Skip to content

research(proximity): add G139 accident census probe - #526

Open
coopertweak wants to merge 1 commit into
lalalune:research/proximity-prizefrom
coopertweak:codex/g139-production-accident-census
Open

research(proximity): add G139 accident census probe#526
coopertweak wants to merge 1 commit into
lalalune:research/proximity-prizefrom
coopertweak:codex/g139-production-accident-census

Conversation

@coopertweak

Copy link
Copy Markdown

Summary

  • add a stdlib-only exact probe for the G136/G139 normalized rung-2 accident census
  • reproduce the G173 nonzero accident witness with the same predicate
  • check production-like first-prime cells for n = 16, 32, 64, 128, 256

Result

The G173 cross-check sees the known nonzero accident cell:

G173 cross-check: n=64 p=17318209 total_solutions=213 accidents=24 witness=(5663213, 17079628, 5424631)

At the production-like first-prime cells near p ~= n^(158/30), the count is zero in every checked row:

n,p,log_p_base_n,v2(p-1),total,lawful,accidents,4|accidents
16,2196401,5.266677511,4,45,45,0,True
32,84551969,5.266667003,5,93,93,0,True
64,3254976577,5.266666669,6,189,189,0,True
128,125306187649,5.266666672,7,381,381,0,True
256,4823887276289,5.266666667,8,765,765,0,True

The script also includes an independent total-solution count by pair sums and checks that subtracting the lawful count 3n-3 gives the same accident count.

Scope

Finite exact-arithmetic evidence only. This is not a proof at production n = 2^30, not an eventual no-accident theorem, and not a prize closure.

Verification

  • python3 scripts/probes/probe_g139_production_accident_census.py
  • python3 -m py_compile scripts/probes/probe_g139_production_accident_census.py
  • git diff --check
  • python3 scripts/forbidden_tokens.py

@coopertweak
coopertweak force-pushed the codex/g139-production-accident-census branch from 8c17bdc to d7d973a Compare August 3, 2026 03:29
@shane9coy

Copy link
Copy Markdown

Independent read: G139 normalized rung-2 accident census probe.
Stdlib-only, exact. Reproduces the G173 nonzero accident witness
(n=64, p=17318209, total=213, accidents=24, witness=(5663213,
17079628, 5424631)) — the same predicate the G173 family uses.
Then sweeps production-like first-prime cells near p ≈ n^(158/30)
for n ∈ {16,32,64,128,256} and reports zero accidents in every
checked row (n=16: 45/45, n=32: 93/93, n=64: 189/189, n=128: 381/381,
n=256: 765/765). This is a meaningful contrast with the G173
nonzero cell — the small-cell accident does NOT survive to the
production-scale first primes. Clean refutation-as-result: the
G173 nonzero accident is a small-field artifact in this lane.
The script also includes an independent total-solution count by
pair sums and checks that subtracting the lawful count 3n-3 gives
the same accident count. Honest-scope: finite exact-arithmetic
evidence only, not a proof at production n=2^30, not a prize
closure. Concur.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants