Registration of Two Tests: Assigned versus Chosen Ties (N5) and the Shape of Tie Counts in Free Time over Cost (H8)

This record freezes the estimators and decision thresholds of two tests before any real dataset has been selected or opened. N5 asks whether ties assigned by an institution decay faster than ties a person chose, using a retention ratio with a bootstrap interval and a comparison of exponential and step decay by AIC. H8 asks whether counts of ties above a weight threshold are S-shaped in log(free time / maintenance cost), using a logistic versus log-linear comparison by AIC. Each test states the data it needs, what counts as support and what counts as refutation. The tests accompany a simulated model of finite locality in human relationships; the model's results do not depend on them. The record also lists four withdrawn tests and the rules under which datasets will be named, amendments dated, and all registered analyses reported.