{
 "tm": "1RB1RA_1LC0RE_1LE0LD_0LB1LB_1RA0LF_1LD---",
 "pass": "funded stage seams (2026-10-10)",
 "confidence": "low",
 "recommend": "park (the funded stop rule fired: the seam alphabet still grows at 1e10); optional: one native closed-language search for the single invariant 'no 00 <B inside the tape' (decision.json)",
 "mechanism": "One bounce per cycle over a sea of u = 01001111. The tape has about 10-15 tokens: the left reservoir u^a, a seam, the run u^b that the wall eats (rule Q: a+1, b-1, c+3), the wall seam 1 (10100)^2, the deposit 1^c, and older segments. Deposits are later rewritten through long (10)-runs (up to (10)^8429) into u-runs of about c/8 blocks; the outcome (u-run in rotation 4, 5 or 7, a (10)-run, or a merge) depends on c mod 8 and the local context. Halting needs C to read 0 inside the tape, which never happens to 1e10: non-halting is equivalent to 'D never steps left onto 0 0 inside the tape'.",
 "rounds": [
  {
   "date": "2026-10-10",
   "option": "seams",
   "confidence": "low",
   "closed_gaps": [
    "Lossless phase-tracked token extractor (scripts/seamscan.c): 33,051 cycle starts to 1e10, verified 3,197/3,197 cycles identical to tools/sim.py to 1e8; 1e10 end state (lo -99,184, hi 1,687, no halt) matches tmsim.",
    "Saturation test done, answer no: distinct seam words 513/783/1,028 at 1e8/1e9/1e10 (verbatim), 449/662/844 (1-runs and (10)-runs abstracted), 454/669/860 (also inner u-pieces), elementary pieces 182/226/270 (longest 17 cells); collision cycles 877/1,797/2,785 and collision types 416/635/842.",
    "Run-length fuzz from real cycle starts: 0/1,200 halts within 5e7 steps (V0 unchanged, V1 residues kept, V2 deposit and (10)-run lengths free, V3 u-run lengths free in cells; 300 each, seed 7).",
    "Near-halt census: C reads 0 33,051 times to 1e10, always at the blank left edge (once per cycle), never inside. B reads 0 inside 66.7M times to 1e9, always with a 1 on its left (50.2M via A0, 16.6M entered moving left from D)."
   ],
   "open": [
    "No finite seam alphabet: every count still grows by a roughly constant amount per decade at 1e10.",
    "No closed language built; the stop rule ended the stage before that step.",
    "Epoch-level map on segment lengths (deposit 1^c -> (10)-run -> u^(c/8), residue branches) not modelled; not tested for Collatz-like behaviour."
   ],
   "numbers": {
    "steps": 10000000000,
    "cycles": 33051,
    "seam_words_A0": {
     "1e8": 513,
     "1e9": 783,
     "1e10": 1028
    },
    "seam_words_A1": {
     "1e8": 449,
     "1e9": 662,
     "1e10": 844
    },
    "elementary_pieces": {
     "1e8": 182,
     "1e9": 226,
     "1e10": 270
    },
    "collision_types": {
     "1e8": 416,
     "1e9": 635,
     "1e10": 842
    },
    "longest_10_run": 8429,
    "fuzz_halts": "0/1200",
    "interior_C0": 0
   },
   "files": [
    "scripts/seamscan.c",
    "scripts/seamtok.py",
    "scripts/seamstats.py",
    "scripts/seamabs.py",
    "scripts/seamelem.py",
    "scripts/collisions.py",
    "scripts/epochs.py",
    "scripts/segmap_tokens.py",
    "scripts/fuzz_lengths.py",
    "scripts/census.c",
    "scripts/census_b.c",
    "work/tokens_1e10.txt",
    "work/seams_growth.txt",
    "work/seams_abs.txt",
    "work/seams_elem.txt",
    "work/collisions.txt",
    "work/fuzz_lengths.out",
    "work/census_1e10.txt",
    "work/census_b_1e9.txt",
    "img/segmap_1e10_right.png"
   ]
  }
 ]
}