{"id":6957,"job_id":13718,"problem_id":6,"lane_id":33,"type":"explore","user_id":90,"model":"muse-spark","provider":"unknown","report_md":"# Self-match: 10-billion vary-M7 cell matches the uniform model, best 8; 8-count exact, 7-count mildly high\n\nJob 13718, explore, self-match. Improvement claim: none. No longer matching\nprefix was found, and none is claimed. Nothing goes to the results page.\n\n## What the accepted scoped answers already cover\n\n- The vary-M7 tally searcher is the fastest local self-match method: about\n  47 M/s solo, about 29-42 M/s under a parallel job. It varies only the\n  last four candidate characters, which form message word m7, so steps 0-6\n  are precomputed per batch.\n- Recent 20-billion cells: tally7 73 and 75 against about 74.5 expected,\n  tally8 3 and 1 against about 4.66, tally9 0 against about 0.29. A prior\n  short cell read tally7 14 against about 7.45, then a follow-up cell\n  regressed to 10, showing short-cell 7-counts fluctuate.\n- The 100-billion tally cell is still unfinished, per the prior explore job.\n\n## Uncovered obligation: is the 8-count systematically low?\n\nTwo recent 20-billion cells gave tally8 3 and 1 against 4.66 expected each.\nSmall numbers cannot settle a systematic deficit from two cells, and the\n7-counts in those same cells matched expectation, so a word-specific leak\nis not indicated either. One more independent counted cell is the cheapest\nway to add an 8-count sample.\n\n## Experiment\n\nA fresh disjoint 10-billion-candidate vary-M7 cell from nonce\n13771500000000, the next disjoint start after the prior cycle, on one\ncore, log interval 120 s, with the tally searcher rebuilt from source and\nits selftest passing (fixture score 12). Tallies are first-word scores:\ntally7 counts first-word score 7 or above, tally8 counts first-word score\nexactly 8, tally9 counts full score 9 or above among score-8 survivors.\n\n- [120s] 40.52 M/s, best 8, tally7 21, tally8 1, tally9 0\n- [240s] 41.08 M/s, best 8, tally7 45, tally8 1, tally9 0\n- [done] 41.09 M/s, total 10,000,000,000, best 8, tally7 47, tally8 2, tally9 0\n\nUniform expectations for 10 billion are about 37.3 at 7 or above, 2.33 at\n8, and 0.15 at 9 or above. All nine new-best candidates in the log were\nre-verified with hashlib by parsing the log programmatically, all matching\nexactly, including the best-8 candidate\nc5a2f3323e5713497ffa95cc3d9f90e0 with digest\nc5a2f332592e3435b5ea98e713c7b80d.\n\n## Result\n\nTally8 matches expectation almost exactly (2 vs 2.33) and tally9 is empty\nas expected (0 vs 0.15). Tally7 at 47 vs 37.3 is about 1.6 standard\ndeviations high, a mild fluctuation of the kind short cells already showed,\nnot a leak: a word-specific leak would move the 8+ tail, which matches.\nAcross the three most recent 10-20-billion cells the 8-counts sum to 6\nagainst 11.65 expected, still low-side but small-number variation, not an\nestablished deficit.\n\n## Cheapest new experiment\n\nContinue disjoint tally cells toward the unfinished 100-billion total;\nnext disjoint start nonce 13781500000000. A further 20-billion cell adds\nan independent 8-count with about 4.66 expected, which is the sample that\ndecides whether the low-side aggregate persists, and it costs no new\nsearcher work.\n","patch":null,"cpu_hours":0,"hashes":{},"author_rung":null,"status":"recorded","final_rung":"recorded","created_at":"2026-11-05T05:37:36.800Z","repo_url":null,"commit":null,"cites":null,"tokens":{"log":"summary","input":0,"models":{},"output":0,"source":"none","entries":0,"cache_read":0,"cache_write":0,"observed_models":[]},"paper_slug":null,"revision_path":null,"revision_sha":null,"recipe_md":"# Recipe\n\nTo repeat, on one core:\n\n1. Rebuild the tally searcher from source and run its selftest\n   (fixture score 12).\n2. Run a counted vary-M7 cell from the next disjoint start nonce,\n   log interval 120 s, and record the [done] tallies.\n3. Parse the run log programmatically and re-verify every new-best\n   candidate with hashlib.\n4. Compare tally7, tally8 and tally9 against N/16^7, N/16^8 and N/16^9,\n   and report the next disjoint start nonce for the following cell.","verification":null,"target":null,"finding":null,"human_md":null,"provisional":false,"effects_applied_at":null,"effort":"high","also_fix":null,"transcript_omitted":{"share":0,"omitted":0,"outputs":0},"patch_hash":null,"superseded_by":null,"duplicate_of":null,"transcript_resubmitted_at":null,"file_notes":null,"research":null,"research_route_id":null,"verification_plan":null,"verification_fingerprint":null,"review_admitted_at":null,"department_id":null,"run_id":null,"triage_lead":null,"revision_base_sha":null,"integration":null,"resolves":null,"paper_exposition":null,"research_evidence":null,"transcript_mode":"summary","known_work":null,"work_disposition":null,"applicability_assessment":null,"negative_result":null,"first_indexed_at":null,"index_held_at":null,"index_ok":null,"index_reason":null,"handle":"sundayweaver","job_brief":"Identify an uncovered obligation or a changed premise on this track; compare the accepted scoped answers before proposing the cheapest new experiment. Deliberate replication needs a stated independence objective.","review_deferred":false,"in_triage":false,"triage":[],"lean_statement_binding":null,"lean_execution_binding":null,"lean_scientific_identity":null,"lean_execution_identity":null,"verification_runs":[],"verification_state":null,"verification_summary":null,"canonical_return":null,"review_history":[],"dependencies":[],"cited_by":[],"route_dependents":[],"research_url":null,"transcript_url":"/projects/md5/return/6957/transcript","files":[],"decided_by_author_handle":false,"reviews":[],"decisions":[],"decision":null,"report_sha256":"3398928b3de27b3b8bea053936bc9772e88a360e74cf39cce7bb336a933b6a01","next_step_sha256":null,"research_authority":{"witness_status":null,"research_status":"recorded","scopes":[]},"research_links":[],"duplicates":[],"cited_messages":[]}