Round x-race

x race

Status · DESCRIPTIVEReviewed status is transcribed from rounds/STATUS.md. The original narrative and its qualifications remain below.

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

Round source README · rounds/x-race/README.md

Round date: HISTORICAL (before 2026-10-09) Publication badge: DESCRIPTIVE; KEPT FOR AUDIT Why not VALID: The historical inventory retains DESCRIPTIVE because of ARITHMETIC_UNRECOMPUTABLE, 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.
  • ARITHMETIC_UNRECOMPUTABLE: The reported 48-cell pass total is not reproducible from retained records.
  • RAW_EVIDENCE_MISSING: Exact-ID raw records are absent.

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.

Terminology: public round glossary.

STATUS: INTERIM

Data completeness: N/A (no exact-ID per-cell records in the public exports). Measurement record; declared gaps.

Pre-registered: not established in the public record; treat this as retrospective exploratory evidence. Hypothesis coverage: H29, H30 (provisional mapping; see the hypothesis register). Question: Do racing multiple workers improve selected task outcomes over simpler arms when a deterministic selector chooses a candidate? Population: Six tasks × two repetitions × four arms = 48 reported cells. Headline: No superiority or deployment claim; the pass total is source-reported and not reproducible.

Reported observations

  • The four-arm panel is reported as 48/48 graded passes, a ceiling result on this task set. (reported data; source-reported, not reproducible from public data.)
  • A selector false-acceptance check is reported as 0/24. (reported data; source-reported, not reproducible from public data.)
  • No two-agent policy was promoted. (reported data; source-reported, not reproducible from public data.)

Reproduction and records

Required field Publicly recoverable value
Harness/Kogen revision Exact harness/Kogen revision was not recovered.
Model and effort Per-arm model and effort values were not recovered in the public snapshot.
Task IDs Six tasks are reported; exact public task IDs are not recovered.
Command Not recovered; public launch command and cell IDs are unavailable.
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.

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: A ceiling panel cannot estimate improvement in success rate, and the selector check does not qualify a live multi-agent route.

Smallest useful next test: Use tasks with measured headroom, compare solo and race under a frozen selection rule, and report wall time and token cost by task.

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; public cells export; public captured-delivery export.

Recomputed from raw records

`recomputed.json`, mined cell records, and the token audit back observed official outcomes, per-arm pass counts/rates, manifest token counters, and wall medians. The source contains 53 rows (pass 50, restore_failed 2, ungraded 1); overall pass rate is 100.0% (50/50) 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. Recovered-source discrepancies: 24 field mismatches across 1 cell IDs (experiment 3, patch.archive_member.5d61af3600d5819e6afd65f6c8cad5822cdea2b1b86fddbef3b85a71f6abe5ba 1, patch.archive_member.ce248f8131723e78bfca8d78f032869d5706e93bb474aa0aa3239e4df52d3059 1, patch.archive_member.f2aa98beee3bceb8e81c28be19b824754fa7ddf26b28d55d076739206937dd32 1, patch.attempt-1 3, usage.cached_input_tokens 3, usage.input_tokens 3, usage.output_tokens 3, usage.reasoning_tokens 3, wall_s 3). Existing values were preserved; each cell, field, kept value, and recovered value is listed below.

  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field experiment: kept "20260930T141459.099867Z-61322-f9e406cf", recovered "20260930T141459.206999Z-61332-a1b32472".
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field experiment: kept "20260930T141459.099867Z-61322-f9e406cf", recovered "20260930T160600.797428Z-8421-27c94025".
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field experiment: kept "20260930T141459.099867Z-61322-f9e406cf", recovered "20260930T160600.906404Z-8424-4b907a91".
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field patch.archive_member.5d61af3600d5819e6afd65f6c8cad5822cdea2b1b86fddbef3b85a71f6abe5ba: kept {"archive_member": null, "sha256": "5d61af3600d5819e6afd65f6c8cad5822cdea2b1b86fddbef3b85a71f6abe5ba", "size_bytes": 3303}, recovered null.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field patch.archive_member.ce248f8131723e78bfca8d78f032869d5706e93bb474aa0aa3239e4df52d3059: kept {"archive_member": null, "sha256": "ce248f8131723e78bfca8d78f032869d5706e93bb474aa0aa3239e4df52d3059", "size_bytes": 4714}, recovered null.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field patch.archive_member.f2aa98beee3bceb8e81c28be19b824754fa7ddf26b28d55d076739206937dd32: kept {"archive_member": null, "sha256": "f2aa98beee3bceb8e81c28be19b824754fa7ddf26b28d55d076739206937dd32", "size_bytes": 4813}, recovered null.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field patch.attempt-1: kept {"sha256": "af7fbae8d5d97d47794f98fcca0706450b10614ac0fb407e5b15fd2684d9d4fc", "size_bytes": 5506}, recovered {"sha256": "5d61af3600d5819e6afd65f6c8cad5822cdea2b1b86fddbef3b85a71f6abe5ba", "size_bytes": 3303}.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field patch.attempt-1: kept {"sha256": "af7fbae8d5d97d47794f98fcca0706450b10614ac0fb407e5b15fd2684d9d4fc", "size_bytes": 5506}, recovered {"sha256": "ce248f8131723e78bfca8d78f032869d5706e93bb474aa0aa3239e4df52d3059", "size_bytes": 4714}.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field patch.attempt-1: kept {"sha256": "af7fbae8d5d97d47794f98fcca0706450b10614ac0fb407e5b15fd2684d9d4fc", "size_bytes": 5506}, recovered {"sha256": "f2aa98beee3bceb8e81c28be19b824754fa7ddf26b28d55d076739206937dd32", "size_bytes": 4813}.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.cached_input_tokens: kept 5120, recovered 24576.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.cached_input_tokens: kept 5120, recovered 61312.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.cached_input_tokens: kept 5120, recovered 89088.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.input_tokens: kept 22131, recovered 18483.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.input_tokens: kept 22131, recovered 34962.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.input_tokens: kept 22131, recovered 35200.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.output_tokens: kept 1849, recovered 1568.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.output_tokens: kept 1849, recovered 3261.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.output_tokens: kept 1849, recovered 3702.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.reasoning_tokens: kept 178, recovered 337.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.reasoning_tokens: kept 178, recovered 586.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field usage.reasoning_tokens: kept 178, recovered 75.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field wall_s: kept 124.299, recovered 107.748.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field wall_s: kept 124.299, recovered 125.358.
  • kh-gpt__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1 field wall_s: kept 124.299, recovered 131.456. 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; 53 missing published metadata). Missing counters remain unknown, not zero. This round also has 53 raw-only or non-public rows; they are kept separate from the published-export denominator.

Captured run records

Standard-schema records assigned to this round in the indexed snapshot.

No per-round file is listed. Check the run-record index for the complete partition list.

Reproduction inventory

Published analysis replay and full experiment replay are reported separately.

Replay type
unavailable
Script
No per-round analysis script is published.
Inputs
No identified public input bundle.
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.

View the complete per-round reproduction inventory · Repository reproduction guide