Coming Soon!

[gill1109]

So here's a nice example:

m = 1
SeedRandom[5]

inputs: 0, 45
outputs: –1, +1

inputs: 0, 135
outputs: –1, –1

inputs: 90, 45
outputs: –1, +1

inputs: 90, 135
outputs: –1, +1

With Bob's input = 135 degrees, changing Alice's input from 0 degrees to 90 degrees switches Bob's output from -1 to +1

Matching must be trivial since Alice and Bob both have one input and one output. It's nice that the code worked even with this unusual value of m. Anyway: the program is evidently non-local.

I was systematically trying all four setting combinations with m = 1 (standard CHSH angles) and the random seed set to 1, then 2, then 3 .... It took me 5 times 4 tests to find one example of non-locality.

https://www.wolframcloud.com/obj/gill1109/Published/test2.nb
The input data comes from files Alice0.txt, Alice1.txt, Bob0.txt, Bob1.txt each containing the single number 0, 90, 45, 135

[FrediFizzx]

@gill1109 More freakin' nonsense. Matching events for A and B is basically a non-local process. That is all your test is showing is that we are matching events between A and B.

Experimenters do not “match events”. Nature does not “match events”. Experimenters pre-define matched time-slots and correlate local data generated locally in the matched time-slots. My test shows what it shows: your model is a hidden variables model, but not a local hidden variables model. A = A(a, b, lambda); B = B(a, b, lambda).

Actually I need to do a few more tests in order to show that, in the example I showed you, the outcomes of trial three only depend on the hidden variables and settings of trial three. No use of memory loophole, no retro-causality. In fact, k = k_A = k_B. The LaTeX formulas which Joy supplied are *not* the formulas actually used to compute outcomes from settings and hidden variables.

Pure nonsense. "matched time-slots" is the same freakin' thing as matching events. When the "orphans" are removed from A_1 and B_1, listA4 and listB4 become matched automatically.
Your test only shows that events are being matched. Period.
.

[gill1109]

@gill1109 More freakin' nonsense. Matching events for A and B is basically a non-local process. That is all your test is showing is that we are matching events between A and B.

Experimenters do not “match events”. Nature does not “match events”. Experimenters pre-define matched time-slots and correlate local data generated locally in the matched time-slots. My test shows what it shows: your model is a hidden variables model, but not a local hidden variables model. A = A(a, b, lambda); B = B(a, b, lambda).

Actually I need to do a few more tests in order to show that, in the example I showed you, the outcomes of trial three only depend on the hidden variables and settings of trial three. No use of memory loophole, no retro-causality. In fact, k = k_A = k_B. The LaTeX formulas which Joy supplied are *not* the formulas actually used to compute outcomes from settings and hidden variables.

Pure nonsense. "matched time-slots" is the same freakin' thing as matching events. When the "orphans" are removed from A_1 and B_1, listA4 and listB4 become matched automatically.
Your test only shows that events are being matched. Period.
.

My new test with m = 1 shows that no orphans were removed. It does show that the outcomes can depend on the other side's setting. Your code does not do what you think it does. You did not implement Joy's LaTeX formulas. k = k_A = k_B.

[FrediFizzx]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.
.

[Justo]

My new test with m = 1 shows that no orphans were removed. It does show that the outcomes can depend on the other side's setting. Your code does not do what you think it does.

The problem is they do not explain what they think their code does. I don't know what the code does but I regret they do not summarily explain what the code does in the introduction. They present the formulas but an explanation of how the model uses them is missing.
It is not obvious how and why they consider the cases . In a theoretical model one just pairs values from the same event, why should we consider cases where k_A is different from k_B.
Those cases are natural if you simulate a real experiment with possible loopholes, but why considering them in a purely theoretical simulation?
I am not saying that it is incorrect, I just point out that it is not evident.

[Joy Christian]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Let him amuse himself. We continue our secret mission for the paper.
.

[gill1109]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Trying out various small values of m and trying out a few settings with a few different random seeds proves that your code does not do what you think it does: it proves that the code is non-local and that it does not implement Joy's formulas.

Of course, if you want proof, then you have Bell's theorem, a bit of elementary probability theory that proves that you are attempting the impossible. But you don't believe in Bell's theorem. So you will keep running up against a brick wall. Perhaps you should take some time out and study some of the elementary proofs of that almost trivial theorem? The general approach was worked out in the 1850's by George Boole. It has since been rediscovered many times.

Let him amuse himself. We continue our secret mission for the paper.

Good luck, have fun!

[FrediFizzx]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Let him amuse himself. We continue our secret mission for the paper.
.

Sure, but he finally broke down and got the Mathematica free trial.
It is a way more powerful math program than using the geeky R.
.

[jreed]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Let him amuse himself. We continue our secret mission for the paper.
.

You are actually going to publish this after all these counterexamples? I'm really impressed!

[gill1109]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Let him amuse himself. We continue our secret mission for the paper.
.

Sure, but he finally broke down and got the Mathematica free trial.
It is a way more powerful math program than using the geeky R.

I had grandchildren to stay for two weeks, so not so much time for my retirement hobbies.

Mathematica is very powerful and very good for some purposes. I find it amusing to use it from time to time. I use the free trial option several times per year.

In your case Mathematica enabled you to write code of such complexity that nobody can understand it, you can’t explain what it does, and it certainly doesn’t do what you say it does.

Nowadays the most widely spread language for scientific computing is Python. I often use it. R is more specialised (statistics, data science). Why don’t you try to program your algorithm in Python so that everyone can understand what you’re trying to do? There is a huge and friendly user community.

[FrediFizzx]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Let him amuse himself. We continue our secret mission for the paper.
.

You are actually going to publish this after all these counterexamples? I'm really impressed!

They are not counterexamples. They are not our model. Typical Bell fanatics strawman attempts.
.

[gill1109]

@gill1109 More freakin' pure nonsense. Setting m=1 proves nothing.

Let him amuse himself. We continue our secret mission for the paper.
.

You are actually going to publish this after all these counterexamples? I'm really impressed!

They are not counterexamples. They are not our model. Typical Bell fanatics strawman attempts.
.

I ran your program, Fred! I changed nothing. No strawman. You can perform the same tests that I did, if you don’t trust me. Do you dare?

[FrediFizzx]

I ran your program, Fred! I changed nothing. No strawman. You can perform the same tests that I did, if you don’t trust me. Do you dare?

CHSH = 2.78608 I already told you why your tests are freakin' nonsense. Pay attention! You are saying that we can't match events. Pure nonsense.
.

[gill1109]

My new test with m = 1 shows that no orphans were removed. It does show that the outcomes can depend on the other side's setting. Your code does not do what you think it does.

The problem is they do not explain what they think their code does. I don't know what the code does but I regret they do not summarily explain what the code does in the introduction. They present the formulas but an explanation of how the model uses them is missing.
It is not obvious how and why they consider the cases . In a theoretical model one just pairs values from the same event, why should we consider cases where k_A is different from k_B.
Those cases are natural if you simulate a real experiment with possible loopholes, but why considering them in a purely theoretical simulation?
I am not saying that it is incorrect, I just point out that it is not evident.

It must be incorrect, because they violate Bell’s inequality. Simple tests show that their program is indeed non local.

The paper does not define k_A or k_B, and the Mathematica code is so complex that the authors cannot explain concisely and accurately what it does. It is therefore likely to contain errors. The authors’ talk about matching shows that they have no idea of present day experimental practice. Early experiments had many defects. Those defects were well known and well understood. They have now been circumvented. Only retro-causality and superdeterminism are available as metaphysical loopholes.

I ran your program, Fred! I changed nothing. No strawman. You can perform the same tests that I did, if you don’t trust me. Do you dare?

CHSH = 2.78608 I already told you why your tests are freakin' nonsense. Pay attention! You are saying that we can't match events. Pure nonsense.
.

Fred, you are not paying attention to me.

No problem!

[Joy Christian]

My new test with m = 1 shows that no orphans were removed. It does show that the outcomes can depend on the other side's setting. Your code does not do what you think it does.

The problem is they do not explain what they think their code does. I don't know what the code does but I regret they do not summarily explain what the code does in the introduction. They present the formulas but an explanation of how the model uses them is missing.
It is not obvious how and why they consider the cases . In a theoretical model one just pairs values from the same event, why should we consider cases where k_A is different from k_B.
Those cases are natural if you simulate a real experiment with possible loopholes, but why considering them in a purely theoretical simulation?
I am not saying that it is incorrect, I just point out that it is not evident.

Both Fred and I have independently explained this in this forum. I am not willing to repeat my explanation. But we are in the process of revising the paper with an improved presentation.
.

[FrediFizzx]

... Fred, you are not paying attention to me. ...

CHSH = 2.78608 Why would I pay attention to you? All you do is spew freakin' nonsense.
.

[FrediFizzx]

My new test with m = 1 shows that no orphans were removed. It does show that the outcomes can depend on the other side's setting. Your code does not do what you think it does.

The problem is they do not explain what they think their code does. I don't know what the code does but I regret they do not summarily explain what the code does in the introduction. They present the formulas but an explanation of how the model uses them is missing.
It is not obvious how and why they consider the cases . In a theoretical model one just pairs values from the same event, why should we consider cases where k_A is different from k_B.
Those cases are natural if you simulate a real experiment with possible loopholes, but why considering them in a purely theoretical simulation?
I am not saying that it is incorrect, I just point out that it is not evident.

Both Fred and I have independently explained this in this forum. I am not willing to repeat my explanation. But we are in the process of revising the paper with an improved presentation.
.

Yes, and we actually put in a suggestion by Justo in the revised paper. Thanks Justo.
.

[gill1109]

I look forward to seeing the revision.

On FaceBook, Fred says that k subscript A = k subscript B = k. So no need to define them. Events are matched if they have the same value of k. Excellent. Now the question is, how are the outcomes defined? *Not* by the formulas presently in the paper, as anyone can easily check.

[FrediFizzx]

I look forward to seeing the revision.

On FaceBook, Fred says that k subscript A = k subscript B = k. So no need to define them. Events are matched if they have the same value of k. Excellent. Now the question is, how are the outcomes defined? *Not* by the formulas presently in the paper, as anyone can easily check.

You have Mathematica so you can easily check the analytical formulas. I have and they match. All angles at 45 degrees is a no-brainer. A = -1 and B =+1. Try it in Mathematica.
.

[Gordon Watson]

I look forward to seeing the revision.

On FaceBook, Fred says that k subscript A = k subscript B = k. So no need to define them. Events are matched if they have the same value of k. Excellent. Now the question is, how are the outcomes defined? *Not* by the formulas presently in the paper, as anyone can easily check.

You have Mathematica so you can easily check the analytical formulas. I have and they match. All angles at 45 degrees is a no-brainer. A = -1 and B =+1. Try it in Mathematica.
.

Fred,

What experiment are you running? I'm not familiar with one that would give A = -1 and B = +1 for all angles at 45 degrees.

Is it the one from Bell (1964)? AKA EPR-Bohm?

Gordon