Research + data · Updates

Discovery status

Report date: 2026-10-06 08:13 UTC. Version: p7-r4-r1.

The local and live atlas each have 538 cases and 550 stops. They have 76 known ordinary cases left and 64 separate holds.

They also have 22 flight notes and 21 update posts. These counts describe the known roster, not the extent of history.

The search matrix has 1,280 records. Five saved states describe bounded work.

A found record has a located result. It may still need more work.

The new found record is not final. Both source channels are partial, and overlap is unknown.

The other 60 cells with complete channels did not change. No search floor or cycle is complete.

A searched-empty record has no result within its stated scope. It does not prove historical absence.

The current ledger has 53 records. It has 32 candidate records, 12 located leads, seven search leads, and two controls.

Candidate type does not mean 32 unique histories or 32 unadmitted records. Nine admissions are tracked apart.

The cycle has 31 partial attempts in 13 blocks and eight eras. Complete floors and cycles remain zero.

The first test recovered zero of three eligible pairs. Three other pairs fell outside the source scope.

This result does not prove that prompts played no role. Seven later repair records get no first-test credit.

The population index has 179 population records, 345 event links, and five population links. It retains 220 unresolved records.

Its intake has 2,358 name leads. It has 409 fully reviewed leads, 38 partial leads, and 1,911 untouched leads.

These are workflow counts. They do not measure how many peoples the atlas accounts for.

The index keeps a dated 44-record ledger context. The separate current ledger has 53 records.

The isolated series has nine accepted admissions toward its report at 30. None is in the active statistical roster. Two already appear in the atlas. Seven still await separate atlas work.

The series has 685 score rows and 780 corpus objects. The report at 30 is never an admission gate.

The priority table says where a later search might look. It does not say what history was found.

Its 128 inputs and four paired checks are incomplete. They do not steer searches.

A society or language entry is only a lead. It does not prove forced removal or date a loss.

Checkpoint dates

Overlap and yield count record references. Missing values are not zero, and references are not unique histories.

The public table uses opaque search IDs. Internal prompt names are not a public axis.