Finite records do not determine the global cone

This is a counterexample to reconstruction from the data of this experiment, not a universal impossibility theorem for all possible geometric measurements.

The same distinction applies to lapse comparisons. In an enlarged comparison class

with constant $N=1$ or $N=2$, the same edges satisfying the unit-lapse causal bound remain causal and the same matrices yield the same channels. Static proper durations differ by a factor of two. The main release still fixes $N=1$; this comparison does not silently change that input.