of Christian.
    Thanks for posting this challange. I've been trying to learn about quantum mechanics for some months now, and I think the Bell experiment and its implications are what trip me up. So I will give the QRC a sincere go, because whatever the end result, I look at it as a win-win situation: either I'll finally understand Bell's theorem, or I win the Nobel prize :)