New clocked EPR Simulation with 100% detection.

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

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 1:04 am

minkwe wrote:You are using the same source file each time with a different setting, which is equivalent to measuring the same particle more than once. I have already explained to you previously that this is a meaningless exercise which can never be measured in any doable experiment.

This thought experiment turned out to be a very meaningful exercise, because it led to a mathematical theorem that applies to your simulation model.

Theorem: your computer program cannot achieve *both* the good value of CHSH *and* a coincidence rate of above 90% (or whatever the threshold is which Jan and I derived).

Moreover, it is not only a thought experiment which can be done in the mind, it is a computer experiment which can can be easily done on your own computer! I already did it and sent you the files.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 1:12 am

minkwe wrote:I do not see any direct response here to the mathematical arguments I made against your paper.

Your arguments are irrelevant. You are not talking about the same things that we are talking about.

minkwe wrote:How can an intersection of 4 disjoint sets not be null?

The four sets *defined in our paper* have a non-null intersection. We assume local realism (local hidden variables). You agree that your model is a local hidden variables model, and allows complete freedom to choose settings (and setting switches) at the will of the experimenters?

minkwe wrote:Secondly if you read my README file you will notice the warning that the particles are not necessarily in pairs only being paired 99.9% of the time.

It should not be difficult for you to fix that defect of your program. Please add in the output lists also identification of the "lost" particles. I am talking about a simulation with 100% detection. "New clocked EPR Simulation with 100% detection." [added later: I fixed this defect myself. I do have truly 100% detection].

minkwe wrote:Thirdly, you are using the same source file each time with a different setting, which is equivalent to measuring the same particle more than once. I have already explained to you previously that this is a meaningless exercise which can never be measured in any doable experiment.

But it is possible to do this with your programs on my computer. It turns out to be a meaningful exercise, because it leads to an interesting deduction.

minkwe wrote:In any case, if this is what your paper is all about, then it confirms my claim that you are talking about impossible experiments.

I did real experiments on your programs on my computer. And you could have done them too. We are not talking about impossible experiments.

Jan-Ake and I proved theorems about the limitations of your programs, through the device of a Gedanken Experiment. Are you saying that Einstein Podolsky and Rosen were fools, for writing a paper about a Gedanken Experiment?

The experimental results which you and I have obtained with your programs confirm our predictions.
Last edited by gill1109 on Sun Mar 02, 2014 2:44 am, edited 1 time in total.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 2:43 am

I would like to put in another request for a new feature to these excellent programs: the possibility to lower the rate of emissions from the source. One can run the experiment for a longer time period, to compensate. When one has done this, one knows pretty much for sure which pairs of detections belong to one another. No events need to be discarded.

The experiment then becomes equivalent to the earlier kind we were simulating (pulsed experiments), except that there are no non-detections! CHSH will drop to below 2.

The other feature which is needed is the ability to freely switch settings at any times, and to any values, chosen in any way they like, externally, by the experimenters.

Of course thanks to the deterministic nature of what goes on in "station.py" it is possible to arrange this by running the program many times on the same set of source particles, but with different settings - and then just pasting different parts of the outputs together.

This is what I have effectively done in recent computer experiments with Minkwe's model.

Actually first I improved the detectors in his model: they are now, for me, 100% efficient, not 99.99%. Makes life a whole lot more simple without fundamentally changing the results.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 4:46 am

PS since my detectors are now 100% efficient and since Michel's data files maintain the order that the particles were generated in, I can easily pair *all* events. Now I get a perfect loophole free experiment ... but unfortunately it has CHSH <= 2, up to statistical error of course.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Sun Mar 02, 2014 9:20 am

gill1109 wrote:
minkwe wrote:I do not see any direct response here to the mathematical arguments I made against your paper.

Your arguments are irrelevant.

Others can judge that and see that I'm exactly correct and you have provided no counter argument other than the assertion without specific justification that my argument is irrelevant. You do not say which part of my argument is wrong and why, as should be expected if your paper could be saved.

minkwe wrote:Secondly if you read my README file you will notice the warning that the particles are not necessarily in pairs only being paired 99.9% of the time.

It should not be difficult for you to fix that defect of your program. Please add in the output lists also identification of the "lost" particles. I am talking about a simulation with 100% detection.

You know very well (or should know), based on our previous discussions AND the README file describing my simulation that the fact that only 99.9% of particles are produced in pairs is a FEATURE of the of my simulation, not a defect. Those lines of code did not accidentally fall in there, I put it there on purpose. If you remove it, you are no longer talking about my simulation. That you would call it a defect while knowing fully well what I have said clearly about it, is revealing. If you ignore my warnings and remove features from my simulation, you would be introducing additional assumptions and any statements you will be making about the result will be deliberate straw-man arguments.

In my simulation, 100% of particles are detected. This is easy to verify using the source file and the station output file. So what you are saying is not true. By modifying the source, you have removed one of the hidden variables.

[added later: I fixed this defect myself. I do have truly 100% detection].

No, there was no defect to begin with, 100% of emmitted particles were already detected. All you did was to eliminate a hidden variable in the source which made sure only 99.9% of particles were emitted in pairs, well above any existing threshold. So now you have a completely different simulation with the additional assumptions and different hidden variables.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 9:23 am

Michel says that I should not have made the detectors 100% effficient. He does insist on this point in the readme file, too. His slightly less than 100% efficiency comes from removing particles if a certain drawing from a normal distribution is more than 4 standard errors away from the mean.

Instead of going for 100%, one could also just change the parameter "4.0" into "10.0". That would make the detectors 99.99....9% efficient, but still less than 100%.

I am quite mystified by why this is such a major issue for Michel, I hope he will explain. The title of this thread, which he started, is "New clocked EPR simulation with 100% detection".
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Sun Mar 02, 2014 9:30 am

gill1109 wrote:Michel says that I should not have made the detectors 100% effficient. He does insist on this point in the readme file, too.

This is false. My detectors are 100% efficient. Any one can verify this. Please read the readme file again.

99.9% of the particles emitted are paired, 100.0000000000000000000000% of emitted particles are detected. This is clearly stated.
Last edited by minkwe on Sun Mar 02, 2014 9:32 am, edited 1 time in total.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 9:31 am

One person's feature is another person's bug! And vice versa.

So you have now explained that when you say 100% detection, you mean that all particles which you intended to be emitted are indeed detected too. But you have told us that 0.01% of the particles do not come in pairs. So your source is imperfect.

I made it perfect. Less realistic, I know. You need to add a whole lot more single particles to make your model realistic by current standards.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 9:32 am

Suppose I change "4.0" to "6.0". Is that a "completely different simulation"?
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Sun Mar 02, 2014 9:37 am

gill1109 wrote:I made it perfect. Less realistic, I know. You need to add a whole lot more single particles to make your model realistic by current standards.

No you did not. My source is already perfect. It behaves the way it is meant to behave. You don't have to like it, but that is the model. You changed the model. You are free to have your own standards of how a source should behave but results you compute according to those assumptions will not tell you anything about my simulation , or the real world experiments it attempts to model.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Sun Mar 02, 2014 9:47 am

gill1109 wrote:Suppose I change "4.0" to "6.0". Is that a "completely different simulation"?


It doesn't matter, what you really want to do is to introduce the assumption that all particles are paired. So that you may analyse the data with that assumption in mind. Once you introduce that assumption, you now have a different simulation. There are many ways to introduce that assumption:

(1) you can modify the source file to remove the code which does that elimination.
(2) you can increase the number of sigmas and run a very short simulation
(3) you can post process the source files to remove unpaired particles before sending them to the station.

All of those accomplish the same thing.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Sun Mar 02, 2014 9:51 am

gill1109 wrote:You need to add a whole lot more single particles to make your model realistic by current standards.

My model is 100% realistic and 100% local. This is easy to see.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Sun Mar 02, 2014 10:19 am

gill1109 wrote:PS since my detectors are now 100% efficient and since Michel's data files maintain the order that the particles were generated in, I can easily pair *all* events. Now I get a perfect loophole free experiment ... but unfortunately it has CHSH <= 2, up to statistical error of course.

My detectors were already 100% perfect, so you are clearly wrong. I would suggest that you make the same assumption that all particles are paired in any real experiment of your choice, including the recent Giustina experiment which claims to have "closed the coincidence loophole". I suggest you use the same data analysis procedure and assumptions you just introduced, and post the results here. It will be revealing.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 9:12 pm

minkwe wrote:
gill1109 wrote:PS since my detectors are now 100% efficient and since Michel's data files maintain the order that the particles were generated in, I can easily pair *all* events. Now I get a perfect loophole free experiment ... but unfortunately it has CHSH <= 2, up to statistical error of course.

My detectors were already 100% perfect, so you are clearly wrong. I would suggest that you make the same assumption that all particles are paired in any real experiment of your choice, including the recent Giustina experiment which claims to have "closed the coincidence loophole". I suggest you use the same data analysis procedure and assumptions you just introduced, and post the results here. It will be revealing.


Sure, your detectors were perfect, but your source was not perfect.

I did not change the data analysis procedure.

I am not making any assumption that *all* particles are paired in any real (quantum optics lab) experiment.

I am explaining how we can simulate such a situation with your program, provided of course we ignore your "warning" not to perfect the source.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 9:15 pm

minkwe wrote:
gill1109 wrote:Suppose I change "4.0" to "6.0". Is that a "completely different simulation"?

It doesn't matter, what you really want to do is to introduce the assumption that all particles are paired. So that you may analyse the data with that assumption in mind.

I do not *want* to analyse the data with that assumption in mind. I do *not* analyse the data with that assumption in mind. I analyse the data with the algorithm which you took from Jan-Ake Larsson's program: compute all absolute differences between times of events left and right; order them; pick the first, pair the two corresponding events; remove these events from the lists of events on both sides; and repeat...

So are you telling me that I am allowed to change 4.0 to 6.0, but I am not allowed to change 4.0 to infinity?
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 9:23 pm

minkwe wrote:
gill1109 wrote:I made it perfect. Less realistic, I know. You need to add a whole lot more single particles to make your model realistic by current standards.

No you did not. My source is already perfect. It behaves the way it is meant to behave. You don't have to like it, but that is the model. You changed the model. You are free to have your own standards of how a source should behave but results you compute according to those assumptions will not tell you anything about my simulation , or the real world experiments it attempts to model.

Your source is not perfect, it also emits unpaired particles. Ask any experimenter which source they would prefer to have.
Regarding the fit to real world experiments: do you think that changing 0.01% of the observed data makes a huge difference to the results of *your* analysis of the data? If your model produces results which mimic closely real world data, then that remains the fact after 0.01 percent of events in the data are altered.

Your experiment generates data with the numbers of events on each side equal to one another to within 0.01 percent. I don't believe real world experiments often got such a perfection. However I am not complaining about this un-realistic aspect of your simulation model and I am not using a pairing of events simply by sequence number to analyse the data - I analyse the data using your algorithms which you took from Jan-Ake which he took to imitate standard laboratory practice.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Sun Mar 02, 2014 10:19 pm

Let me remind you of the Larsson-Gill result: it is only possible to get CHSH approx equal to 2 sqrt 2 with a simulation like yours, with a coincidence rate below 87.87%.

In other words: it is not possible to get the coincidence rate to above 87.87%, *and* CHSH = 2.828

Both numbers refering to the limit as the experiment gets indefinitely large. There can obviously be chance fluctuations in both numbers in a finite experiment - you know the 1 over square root of N law, of course.

Your own simulation results confirm our predictions.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Mon Mar 03, 2014 5:39 am

minkwe wrote:
Larsson & Gill wrote:Correlations are obtained on subsets of Λ, namely on

ΛAC′ , ΛAD′ , ΛBC′ , or ΛBD′ . (iv)

Then

E(AC′|ΛAC′ ) + E(AD′|ΛAD′ ) + E(BC′|ΛBC′ ) − E(BD′|ΛBD′ ) ≤ 4 − 2δ. (11)

Proof. The proof consists of two steps; the first part is similar to the proof of Theorem 1,
using the intersection ΛI = ΛAC′ ∩ ΛAD′ ∩ ΛBC′ ∩ ΛBD′ , (12)

on which coincidences occur for all relevant settings. This ensemble may be empty, but only
when δ = 0 and then the inequality is trivial, so δ > 0 can be assumed in the rest of the proof.


E(AC′|ΛI) + E(AD′|ΛI) +E(BC′|ΛI) − E(BD′|ΛI) ≤ 2. (13)


Now it is easy to see that the set is empty: None of the particle pairs in any of the 4 sets measured in any EPR-experiments ever performed or performable in the future is a member of any of the other sets. The sets used for calculating each of the terms is disjoint from all the others, therefore ΛI is a null set. Do you deny this?

No response.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby minkwe » Mon Mar 03, 2014 5:40 am

minkwe wrote: I would suggest that you make the same assumption that all particles are paired in any real experiment of your choice, including the recent Giustina experiment which claims to have "closed the coincidence loophole". I suggest you use the same data analysis procedure and assumptions you just introduced, and post the results here. It will be revealing.

No response.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: New clocked EPR Simulation with 100% detection.

Postby gill1109 » Mon Mar 03, 2014 5:51 am

minkwe wrote:
minkwe wrote:
Larsson & Gill wrote:Correlations are obtained on subsets of Λ, namely on

ΛAC′ , ΛAD′ , ΛBC′ , or ΛBD′ . (iv)

Then

E(AC′|ΛAC′ ) + E(AD′|ΛAD′ ) + E(BC′|ΛBC′ ) − E(BD′|ΛBD′ ) ≤ 4 − 2δ. (11)

Proof. The proof consists of two steps; the first part is similar to the proof of Theorem 1,
using the intersection ΛI = ΛAC′ ∩ ΛAD′ ∩ ΛBC′ ∩ ΛBD′ , (12)

on which coincidences occur for all relevant settings. This ensemble may be empty, but only
when δ = 0 and then the inequality is trivial, so δ > 0 can be assumed in the rest of the proof.


E(AC′|ΛI) + E(AD′|ΛI) +E(BC′|ΛI) − E(BD′|ΛI) ≤ 2. (13)


Now it is easy to see that the set is empty: None of the particle pairs in any of the 4 sets measured in any EPR-experiments ever performed or performable in the future is a member of any of the other sets. The sets used for calculating each of the terms is disjoint from all the others, therefore ΛI is a null set. Do you deny this?

No response.

My response is, yet again, that you are wrong and clearly haven't read Larsson-Gill closely enough, nor taken any notice of my explanations here. So I don't feel obliged to add anything to that.
gill1109
Mathematical Statistician
 
Posts: 2049
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

PreviousNext

Return to Sci.Physics.Foundations

Who is online

Users browsing this forum: ahrefs [Bot] and 7 guests

CodeCogs - An Open Source Scientific Library