Talk:Loopholes in Bell tests

Started a complete rewrite to make the page up to par with current verifiable scientific position. Jan-Åke Larsson 20:47, 5 December 2006 (UTC)

Other Loopholes
Harald88: "ruled out" vs "reduced to a minimum": Fair enough. But note that these "Other loopholes" can be put in categories: Those who can be avoided completely (accidentals, synchronisation, double detection), one which does not apply (rotational invariance), and one which goes in under the efficiency problem (no enhancement). Gisin did say that there are no loophole-free experiments yet, but he was talking about efficiency/locality. Or perhaps it was unclear that it was the "other loopholes" that can be ruled out. Encyclopedic as this is, they will need to be mentioned in the article anyway. I'll write something proper on this when I get the time. Jan-Åke Larsson 22:38, 7 December 2006 (UTC)


 * Good - thanks for the effort, this article was in a poor shape for a too long time. Harald88 23:54, 7 December 2006 (UTC)

The entry on Digital physics mentions a Loophole not listed on this page, called "pre-determinism". Can anyone with more knowledge provide a description? Naasking 04:49, 26 December 2006 (UTC)


 * This would be what Bell himself referred to as "Superdeterminism", that everything is pre-determined. One of Bell's "four possible positions". I'll add it when I get the time. Jan-Åke Larsson 15:30, 1 January 2007 (UTC)


 * I am a physicist, and had some training on this subject matter, but I cannot make heads or tail of what is being said in this section. I think it is too much jargon for wikipedia in its present form. &#151; Xiutwel ♫☺♥♪ (talk) 12:34, 21 August 2007 (UTC)

I am not a physicist and I find the current form of this page along with CHSH_inequality the most lucid presentations I've read about real-world tests of inequality! Not trying to argue, just praising. Jwoehr (talk) 21:37, 1 August 2009 (UTC)

Proposed merge with Superdeterminism
Thankyou to User:Jim E. Black for tidying up Superdeterminism, it reads much better now. I support the proposed merge. Marasmusine 07:15, 14 May 2007 (UTC)

I do not. Superdeterminism is a separate subject Jan-Åke Larsson (talk) 12:07, 18 September 2008 (UTC)

Some background information
I thought it could be a good Idea to draw a sketch (I did not draw it myself) and explain some of the technical vocabulary used in this and similar pages, so it would be more comprehensible to a non-expert. As well as some basics on measurement errors and a rough explanation of a typical experiment. I am not really an expert myself though.Agge1000 00:53, 20 September 2007 (UTC)

Is this a POV fork or what?
Is this a POV fork or what? 1Z 00:05, 30 October 2007 (UTC)
 * From what I understand from this (see the archive) talkpage and those in related articles there used to be an editor, User:Caroline_Thompson with a special interest in "loopholes", and this article was created to avoid getting the main article on Bell test experiments to be all about loopholes. More like a solution to an edit war than a pov fork. Agge1000 23:12, 30 October 2007 (UTC)


 * Indeed, it's not a POV fork but instead it's a spin-off from Bell test experiments, after one or two editors complained that this subtopic took too much space on that page. Of course, care must be taken that the two pages continue to link to each other. Harald88 18:18, 4 November 2007 (UTC)

Removed the "circular logic loophole"
There is no reference to this other than being an argument on a webpage somewhere. The argument sounds quite confused, and makes statements about assumptions used in the derivation of the CHSH inequality that are certainly not necessary. Wikipedia pages do not constitute an authoritative source for the assumptions required to derive the CHSH inequality, so objecting to statements in the Wikipedia page doesn't constitute the discovery of a loophole. This seems to be a combination of POV, original research and unverifiable. I have reverted to before it's being added. --Drojem (talk) 06:18, 14 January 2008 (UTC)

Removal of a lot of original research
A significant portion of this article had what seemed to be the user Caroline Thompson pushing their articles (most of which were not published in any reputable journal) from arXiv. It appeared to be a violation of verifiability and no original research. I have removed the most egregious of examples, however I have left the reference to the one published article (from the journal Foundations of Physics) and the related content in. Here's the diff of the edits in question: []. Someone42 (talk) 11:15, 20 June 2008 (UTC)

Local causality NOT "local realism"
I replaced all instances of the term "Local Realism" with the terms "Local Causality" or "locally causal". Bell's theorem has no negative bearing on realism, as many nonlocal hidden variable theories (most notably the de Broglie-Bohm theory) can violate the inequalities. The central assumptions being tested in these experiments are locality and causality. Bell makes this very clear in all his papers, many of which are cited in this article. (by Unknown)


 * I disagree strongly. "Local Realism" is a technical term used in the literature. The de Broglie-Bohm theory is realist but not local realist. Indeed Bell's theorem has no negative bearing on realism, just as quantum mechanics is local. But Bell's theorem does have an impact on the possibility of local realist models. Jan-Åke Larsson (talk) 12:12, 18 September 2008 (UTC)

Need for an update!
The most recent reference dates more than 10 years back. I just saw a 10 year old article by Sulcs et al,

"Effective linearization by noise addition in threshold detection and implications for stochastic optics" - J.Phys. A: Math. Gen.33 (2000) 3997-4007 http://iopscience.iop.org/0305-4470/33/22/304 :

Abstract: "We present some computer simulation results pertaining to the restoration of a weak sinusoidal input signal after threshold detection with additive random noise. We show how, if stochastic optics is regarded as a viable theory, these results may cast doubt upon some conclusions drawn from seven experimental realizations of the Einstein-Podolsky-Rosen thought experiment performed between 1982 and 1998."

Obviously, that article is very relevant for this "Loopholes" article so that it should be mentioned. Probably some of the information in this Wikipedia article has to be corrected based on the information contained in that article, as the authors concluded that at that time the "detector loophole" remained open and was even explained.

Moreover, it would be very surprising if no other new information has been published in the last decade.

Harald88 (talk) 11:28, 26 July 2010 (UTC)

Speed of light loophole
I recently published a paper in Complex Systems,, Emergent Properties of Discretized Wave Equations that suggest an additional loophole in tests of Bell's inequality related to the necessity of any discrete model for physics with a regular spatial grid to approximate the wave equation to very high accuracy. This requires a velocity of wave propagation slower, perhaps much slower, then the one time step in one spatial step that is the maximum rate that a physical effect can propagate in these models. There is a web page that summarizes this argument: http://www.mtnmath.com/discr_wave.html Paul Budnik Mtnmath (talk) 16:45, 22 April 2011 (UTC)

Coincidence loophole
The minimum efficiency for overcoming the detection loophole (82%) is based on the idea that the decision whether or not detection events belong together is based on external timing criteria, e.g. In a pulsed emission experiment, the time of emission of two photons determines a detection time window at each detector for the measurement of the two photons in that pair. However, in typical experiments, two events are considered part of a pair merely if their intearrval time is shorter than some threshold. Thus the time window for determining which events are paired is partly fixed by the detection events themselves, I.e., "after" settings have been fixed in each wing of the experiment. This gives further room for a local realist explanation of observed data. According to Gill and Larsson (2004) the minimum efficiency threshold needs to be raised from about 80% to about 90% to take account of this. http://arxiv.org/abs/quant-ph/0312035, "Bell's inequality and the coincidence-time loophole", Jan-Ake Larsson, Richard Gill, Europhysics  Letters, vol 67, pp. 707-713 (2004). See for example de Raedt et al. (2011), http://arxiv.org/abs/1112.2629     Richard Gill (talk) 07:41, 7 June 2012 (UTC)


 * I wrote a small section on this loophole. Still need to add proper literature references. Richard Gill (talk) 13:28, 4 September 2015 (UTC)

Communication, or locality
The sentence "An experimental set-up without any such provision effectively becomes entirely "local", and therefore cannot rule out local realism." in the section "Communication, or locality" does not make sense and appears erroneous. The word "local" does not really apply to the experimental setup, and the intended meaning is far from clear. The last part apperars to be non-sequitur. As the intent is unclear, I cannot propose a correction. If no one else can either, the text should be removed. — Preceding unsigned comment added by 85.23.38.101 (talk) 11:19, 7 February 2014 (UTC)

Memory loophole
I also added a section on the memory loophole. Including references to my own work ... which became refined by other authors, for instance Peter Bierhorst, and is nowadays actually used in experiment, e.g. the recent Delft experiment. Richard Gill (talk) 09:31, 5 September 2015 (UTC)

Referencing or rephrasing
I am adding a few landmark research articles or ones that specifically developed arguments on points made where citations are needed. Double-checks on my additions would be welcome (I have just started contributing to Wikipedia here)

It seems to me, however, that an adequate reference cannot be found to corroborate these two points:
 * 1) "There is no way to test experimentally whether a given experiment does fair sampling, as the number of emitted but undetected pairs is by definition unknown."
 * 2) "Entanglement and local realism give different predicted values on S, thus the experiment (if there are no substantial sources of error) gives an indication to which of the two theories better corresponds to reality."

I think rephrasing into a weaker statement would be acceptable without citation. What do you think?

Sharasque (talk) 14:48, 1 December 2015 (UTC)

Disjoint sampling
I removed the section which claimed (without reference to any sources) that it is a loophole that different particle pairs are needed to estimate different correlations. What Bell assumed, is that the local hidden variables are drawn from the same probability distribution, whatever pair of measurement settings are used. Richard Gill (talk) 20:02, 17 April 2016 (UTC)

Ensemble interpretation avoids loopholes
This sentence should be included in Wikipedia:

An ensemble interpretation of quantum mechanics explains quantum correlations without any instantaneous action-at-a-distance and completely avoids any question of loopholes. Published References: http://www.worldscientific.com/doi/abs/10.1142/S0217979206034078 https://arxiv.org/ftp/quant-ph/papers/0404/0404011.pdf http://physicsessays.org/browse-journal-2/product/47-3-dean-l-mamas-an-intrinsic-quantum-state-interpretation-of-quantum-mechanics.html


 * The point is, can we explain quantum phenomena by supposing an underlying classical-like world? An ensemble interpretation of quantum mechanics does not help answer this question. It just refuses to pose the question. As Bell himself remarked, Bohr's reaction to Bell's findings would have been "so what?". Richard Gill (talk) 16:43, 25 December 2018 (UTC)

Nonrecurrence
I removed the section on non-recurrence. This refers to a paper by Danforth in which it is argued that if the hidden variables are real numbers, then any particular value will never be repeated. According to that paper, this assumption is however needed in the proof of Bell's theorem. (I believe this claim is not true, but of course, my personal opinion is irrelevant).

As far as I can see, this point of view has not achieved any further support in the literature. Hence it should not be included as a notable loophole. Richard Gill (talk)

PS here is the deleted text: When hidden variables belong to the reals they do not recur from one instance to the next. That makes it impossible to derive Bell's inequality that assumes the hidden variables are shared between different experimental configurations, e.g. (a,b), (a,b'), (a',b), (a',b') as specified in the 4 experiments used to define the CHSH inequality.

this article is awful
As a layman I came here to read about “freedom-of-choice” loopholes. I am more lost now then when I encountered the term in a paper.

"disjoint measurement" loophole? (and misc.)
what is the "disjoint measurement loophole" that is mentioned in the introduction but the not discussed in the main part of the article? I have never heard of it and Google only finds WP-related results. Does it refer to the "disjoint sample loophole" (mentioned, e.g., here and her) and attributed to Larsson? In that case, wouldn't itbelong the detection/fair sampling loopholes? In any case, is this loophole really of the same importance as detection/fair sampling so that it should be mentioned so prominently?

Two unrelated comments: (1) Hanson's experiment is mentioned twice in the introduction, I think the 2nd one suffices. (2) The reference to Eberly's "classical entanglement" does not belong in this article, since it has nothing to do with "loopholes". (there is no spatial separation (at all) between the degrees of freedom considered there) --Qcomp (talk) 16:24, 18 January 2020 (UTC)
 * Not only the introduction, but most of the article was a dumpster fire. I started rewriting it. Tercer (talk) 14:09, 8 February 2022 (UTC)
 * Thanks. XOR&#39;easter (talk) 04:07, 11 February 2022 (UTC)