> Data revision: baf7c2f4307d3a989a8fb9363df8000fda3d9310
> Source files:
- [rounds/kh-climb/README.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/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)

# kh climb

**Round ID:** kh-climb  
**Status:** DESCRIPTIVE

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

> 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: DESCRIPTIVE; KEPT FOR AUDIT
Why not VALID: The historical inventory retains DESCRIPTIVE because of DENOMINATOR_UNRESOLVED, PRE_REGISTRATION_MISSING, RAW_EVIDENCE_MISSING. 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

**DESCRIPTIVE**

### Why not VALID
- **PRE_REGISTRATION_MISSING:** No pre-registered rule predating the first result is established.
- **DENOMINATOR_UNRESOLVED:** The exact V3 plan and cohort denominators are not recovered.
- **RAW_EVIDENCE_MISSING:** Exact-ID sanitized raw records are absent.


## Required reproduction metadata

- Kogen commit: not recorded as a resolvable commit in the cited round source; the source carries the unresolved identifier `b3009982`.
- Harness commit: not recorded in the cited round source as a Git commit; any adapter or runtime fingerprints are not Git commits.


Terminology: [public round glossary](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/GLOSSARY.md).

STATUS: **INTERIM**

Data completeness: **N/A** (no exact-ID per-cell records in the public exports). [Measurement record](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/MEASURED.md); [declared gaps](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/MISSING.md).

Pre-registered: not established in the public record; treat this as retrospective exploratory evidence.
Hypothesis coverage: H21, H32, H33 (provisional mapping; see the hypothesis register).
Question: Did one-factor changes to the kh coding loop improve full-task completion over the same-binary baseline on the selected development set?
Population: 21 planned per arm for baseline and the named V1/V2/V4/V5 screens; exact V3 plan not recovered.
Headline: No promotion claim; all reported numeric results are documentary only.

## Reported observations

- The B0 baseline is reported at 12/21 passes; unchanged H0 is reported at 14/21, a two-pass difference described as the same-binary noise floor. ([reported data](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); source-reported, not reproducible from public data.)
- V1 boundary prompt has a denominator conflict: one summary reports 10/18 usable against B0 12/18; another reports V1 as 10 pass, 11 fail, 0 infrastructure outcomes across 21. These are not reconciled. ([reported data](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); source-reported, not reproducible from public data.)
- V2 mix-test is reported at 10/21 passes. ([reported data](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); source-reported, not reproducible from public data.)
- V4 verification-gate screen: 6/10 started versus H0 7/10; 11 planned cells were unstarted. ([reported data](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); source-reported, not reproducible from public data.)
- V5 requirement-to-diff screen: 2/6 started versus H0 4/6; 15 planned cells were unstarted. ([reported data](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); source-reported, not reproducible from public data.)
- V3 was built but not run. No variant was selected. ([reported data](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); source-reported, not reproducible from public data.)

## Reproduction and records

| Required field | Publicly recoverable value |
|---|---|
| Harness/Kogen revision | kh@b3009982 is reported as an 8-character prefix. A full exact SHA and public source mapping were not recovered. |
| Model and effort | gpt-6-luna, requested max, as reported. Effective effort attestation is not available for every cell. |
| Task IDs | Seven development tasks are reported; the exact public task IDs are not recovered. |
| Command | Not recovered. The launch command and exact selected cell IDs are absent from the public snapshot. |
| Raw records | No sanitized raw records for this exact round ID are in this snapshot. The exact-ID query returned zero rows in both `results/cells.jsonl` and `results/run-records/index.json`; see [reported-data.json](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json). |

The numerical observations above are documentary summaries, not a public exact-cell analysis. `reported-data.json` records each reported quantity and its status. The current public result files were checked for this exact round identifier; they do not provide this round's raw records.

## Interpretation

**Limitation:** These exploratory single-factor screens do not establish a benefit for a verification gate or requirement-to-diff step. V4 and V5 are partial, V3 is not run, and V1 counts conflict.

**Smallest useful next test:** Recover exact cell IDs and one reconciled grading ledger, then run a preregistered matched comparison with completed denominators and an independent task set.

No individual outcome should be promoted beyond the status above until the exact run records, grade identities, denominators and source revisions are reconciled.

Sources: [reported data file](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/reported-data.json); [public cells export](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/cells.jsonl); [public captured-delivery export](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/index.json).

## Studio source recovery

Studio retains `state-current.json` and log rows only. Log content is excluded from the mined records, and no per-cell model manifests were present. This source remains in [the source inventory](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/data/mined/STUDIO-SOURCES.json); the round’s published status and numbers are unchanged.

## Recomputed from raw records

[`recomputed.json`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/kh-climb/recomputed.json), [mined cell records](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/data/mined/kh-climb.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 18 rows (fail 11, pass 7); overall pass rate is 38.9% (7/18) 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; 18 missing published metadata). Missing counters remain unknown, not zero.
This round also has 18 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/)
- [H32](/hypotheses/h32/)
- [H33](/hypotheses/h33/)

### Related experiments

- [kh-climb](/rounds/kh-climb/)

### Related claims

- [CL-WA-kh-climb-reported-kh-climb-baseline](/claims/cl-wa-kh-climb-reported-kh-climb-baseline/)
- [CL-WA-kh-climb-reported-kh-climb-control](/claims/cl-wa-kh-climb-reported-kh-climb-control/)
- [CL-WA-kh-climb-reported-kh-climb-v1](/claims/cl-wa-kh-climb-reported-kh-climb-v1/)
- [CL-WA-kh-climb-reported-kh-climb-v2](/claims/cl-wa-kh-climb-reported-kh-climb-v2/)
- [CL-WA-kh-climb-reported-kh-climb-v3](/claims/cl-wa-kh-climb-reported-kh-climb-v3/)
- [CL-WA-kh-climb-reported-kh-climb-v4](/claims/cl-wa-kh-climb-reported-kh-climb-v4/)
- [CL-WA-kh-climb-reported-kh-climb-v5](/claims/cl-wa-kh-climb-reported-kh-climb-v5/)
