Another Response to Richard Gill's 10,000 Euros Challenge

Foundations of physics and/or philosophy of physics, and in particular, posts on unresolved or controversial issues

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Fri Jun 13, 2014 1:12 am

Joy Christian wrote:
gill1109 wrote:Please explain to us why there is no "s" or "p" mentioned in the challenge.

Ask the challenger.

Pity. Too late.

Too bad. The challenge is the challenge.

If "s" and "p" had been introduced explicitly in page 4 of the 2008 experimental paper, then there possibly never would have been a challenge. So both ways, Christian doesn't get 10 000 Euro.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Fri Jun 13, 2014 1:16 am

gill1109 wrote:
Joy Christian wrote:
gill1109 wrote:Please explain to us why there is no "s" or "p" mentioned in the challenge.

Ask the challenger.

Pity. Too late.

Pity, indeed. Because of your foolishness, you now owe me 10,000 Euros.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Fri Jun 13, 2014 1:30 am

Joy Christian wrote:
gill1109 wrote:
Joy Christian wrote:gill1109 said "Please explain to us why there is no "s" or "p" mentioned in the challenge."

Ask the challenger.

Pity. Too late.

Pity, indeed. Because of your foolishness, you now owe me 10,000 Euros.[/quote]
I'm afraid that the adjudicators probably disagree with you.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Fri Jun 13, 2014 2:17 am

gill1109 wrote:I'm afraid that the adjudicators probably disagree with you.

We don't know that.

In any case, my model and my numbers fully agree with me and disagree with you: http://rpubs.com/jjc/19298.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Fri Jun 13, 2014 7:37 am

Joy Christian wrote:
gill1109 wrote:I'm afraid that the adjudicators probably disagree with you.

We don't know that.

In any case, my model and my numbers fully agree with me and disagree with you: http://rpubs.com/jjc/19298.


Everything in your world appears self-consistent to you.

Everything in my world appears self-consistent to me.

I hope we'll know soon what the adjudicators think. Hans de Raedt told me he thought he should be ready very soon. Gregor seems to be already done. Andrei Khrennikov is hopefully relaxing after a fantastic conference. He predicted that by the next one (second week in June 2015) a loophole free Bell experiment will already have been done.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Fri Jun 13, 2014 8:42 am

gill1109 wrote:Everything in my world appears self-consistent to me.

But everything in your world is hardly self-consistent. Among many other things, two of the four correlations you have been calculating do not even reside on the local-realistic correlation surface generated in this simulation in response to your challenge: http://rpubs.com/jjc/19298. I would hardly call that self-consistent.

This contradiction should have made you realize that all the calculations you have been doing are wrong. There is a fundamental inconsistency in your statistical worldview, and my simulation gives you a brilliant opportunity to correct it (albeit at a price of 10,000 Euros).

The correct calculations fully consistent with both my local model as well as my proposed experiment are done in the above simulation. Note the consistency between the four points on the 2D surface (with random a and b) and the four separate correlations calculated in the simulation with fixed a and b. The correct results are:

Code: Select all
 
E(0, 45)  E(0, 135)  E(90, 45)  E(90, 135)
   -0.6993     0.703     -0.699      -0.7276

Now this I call self-consistent.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby FrediFizzx » Fri Jun 13, 2014 9:03 pm

gill1109 wrote:
FrediFizzx wrote:Then please tell us why the experimenters on the real experiment don't need to do anything with "s" and "p" from the simulation.

Ah, good question!

It wasn't a question; it was a directive for you to try to explain. Which you didn't even try so I guess you can't explain it and thus you really don't know what a hidden variable is.
FrediFizzx
Independent Physics Researcher
 
Posts: 2905
Joined: Tue Mar 19, 2013 7:12 pm
Location: N. California, USA

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Fri Jun 13, 2014 10:39 pm

FrediFizzx wrote:
gill1109 wrote:
FrediFizzx wrote:Then please tell us why the experimenters on the real experiment don't need to do anything with "s" and "p" from the simulation.

Ah, good question!

It wasn't a question; it was a directive for you to try to explain. Which you didn't even try so I guess you can't explain it and thus you really don't know what a hidden variable is.

What's in a name? A variable is a variable. The word "variable" belongs to mathematics. When we call a variable hidden we are adding physical interpretations to a mathematical model. Not directly observable yet "behind the scenes" causing things to be how they are.

I don't have to explain Christian's model. Christian needs to explain to his experimenter, what the experimenter needs to do. Unfortunately, the present instructions cannot result in the desired output.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby FrediFizzx » Fri Jun 13, 2014 11:14 pm

gill1109 wrote:What's in a name? A variable is a variable. The word "variable" belongs to mathematics. When we call a variable hidden we are adding physical interpretations to a mathematical model. Not directly observable yet "behind the scenes" causing things to be how they are.

Do you still think the experimenters in the real experiment need to know something or anything about the hidden variables? If so, why?
FrediFizzx
Independent Physics Researcher
 
Posts: 2905
Joined: Tue Mar 19, 2013 7:12 pm
Location: N. California, USA

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Fri Jun 13, 2014 11:45 pm

FrediFizzx wrote:
gill1109 wrote:What's in a name? A variable is a variable. The word "variable" belongs to mathematics. When we call a variable hidden we are adding physical interpretations to a mathematical model. Not directly observable yet "behind the scenes" causing things to be how they are.

Do you still think the experimenters in the real experiment need to know something or anything about the hidden variables? If so, why?

The experimenters need to follow unambiguous instructions. They don't need to "know" about the different theories they are testing. Not once the experimental protocol is written down.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby FrediFizzx » Sat Jun 14, 2014 12:06 am

gill1109 wrote:
FrediFizzx wrote:
gill1109 wrote:What's in a name? A variable is a variable. The word "variable" belongs to mathematics. When we call a variable hidden we are adding physical interpretations to a mathematical model. Not directly observable yet "behind the scenes" causing things to be how they are.

Do you still think the experimenters in the real experiment need to know something or anything about the hidden variables? If so, why?

The experimenters need to follow unambiguous instructions. They don't need to "know" about the different theories they are testing. Not once the experimental protocol is written down.

So they don't need to know anything about the hidden variables at all. Correct? If you don't think that is correct, then please explain why they would need to know about the hidden variables.

It seems pretty simple to me. They record the angular momentum directions at particular angles for a and b, then take those results and calculate the correlations per Joy's eq. (16). That is all they have to do. There is nothing else. Their results will either show -a.b, or not or maybe they might discover something inbetween. They don't need to know anything about the simulations if they don't want to. But if you are trying to simulate what Nature might be doing to suit a certain theory, you will have to have those hidden variables in the simulation that the theory says are in Nature.

Ok, that is the last time I am going to explain it. Either you get it or you don't.
FrediFizzx
Independent Physics Researcher
 
Posts: 2905
Joined: Tue Mar 19, 2013 7:12 pm
Location: N. California, USA

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Sat Jun 14, 2014 12:13 am

gill1109 wrote:The experimenters need to follow unambiguous instructions. They don't need to "know" about the different theories they are testing. Not once the experimental protocol is written down.

It is beyond my understanding why you continue to lie shamelessly about the experimental protocol for my proposed experiment. It has been written down completely unambiguously by me, on the page 4 of my 2008 paper (this one: http://arxiv.org/abs/0806.3078):

Image

You seem to be suffering from severe reading difficulties. Consequently, you have been calculating the four correlations incorrectly, as I have explained to you here:
Joy Christian wrote:
gill1109 wrote:Everything in my world appears self-consistent to me.

But everything in your world is hardly self-consistent. Among many other things, two of the four correlations you have been calculating do not even reside on the local-realistic correlation surface generated in this simulation in response to your challenge: http://rpubs.com/jjc/19298. I would hardly call that self-consistent.

This contradiction should have made you realize that all the calculations you have been doing are wrong. There is a fundamental inconsistency in your statistical worldview, and my simulation gives you a brilliant opportunity to correct it (albeit at a price of 10,000 Euros).

The correct calculations fully consistent with both my local model as well as my proposed experiment are done in the above simulation. Note the consistency between the four points on the 2D surface (with random a and b) and the four separate correlations calculated in the simulation with fixed a and b. The correct results are:

Code: Select all
 
E(0, 45)  E(0, 135)  E(90, 45)  E(90, 135)
   -0.6993     0.703     -0.699      -0.7276

Now this I call self-consistent.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Sun Jun 15, 2014 5:18 am

I have cleaned up some of my other simulations by incorporating what we have learned during the past few weeks.

Here is an example of a cleaned-up simulation of my 3-sphere model: http://rpubs.com/jjc/16567.

The most important conceptual clarification in this simulation is that, what appeared as a "post-selection" of states to the flatlanders, has now been relegated to the "state preparation" of the pre-selected initial states of the system. This was never a problem from the perspective of the 3-sphere, but the flatlanders were having a great deal of difficulty understanding my model because of it. But now the N spin directions, "v", which are actually observed by Alice and Bob, are all pre-selected, thereby removing the last vestige of the Bell-baggage from the simulation.

In summary, N spin directions, "v", as an ensemble of the initial or complete states of the physical system, are pre-selected in the simulation, after the state preparation within S^3, just as we prepare initial states in quantum mechanics and experiments. The correlations are then calculated using the standard formula,

E(a, b) = (1/N) Sum_(k=1)^(k=N) [ sign(+a.v_k) * sign(-b.v_k) ] = -a.b ,

where the measurement directions a and b are randomly selected, respectively, by Alice and Bob. Note that not a single one of Bell's assumptions has been dropped.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby FrediFizzx » Sun Jun 15, 2014 1:45 pm

Below is a version of Joy's simulation showing that creating the direction files for Alice and Bob that they observe and then reading them back in works in the simulation for when the angles a and b are random. However, if I comment out the write.csv lines and try to read the files back in on a subsequent run, it doesn't work right. The graphic has some distortion in it. Is there some kind of problem with the files syncing with the loop?

Code: Select all
set.seed(9875)

M <- 10^4  ## Sample size. Next, try 10^5, or even 10^6

angles <- seq(from = 0, to = 360, by = 10) * pi/180

K <- length(angles)
Ns <- numeric(K)  ## Container for number of states
corrs <- matrix(nrow = K, ncol = K, data = 0)  ## Container for correlations

r <- runif(M, 0, 2 * pi)
s <- runif(M, 0, pi)

x <- cos(r)
y <- sin(r)
u <- rbind(x, y)

## 'u' is a 2xM matrix. The M columns of 'u' represent the x and y
## coordinates of points on a unit circle in the equatorial plane.

p <- 1.21 * (-1 + (2/(sqrt(1 + (3 * s/pi)))))

for (i in 1:K) {
    alpha <- angles[i]
    a <- c(cos(alpha), sin(alpha))
    for (j in 1:K) {
        beta <- angles[j]
        b <- c(cos(beta), sin(beta))
        ua <- colSums(u * a)  ## Inner products of cols of 'u' with 'a'
        ub <- colSums(u * b)  ## Inner products of cols of 'u' with 'b'
        good <- abs(ua) > p & abs(ub) > p  ## Sets the topology to that of S^3
        N <- sum(good)
        q <- x[good]
        w <- y[good]
        e <- cbind(q, w)  ## The spin directions observed by Alice and Bob
        write.csv(e, file = "AliceDirectionse1.txt")
        write.csv(-e, file = "BobDirectionse1.txt")
        ad <- read.csv("AliceDirectionse1.txt")  ## alice directions file
        v <- rbind(ad$q, ad$w)           ## just the x and y coordinates as a 2xN matrix
        rownames(v) <- c("q", "w")
        bd <- read.csv("BobDirectionse1.txt")   # bob directions file
        bv <- rbind(bd$q, bd$w)           # just the x and y coordinates as a 2xN matrix
        rownames(bv) <- c("q", "w")
        va <- colSums(v * a)  ## Inner products of cols of 'v' with 'a'
        vb <- colSums(bv * b)  ## Inner products of cols of 'bv' with 'b'
        corrs[i, j] <- sum(sign(va) * sign(vb))/N
        Ns[i] <- N
    }
}

(N)

par(mar = c(0, 0, 2, 0))
persp(x = angles, y = angles, z = corrs, zlim = c(-1, 1), col = "pink", theta = 135,
    phi = 30, scale = FALSE, xlab = "alpha", ylab = "beta")

And the result is,
Image
FrediFizzx
Independent Physics Researcher
 
Posts: 2905
Joined: Tue Mar 19, 2013 7:12 pm
Location: N. California, USA

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Mon Jun 16, 2014 1:10 am

FrediFizzx wrote:Below is a version of Joy's simulation showing that creating the direction files for Alice and Bob that they observe and then reading them back in works in the simulation for when the angles a and b are random.

It is good to know that by feeding the spin directions back into the simulation the correlation surface is reproduced exactly. This proves that the spin directions we extracted from the simulation are correct, and Richard is simply doing his calculations incorrectly by fixing the measurement directions a and b "outside the loop."
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Mon Jun 16, 2014 7:32 am

Joy Christian wrote:
FrediFizzx wrote:Below is a version of Joy's simulation showing that creating the direction files for Alice and Bob that they observe and then reading them back in works in the simulation for when the angles a and b are random.

It is good to know that by feeding the spin directions back into the simulation the correlation surface is reproduced exactly. This proves that the spin directions we extracted from the simulation are correct, and Richard is simply doing his calculations incorrectly by fixing the measurement directions a and b "outside the loop."

Interesting. Here Christian admits that his model works by exploiting the conspiracy loophole: the measurement directions need to be known "inside the loop". [I first wrote "inside the loophole"!]

Incidentally, two of our three adjudicators, Hans de Raedt and Gregor Weihs have come out in support of me: the files of directions which Christian submitted did not win the 10 000 Euro challenge - I talk here of the challenge as formulated, not as how it might have been in another parallel universe.

We are still waiting for Andrei Khrennikov, who is probably still recuperating after the excitement of last week, and moreover, I imagine, is a bit computer-challenged, compared to the other two. So no official pronouncement yet.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby gill1109 » Mon Jun 16, 2014 7:45 am

FrediFizzx wrote:
gill1109 wrote:
FrediFizzx wrote:gill1109 said "What's in a name? A variable is a variable. The word "variable" belongs to mathematics. When we call a variable hidden we are adding physical interpretations to a mathematical model. Not directly observable yet "behind the scenes" causing things to be how they are."
Do you still think the experimenters in the real experiment need to know something or anything about the hidden variables? If so, why?

The experimenters need to follow unambiguous instructions. They don't need to "know" about the different theories they are testing. Not once the experimental protocol is written down.

So they don't need to know anything about the hidden variables at all. Correct? If you don't think that is correct, then please explain why they would need to know about the hidden variables.

It seems pretty simple to me. They record the angular momentum directions at particular angles for a and b, then take those results and calculate the correlations per Joy's eq. (16). That is all they have to do. There is nothing else. Their results will either show -a.b, or not or maybe they might discover something inbetween. They don't need to know anything about the simulations if they don't want to. But if you are trying to simulate what Nature might be doing to suit a certain theory, you will have to have those hidden variables in the simulation that the theory says are in Nature.

Now sure, Joy may simulate the hidden variables of his model to simulate the data which his experiment might generate. But unfortunately his simulation has failed to generate data having the properties which he said the data coming out of his experiment would have. Once the experiment is done, some video files are processed. Directions u and v are calculated using image reconstruction software. Then, correlations are computed according to formulas which Christian wrote down on page 4 of his experimental paper. There are no hidden variables involved on the experimental side after the video files have been created. From then on we are in flatland ...
Ok, that is the last time I am going to explain it. Either you get it or you don't.

I can read what is written on page 4 of the experimental paper. I can't make anything different out of it from what I made out of it, and described in the challenge formulated elsewhere on this forum. Christian spent weeks trying to win the challenge as formulated. Now he claims he has won by doing something completely different to what is written in his experimental paper, and moreover something we all here only have known about since a few months ... since I decoded Pearle (1970) and expressed it as a computer simulation.

Seems a bit fishy to me. But everyone can believe just what they like, sure.

Now for sure, Christian can simulate his hidden variables in order to simulate the data which he believes his experiment would generate. But in that experiment, after the video films have been taken of the exploding balls, everything is in flatland. Some image processing software is used to calculate directions of angular momentum u and v of some macroscopic objects. Then a calculation is done following formulas in the paper. Flatland formulas. No more hidden variables.
gill1109
Mathematical Statistician
 
Posts: 2812
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Mon Jun 16, 2014 8:50 am

I am afraid I am unimpressed by Richard Gill’s comments above. In my opinion I have defeated his challenge decisively, by reproducing the entire correlation surface predicted by QM correctly. What is more, I have been able to accomplish this by providing N pre-selected spin directions, as demonstrated in this simulation. This suggests that my proposed experiment will be a spectacular success. It will reproduce the strong correlations exactly as I have predicted in my papers. Contrary to what Richard Gill says, my model has nothing whatsoever to do with the detection loophole or conspiracy theory. It simply points out the correct topology of the physical space, which has been ignored by the Bell community---rather surprisingly---for the past 50 years, as I explain here: http://arxiv.org/abs/1405.2355.

The evidence I have presented so far has been neatly summarized in this simulation of my 3-sphere model: http://rpubs.com/jjc/16567.

The most important conceptual point here is that, what appeared as a "post-selection" of states to the flatlanders like Richard Gill, has now been formulated as a "state preparation" of the pre-selected initial states of the system, originating at the source. This was never a problem from the perspective of the 3-sphere, but the flatlanders were having a great deal of difficulty understanding my model because of it. But now the N spin directions, "v", which are actually observed by Alice and Bob, are all pre-selected at the source (rather than post-selected at the detector), thereby removing the last vestige of the Bell-baggage from the simulation.

In summary, N spin directions, "v", as an ensemble of the initial or complete states of the physical system, are pre-selected at the source, after the state preparation within the 3-sphere, just as we prepare initial states in quantum mechanics and experiments. The correlations are then calculated using the standard formula,

E(a, b) = (1/N) Sum_(k=1)^(k=N) [ sign(+a.v_k) * sign(-b.v_k) ] = -a.b ,

where the measurement directions a and b are randomly selected, respectively, by Alice and Bob. Note that not a single one of Bell's assumptions has been dropped.

On the other hand, Richard Gill's incorrect calculations of the four correlations contradict the evidence presented in this simulation. Note that two of his four correlations do not reside on the correlation surface presented in the simulation. Evidently, he has got his calculations badly mixed up, as I explained previously.
Last edited by Joy Christian on Mon Jun 16, 2014 9:40 am, edited 1 time in total.
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Heinera » Mon Jun 16, 2014 9:39 am

And the fact that the pre-selection at the source uses a formula that depends heavily on detector settings a and b doesn't bother you at all?
Heinera
 
Posts: 917
Joined: Thu Feb 06, 2014 1:50 am

Re: Another Response to Richard Gill's 10,000 Euros Challeng

Postby Joy Christian » Mon Jun 16, 2014 9:49 am

Heinera wrote:And the fact that the pre-selection at the source uses a formula that depends heavily on detector settings a and b doesn't bother you at all?

Why should the fact that the pre-selection of N vectors "v" at the source does not depend at all, even lightly, on the future detector settings a and b bother me?
Joy Christian
Research Physicist
 
Posts: 2793
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

PreviousNext

Return to Sci.Physics.Foundations

Who is online

Users browsing this forum: No registered users and 81 guests

cron
CodeCogs - An Open Source Scientific Library