Karma Peny Simulation Appears to Violate Bell

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

Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Fri Nov 22, 2019 6:48 pm

After much effort the simulation at https://codeserver.net/bell/2/ now appears to violate Bell's in-equalities using CHSH experiment.

Math is based on Karma Peny's presentation here: https://www.youtube.com/watch?v=yOtsEgbg1-s

Modes are:
Quantum (simulated 'spooky action at a distance', random 50% detection)
Classic (local hidden variables, cos²(θ) probability match, |cos(2θ)| probability detection)
Karma Peny (local hidden variables, cos²(θ) simple match, |cos(2θ)| probability)

The Quantum mode is nothing special (particles communicate), and achieves an almost perfect 2.8 CHSH score.

But the other 2 modes Classic and Karma Peny are special, the particles have "local hidden variables" that do not communicate (each detection knows only properties of 1 particle and 1 detector).

The Classic Model exceeds QM expected values at approximately 2.5.
The Karma Peny model actually far exceeds the QM expected 2.8 with values in the 3+ range.

Image

Simulation at https://codeserver.net/bell/2
Code is available at https://github.com/tankersley/bell
Last edited by jtankersley on Fri Nov 22, 2019 7:59 pm, edited 1 time in total.
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Fri Nov 22, 2019 7:12 pm

Math used to apparently violate Bell's Inequalities with local hidden variables with CHSH experiments
('Classic' and 'Karma Peny' modes of https://codeserver.net/bell/2/).

Classic (minor variant of Karma Peny)
Code: Select all
        function getClassicResult(prt,pol,debug) {
            let result=0;
            const delta=Math.abs(Math.abs(pol.axis)-Math.abs(prt.axis)), deltaCos=Math.abs(Math.cos(delta*Math.PI/180)), posProb=deltaCos*deltaCos;
            result=(Math.random()<=posProb)?1:0;
            debug.math='getClassicResult(prt,pol)';
            return result
        }

        function detectClassic(prt,pol,debug) {
            const delta=Math.abs(Math.abs(pol.axis)-Math.abs(prt.axis)), delta2Cos=Math.abs(Math.cos((delta+delta)*Math.PI/180));
            return (Math.random()<=delta2Cos)?true:false;
        }


Karma Peny as presented here: https://www.youtube.com/watch?v=yOtsEgbg1-s
Code: Select all
        function getKarmaPenyResult(prt,pol,debug) {
            let result=0;
            const delta=Math.abs(Math.abs(pol.axis)-Math.abs(prt.axis)), deltaCos=Math.abs(Math.cos(delta*Math.PI/180)), deltaCos2=deltaCos*deltaCos;
            result=deltaCos2>=0.5?1:0;
            debug.math='getKarmaPenyResult(prt,pol)';
            return result;
        }

        function detectKarmaPeny(prt,pol,debug) {
            const delta=Math.abs(Math.abs(pol.axis)-Math.abs(prt.axis)), delta2Cos=Math.abs(Math.cos((delta+delta)*Math.PI/180));
            return (Math.random()<=delta2Cos)?true:false;
        }
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby local » Fri Nov 22, 2019 7:47 pm

Your code link is broken. Anyway, your simulation clearly has missed detections so it is just the good old detection loophole.

Honestly, nobody cares about this. The detection loophole has been known for many decades and protocols to avoid it are also well known and standard in modern experiments. Boring!
local
 
Posts: 83
Joined: Mon Aug 05, 2019 1:19 pm

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Fri Nov 22, 2019 7:56 pm

You may need to clear your browsers Cache.
(On Chrome browser, Right-click on browser and select "Inspect", then Control-Click on "Refresh" button and select "Empty Cache and Hard Reload")

Or try this alternate link: https://codeserver.net/bell/2/chsh100.html
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby local » Fri Nov 22, 2019 8:53 pm

Why don't you address the substance of my remarks? Why is this interesting? There are boatloads of such simulations out there. They are all stupid and boring now.
local
 
Posts: 83
Joined: Mon Aug 05, 2019 1:19 pm

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Fri Nov 22, 2019 9:11 pm

Hello local,

Karma Peny appears to address similar about 2 months ago on the blog under his video (at https://www.youtube.com/watch?v=yOtsEgbg1-s).
They call the 2-sided detections 'coincidences' and the 1-sided detections 'singles'. I'm sorry to say that it is not easy to find data about this. You may wish to examine the two 'loophole-free' Bell test experiments perfored in 2015, one led by Marissa Giustina and the other led by Bas Hensen. In the Bas Hensen experiment they use electron spin instead of photon polarisation. It is a very complicated experiment in which they always get a reading, and so there are no 'singles'. However, I spent one evening analysing their results and found statistically significant figures that indicate that the correlation is caused by some influence from the setting used in the previous trial attempt. In the Marissa Giustina experiment I found that they were not using the Eberhard inequality that they claimed to be using. They had derrived a simpler inequality from it and I could easily violate the simpler inequality with local reality detection probability.


I would be interested in links to similar Bell Violating simulations using local hidden variables that do not communicate.
(It is not particularly difficult to replicate Karma Peny's work, particularly since he is such an effective communicator of the concepts involved, so not too surprised to hear its been done before).
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby Joy Christian » Sat Nov 23, 2019 2:34 am

jtankersley wrote:
I would be interested in links to similar Bell Violating simulations using local hidden variables that do not communicate.

Here is the very first simulation of Pearle's detection loophole model done by Michel Fodje: https://github.com/michel4j/epr-simple.

Later on (in 2015), Richard D. Gill reproduced Michel's simulation in R without giving him credit for priority. You won't have difficulty finding Gill's version.

I have cited Michel Fodje's simulation and several more in my IEEE paper: https://ieeexplore.ieee.org/document/8836453 (Michel's simulation was produced in 2013).

***
Joy Christian
Research Physicist
 
Posts: 2132
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Sat Nov 23, 2019 6:08 am

Thank you Joy,
Richard Gill and his group is reviewing the code implementation of Karma Peny's description now at https://groups.google.com/forum/#!topic ... dvTO7aIOXs
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Sat Nov 23, 2019 11:18 am

Image

The "Classic" mode calculations (graphed above) look particularly interesting, and generate CHSH "S" results around 2.3.

x = θ as difference between photon polarization angle and polarizer angle
Green = Detection Probability, |cos(2x)|
Red = Probability of passing through polarizer (+) instead of being reflected (-), |cos²(θ)|

(this varies from Karma Peny "polarizing" math, which uses linear |cos²(θ)| >= 0.5)
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby local » Sat Nov 23, 2019 2:45 pm

jtankersley wrote:Karma Peny appears to address similar about 2 months ago on the blog under his video (at https://www.youtube.com/watch?v=yOtsEgbg1-s).
They call the 2-sided detections 'coincidences' and the 1-sided detections 'singles'. I'm sorry to say that it is not easy to find data about this. You may wish to examine the two 'loophole-free' Bell test experiments perfored in 2015, one led by Marissa Giustina and the other led by Bas Hensen. In the Bas Hensen experiment they use electron spin instead of photon polarisation. It is a very complicated experiment in which they always get a reading, and so there are no 'singles'. However, I spent one evening analysing their results and found statistically significant figures that indicate that the correlation is caused by some influence from the setting used in the previous trial attempt. In the Marissa Giustina experiment I found that they were not using the Eberhard inequality that they claimed to be using. They had derrived a simpler inequality from it and I could easily violate the simpler inequality with local reality detection probability.

It's very easy to make glib claims in youtube comments. Where are the peer-reviewed publications or other documentation of his "findings"?

BTW, you'll probably not find much interest here in Gill's nonsense.
local
 
Posts: 83
Joined: Mon Aug 05, 2019 1:19 pm

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Sun Nov 24, 2019 1:09 pm

After discussions with Karma Peny (after the discrepancy was discovered by Chantal Roth), the Karma Peny mode now calculates S around 2.8 (realistic mode calculates around 2.4).
(version 1.1.03 https://codeserver.net/bell/2/, paper with formulas https://sites.google.com/site/physicschecker/unsettled-physics/testing-bells-theorem-paper)

The experiment models low detection rates at theta angles (difference in angle between photon polarization angle and polarizer angle) near 45°(in-between vertical and horizontal), causing a false appearance of strong correlations between non-communicating particles.

This is referred to as the "detection loophole", and appears to provide reasonable experimental proof of an alternate explanation for results of most Bell CHSH experiments prior to 2015.
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby local » Mon Nov 25, 2019 4:20 pm

jtankersley wrote: This is referred to as the "detection loophole"

As Dr Christian pointed out, it's been known since Philip M. Pearle (1970). Stop pretending it is some big deal.

an alternate explanation for results of most Bell CHSH experiments prior to 2015

Nobody cares, this is 2019.
local
 
Posts: 83
Joined: Mon Aug 05, 2019 1:19 pm

Re: Karma Peny Simulation Appears to Violate Bell

Postby Joy Christian » Tue Nov 26, 2019 1:19 am

***
Detection loophole discovered by Pearle in 1970 violates some predictions of quantum mechanics. Last year I made a post (in another thread) about which predictions of quantum mechanics are violated by the detection loophole: viewtopic.php?f=6&t=330#p8009.

In my view, Phil's 1970 paper is a brilliant paper but he misinterpreted his findings. His distribution function, in fact, describes the geometry and topology of the 3-sphere.

***
Joy Christian
Research Physicist
 
Posts: 2132
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Sat Nov 30, 2019 5:08 pm

Hi local,

The detection loophole was news to me, I was just 50 years late in learning about it :D, as Joy notes at http://www.sciphysicsforums.com/spfbb1/viewtopic.php?f=6&t=330#p8009.

This computer experiment https://codeserver.net/bell/2 is updated to use more standard symbols, somewhat simplified JavaScript code and comments, slightly improved animation, and should run reliably in modern browsers including smart phones.

Probability math, graphed below, generates S values of 2.4 after 10's of thousands of detections.

Image
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA

Re: Karma Peny Simulation Appears to Violate Bell

Postby jtankersley » Sat Nov 30, 2019 7:40 pm

Correction, S values approaching 2.83 :D (2.80, 2.82 and 2.86 from 3 concurrent experiments) after changing the detection probability to cos²(2Δ) (with polarization probability of cos²(Δ)), after about 10k detections...

Image
jtankersley
 
Posts: 18
Joined: Sun Nov 10, 2019 4:47 pm
Location: Middleton, WI USA


Return to Sci.Physics.Foundations

Who is online

Users browsing this forum: No registered users and 2 guests

cron
CodeCogs - An Open Source Scientific Library