Archive a lazy-position-only experiment so the study can answer the missing question from the original plan: whether lazy `position` materialization is the real cost by itself, rather than only in combination with shared props. This snapshot keeps the candidate code, its report, the baseline comparison, and the matching `16 MiB` smoke artifact together in one place. That makes the result easy to audit later against the exact implementation that produced it. The result is a clear rejection. Against `current-baseline`, the candidate lands at about -112.62 percent target timing, about -185.19 percent memory, and a -124.96 percent worst critical timing case. That is worse than the combined lazy-position-plus-shared-props lane, which makes the getter-based lazy `position` strategy itself look like the dominant problem on this code path. Methodology stays consistent with the rest of the study: 10 timing runs, 5 memory repeats, bootstrap comparison against the checked-in baseline report, and a comparable `16 MiB` mixed-article smoke artifact. Verification: the archived comparison comes from `experiments/event-shape-study/lazy-position/artifacts/comparison.txt` and the branch preflight now recognizes the new approach surface