“…Geometrical and other inefficiencies lead to counters registering (no, no) with overwhelming probability, (yes, yes ) very seldom , and (yes, no) and (no, yes) with probabilities only weakly dependent on the settings of the instruments . Then from (21) P=1-E2 with e2 weakly dependent on the variables , so that ( 16) is trivially satisfied. The authors in general make some more or less ad hoc extrapolation to connect the results of the pratical with the result of the ideal experiment.…”