At temperature one, adding heading weight 0.5 lowers mean translation-heading cost by 0.02884, but changes endpoint gain by −0.01260 and increases absolute yaw/gaze change. The 30 September screen therefore supplies a tradeoff, not a preferred selector or a human-realism result.
Eight cost-selected scenes each received eight arms: temperatures 0.25, 1, 4 and 16, crossed with heading weights zero and 0.5. All 64 attempts produced validated stores and independent endpoint audits. Only 62 reached H16: both cold-temperature arms on scene 83036 stopped after six acquisitions. Successful receipts must not be mistaken for full-length completion.
Open the decisive interactive selector comparison. Read the heading-cost contrast alongside endpoint gain and prefix support. At fixed temperature, adding heading weight 0.5 changes no paths at T=4 and changes one of eight scene paths at T=16. The cold-temperature survivor contrast covers seven scenes per arm and remains separate from the eight-scene operational endpoint comparison.
The trajectory companion explains target-relative paths, bends and wrapped bearing. These have acquisition order but no physical duration: neither a reduced heading proxy nor a visually smooth path establishes speed, acceleration, safety or human realism.
Next: predeclare the intended tradeoff and evaluate matched endpoints and continuation support before selecting a motion preference. Saved-state reweighting estimates conditional immediate expectations, not new successor trajectories. No selector is promoted.
Use the current pinned bundle readback command for the complete archived E28 publication, including all 64 attempts. The frozen guide below retains paths from before relocation.
The frozen
protocol and readback commands expose
aria_nbv.reporting.selector_screen, the publication pin and
captured summaries. Portable verification does not refresh original
raw-store seals; live deep readback requires exact original paths and
receipts. The single seed and cost-selected roster limit population
inference.
| Finding / evidence | RQ | Thesis section | Coverage / gap |
|---|---|---|---|
| Heading cost versus endpoint tradeoff | RQ1 | ssec:selected-camera-coverage | Partial: camera coverage explained; full selector screen remains supporting evidence |
Previous primary report retains its complete narrative, exact tables and original figures. This 15 October revision is an authored synthesis, not a new experiment or a change to the numerical snapshot. Interactive temperature and heading readout.
Experiment index and evidence routes. Missing/private parents remain explicit in the native campaign owner; historical versions retain their original dates and scientific identities.
Scientific snapshot and attempt identities remain those of the cited experiment. Report synthesis: 15 October 2026; restored to the domain/date owner: 16 October 2026.