Complete binary records and relation to Em03-C

For independent repeated preparations, a binary sample mean has

This follows by the exponential-moment bound for independent variables in $[0,1]$, followed by Markov inequality. Thus the confidence interval for the ideal packet probability enlarges by ${ϵ}_{\mathrm{local}}$. Independence of repeated preparations is not inferred from the one-shot construction.

The outcome says whether a specified bounded exterior observable departs from its stated baseline. It does not identify the cause as quantum gravity, microscopic information recovery, or evaporation.