> Data revision: baf7c2f4307d3a989a8fb9363df8000fda3d9310
> Source files:
- [rounds/r36/README.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r36/README.md)
- [results/run-records/index.json](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/index.json)
- [results/run-records/r36.jsonl](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/r36.jsonl)
- [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)

# r36

**Round ID:** r36  
**Status:** INVALID

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

> 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: INVALID; KEPT FOR AUDIT
Why not VALID: The historical inventory retains INVALID; the available public record does not support upgrading that classification.
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

**INVALID**

- **NO_PREREG** — The README says this round was not pre-registered, and its files show no timestamped rule predating the first result.
- **NO_DECISION_RULE** — The README says no public predeclared decision rule was recovered.
- **RAW_BUNDLE_INCOMPLETE** — The README says the public sources do not provide a complete round-specific request and grade bundle.
- **OUTCOME_CROSSWALK** — The README says official cells outcomes do not match a prior captured pass count.

## 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.
- Model and effort: a complete per-cell mapping is not recorded in the cited round source; values present in the source summary remain as stated above.
- Task IDs (source-listed plan; execution mapping may differ): elx-07-cli-stats, elx-12-retry-api-deprecation, rails-ac-throttle-search, rails-aj-enqueue-after-commit, rails-ar-bulk-access-grants, rails-hw-scoped-broadcast, rails-sup-legacy-conversions, syn-14-bug-sla-business-hours.
- Historical execution command: not recorded in the cited round source.
- Raw records: the cited sources do not provide a complete public round-specific request and grade bundle; public delivery and outcome exports are linked above where applicable.


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

STATUS: **INVALID**

Data completeness: **44.7%** (33 deliveries; PASS WITH DECLARED GAPS). [Measurement contract](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r36/MEASURED.md); [declared gaps](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r36/MISSING.md).













Status reason: Official cells outcomes do not match one prior captured pass count; the table and result summary are removed.





Pre-registered: no


Lifecycle: Historical capture; closure is not independently reconciled from the public record.

Question: Can a lower-cost ensemble reach the observed pass rates of the Sol-helped pipeline on the selected tasks?
Design: The historical recipe describes a Luna high pack plan, five Luna low builds, Luna high selection and scoped review. The selected-build criterion is not re-derivable from the public record. Studio worker tasks: rails-aj-enqueue-after-commit, rails-sup-legacy-conversions, rails-hw-scoped-broadcast, rails-ar-bulk-access-grants and rails-ac-throttle-search, three reps each; worker tasks: elx-07-cli-stats, elx-12-retry-api-deprecation and syn-14-bug-sla-business-hours, three reps each. The Sol-assisted reproduction-and-review comparison used the same worker tasks.

Venue: Public host identifiers are listed in the reconciliation table. The surviving record does not establish every planned task-to-host assignment or the execution-environment details.

Decision rule: No public predeclared decision rule was recovered; no outcome comparison is reported.

Arms: Exact arm recipes are not recoverable from the public record; the previous label table is removed pending a cell-level crosswalk.

Planned tasks in the surviving round note: elx-07-cli-stats, elx-12-retry-api-deprecation, rails-ac-throttle-search, rails-aj-enqueue-after-commit, rails-ar-bulk-access-grants, rails-hw-scoped-broadcast, rails-sup-legacy-conversions, syn-14-bug-sla-business-hours

Verdict: Outcome interpretation is withheld because the official cells export and captured run-record export cannot be crosswalked for this round. No benchmark comparison is reported.

## Public outcome reconciliation status

The prior delivery and outcome table is removed because its arm labels or outcome totals do not match the official [cells export](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/cells.jsonl). The captured-delivery ledger remains available in [r36.jsonl](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/r36.jsonl); no combined result is reported here.

## Recomputed from raw records

[`recomputed.json`](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/rounds/r36/recomputed.json), [mined cell records](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/data/mined/r36.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 30 rows (fail 10, grader_error 1, pass 19); overall pass rate is 65.5% (19/29) among pass/fail grades. Per-arm detail and coverage counts are in the JSON.

Published-export outcome cross-check: 30/30 shared cell IDs match; 0/30 public-export rows have no mined source record. Exact IDs and results for unpaired rows are in `recomputed.json`.
No per-cell outcome mismatches were found in the shared IDs.
Token reconciliation under [METHOD.md](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/METHOD.md): 30/30 published metadata totals equal the manifest `input + cached input + output` sum; 0 differ (no mismatches). Reasoning remains a separate reported component and is not added to output.
The recompute keeps both numbers visible. 0 unexplained single-response numeric errors were identified; multi-response rows retain the published full-cell total and the manifest counter separately because per-response usage records are excluded.


## Captured run records

[results/run-records/r36.jsonl](/downloads/results/run-records/r36.jsonl) · 33 records · 225,349 bytes · [immutable source](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/run-records/r36.jsonl).


## 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 experiments

- [r36](/rounds/r36/)
