Clauser-Horne/Eberhard inequality violation by a local model

[minkwe]

http://arxiv.org/abs/1507.06231

This finding has implications for the design and interpretation of experiments, as it shows that the coincidence window loophole cannot be eliminated by a demonstration of independence of the detection times and settings.
...
Recent experimental tests of local realism based on the Clauser-Horne/Eberhard inequality are considered and it is shown that in one case (Christensen et al.) the emission rate is appropriately limited to ensure valid counting, and the data confirms local realism; in a second case (Giustina et al.) the experiment fails to appropriately limit the emission rate, and the claimed violation can be accounted for locally.

http://www.ingentaconnect.com/content/a ... 3/art00021

The recent Clauser-Horne (CH)-based Einstein-Podolsky-Rosen-Bohm (EPRB) experiment performed by Christensen et al. is analyzed. The data analysis performed by Christensen et al. is shown to rely upon postselection (data discarding) that produces an artifactual violation of the CH inequality. A corrected analysis that properly includes all of the experimental data is developed, showing that the Christensen et al. experiment confirms locality and disconfirms the quantum joint prediction.

No where left to hide. The application of the CH inequality to experiments is just as flawed as the CHSH. Perhaps we need a new thread to discuss the derivation of the CH inequality to reveal in detail how the same fudging is done in order to compare QM expectations with the inequality to claim phantom violations.

[Joy Christian]

This result is in addition to our latest simulation, which too exhibits manifestly local violations of the Clauser-Horne/Eberhard (or CH) inequality quite explicitly!!!

[Joy Christian]

Let me reproduce here the essential part of the code in the above simulation: http://rpubs.com/jjc/84238. The measurement functions = +/-1 of Alice and = +/-1 of Bob generated in the simulation are exactly the local functions demanded by Bell in his 1964 paper, and the correlations are then calculated using the coincidence counts in the same manner as done in the actual experiments. The displayed plots put to rest any lingering doubt that there might be somehow some uncounted "0 outcomes" in the simulation. But in the 3-sphere there are no "0 outcomes", as is evident from Eq. (B10) of this paper consolidating the analytical model:

Code: Select all

A = +sign(g(a,e,s))  # Alice's measurement results A(a, e, s) = +/-1  # Here g(u,v,s) is a metric on S^3, reducing to the usual g on R^3
B = -sign(g(b,e,s))  # Bob's measurement results B(b, e, s) = -/+1    # A(a, L) = +/-1 and B(b, L) = +/-1 are exactly as defined by Bell
     
Cuu = length((A*B)[A > 0 & B > 0])   # Coincidence count of (+,+) events
Cdd = length((A*B)[A < 0 & B < 0])   # Coincidence count of (-,-) events
Cud = length((A*B)[A > 0 & B < 0])   # Coincidence count of (+,-) events
Cdu = length((A*B)[A < 0 & B > 0])   # Coincidence count of (-,+) events
       
corrs[i,j] = (Cuu + Cdd - Cud - Cdu) / (Cuu + Cdd + Cud + Cdu) # = -a.b

# There are no "0 outcomes" within S^3: Cou = Cod = Cuo = Cdo = Coo = 0
CoB = length(A[g(a,e,s) & A == 0])   # Number of A = 0 events within S^3 (regardless of B events) = 0
CAo = length(B[g(b,e,s) & B == 0])   # Number of B = 0 events within S^3 (regardless of A events) = 0


Joy Christian