Archive the control snapshot alongside the shared-props and lazy-position-plus-shared-props experiments so the study keeps the exact code, reports, and notes that produced each result. The point of this commit is to give the event-shape study a stable baseline and to preserve two early object-layout hypotheses next to the evidence that judged them. That matters because later parser changes would otherwise blur what was actually measured. The results are useful even though both candidates are rejected. Shared props improves retained memory by about 4.02 percent, but the median timing gain only reaches about 0.81 percent, far below the +5 percent bar. Lazy position plus shared props is a clear miss, with about -107.05 percent target timing and a -149.81 percent worst critical case. Keeping those negative results in the archive narrows the search space and makes the study easier to trust. These reports use the same standard settings as the rest of the study: 10 timing runs, 5 memory repeats, and baseline-versus-candidate comparisons backed by machine-readable artifacts. Verification: the commit is archival data built with the already-checked study tooling; no extra executable check was run for this slice