our fr10 advances the pocket before the block is down

Observed. Block held, next pocket saved. The destination fixture was still empty.

Our review needed planned source, confirmed pickup and confirmed placement kept apart. Ask controls to describe your held-block state explicitly, with unknown outcomes left for the agreed recovery decision rather than inferred from the next pocket.

Walk through the opposite gap too, where placement has happened but the saved record has not advanced, before changing when the counter is written

Imran, has the review included which source tray the saved pocket belongs to?

Tray identity added, Victor. Leo, controls and the shift lead own the recovery design together.

Include reopening after the review decision has been saved. Otherwise the first careful recovery can be undone by the next application restart, even if the interrupted case looked right in the initial check.

Both interruption gaps and post-review reopening are in the offline tests. No automatic repeat for uncertain outcomes.

Did the shift lead understand the proposed screen without your explanation? I would be interested in which state was hardest to distinguish.

And do the tests include a late result for the original attempt? That should not silently become confirmation for whatever is next.

Late result keeps its original identity. Screen read-through exposed confusion between planned and confirmed placement; wording still being revised.

Keep the confused example in the handover check, with the operational recovery still subject to the owners' review