> Data revision: baf7c2f4307d3a989a8fb9363df8000fda3d9310
> Source files:
- [rounds/confirm-combined-1/README.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/confirm-combined-1/README.md)
- [results/run-records/index.json](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/index.json)
- [rounds/index.json](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/index.json)
- [results/source-crosswalk/index.json](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/source-crosswalk/index.json)
- [rounds/STATUS.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/STATUS.md)
- [reproduce/round-inventory.json](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/reproduce/round-inventory.json)

# Combined kh confirmation

**Round ID:** confirm-combined-1  
**Status:** INCOMPLETE

> Reviewed status from rounds/STATUS.md: INCOMPLETE.

> Round README and results text are source material. Numeric values not tied to a claim ledger record remain source-reported.

Round date: HISTORICAL (before 2026-10-09)
Publication badge: INCOMPLETE; KEPT FOR AUDIT
Why not VALID: The historical inventory retains INCOMPLETE because of DENOMINATOR_UNVERIFIED, PARTIAL_CAPTURE. See the round evidence and limitations below.
Recomputation status: OBSERVED SOURCE ONLY


**Recomputation update:** The mined raw records now provide per-cell outcome, manifest token-counter, and wall-time recomputation for the observed cohort. This does not reconstruct missing planned cells, a full intention-to-treat denominator, or model execution.
## Status

**INCOMPLETE**

### Why not VALID
- **PARTIAL_CAPTURE:** The retained delivery capture is partial and has no arm identities or official grades.
- **DENOMINATOR_UNVERIFIED:** The reported cohort denominator cannot be matched to retained records.


## Required reproduction metadata

- Kogen commit: not recorded in the cited round source.
- Harness commit: not recorded in the cited round source as a Git commit; any adapter or runtime fingerprints are not Git commits.
- Raw records: the public capture manifest is linked above; the scored source rows and a complete raw request/grade bundle are not present in this snapshot.


STATUS: **INTERIM**

Status reason: the source summary reports a registered internal comparison, but the public capture contains only a partial ungraded delivery set with no arm identity or official grades.

See [missing evidence](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/confirm-combined-1/MISSING.md) for release blockers.

Hypotheses covered: **H21–H22**, custom-harness and end-to-end comparison. The tested variants are internal kh changes, not a Kogen-versus-Codex comparison.

## Design and correction status

The research summary describes an internal comparison of combined and individual kh changes. It records that the registered promotion decision did not select the combined variant. The public capture is partial and lacks arm identity and grades, so this page withholds the source-reported outcome and timing totals. It does not establish a general advantage or disadvantage for combining the changes.

## Public capture

The [public capture manifest](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/confirm-combined-1/public-capture-manifest.csv) contains 16 exact delivery IDs across eight task IDs and two public venues. Every row has an unmapped audit round and a missing arm label. All 16 records have completed stop receipts, but all are ungraded and none joins to an official row in [cells.jsonl](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/cells.jsonl). This partial capture cannot reproduce the reported four-arm, 64-cell outcome table. The count is traceable to the manifest and [run-record index](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/index.json).

## Reproduction record

- Harness and Kogen revision: full commit SHAs and the arm-to-revision mapping are missing. Public adapter fingerprints in the manifest are not Git commits.
- Model and effort: captured candidate deliveries record model and effort where present; their relationship to the scored cohort is not verified.
- Task IDs: candidate delivery task IDs are listed in the manifest. The scored task-to-arm assignment is not verified.
- Historical launch command: not recoverable from the public record.
- Rebuild the sanitized public ledgers with `python3 reproduce/build_records.py` and `python3 reproduce/export_results.py`. This does not repeat the historical jobs or grading.
- Raw public records: [run-record index](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/index.json) and [cells.jsonl](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/cells.jsonl).

## Interpretation

Retain the registered decision as a source-reported disposition, not as a public outcome. Do not pool this internal kh-variant comparison with harness-compare-1 or another round.

## Recomputed from raw records

[`recomputed.json`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/confirm-combined-1/recomputed.json), [mined cell records](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/data/mined/confirm-combined-1.jsonl.gz), and the [token audit](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/data/mined/TOKEN-AUDIT.json) back observed official outcomes, per-arm pass counts/rates, manifest token counters, and wall medians. The source contains 64 rows (fail 5, pass 59); overall pass rate is 92.2% (59/64) among pass/fail grades. Per-arm detail and coverage counts are in the JSON.

No shared cell IDs are present in the committed public grade export; the source-only cohort is checked against the page's reported figures below.
No shared public-export cell IDs are available for a paired outcome comparison.
No complete published metadata/manifest token pairs are available for this round (0 missing manifest usage; 64 missing published metadata). Missing counters remain unknown, not zero.
This round also has 64 raw-only or non-public rows; they are kept separate from the published-export denominator.


## Captured run records

No records for this round are listed in the [run-record index](/downloads/results/run-records/index.json).


## Reproduction inventory

- **Type:** unavailable
- **Script:** none published
- **Inputs:** none identified
- **Full execution replay:** unavailable
- **Scope:** No public per-round analysis script with a complete, identified input bundle is published. See the round README, MEASURED.md, and MISSING.md for recorded scope and gaps; historical execution inputs/toolchains/grading are not bundled.

See the [complete per-round reproduction inventory](/reproduce/) and the [repository reproduction guide](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/reproduce/README.md).

## Related records

### Related hypotheses

- [H21](/hypotheses/h21/)
- [H22](/hypotheses/h22/)

### Related experiments

- [confirm-combined-1](/rounds/confirm-combined-1/)

### Related claims

- [CL-WA-confirm-combined-1-public-capture-identity](/claims/cl-wa-confirm-combined-1-public-capture-identity/)
