RESEARCH / Focused target
The exact halting time of Wolfram machine 3626
RELEASED FOUNDATIONS / PROPOSED CONNECTION
Research connections
All source anchors (1)
- D5/S0/Computability/BinaryMachine/Machine3626.resultAbsent from this graph
Problem
The question
Peter Bala, OEIS A393210, conjecture of September 11, 2026, revision 18:
Conjecture: a(n) = Sum_{d | n} m(d)^2 * phi(d), where m(n) = Sum_{d | n} (-1)^(d+1) * moebius(d) = A069517(n) and phi(n) denotes Euler's totient function A000010(n). Cf. A296306.
The sequence has offset 1. The equivalent conjectural relation in A069517, revision 31, is the same question, counted once. Here a(n) is the actual number of transitions of the one-sided two-state two-symbol machine 3626 on every positive natural input n in binary. Its half-tape extends infinitely leftward, blank zero beyond the most significant bit. The head starts in state A at the right edge on the least significant bit. Its rules are A0 → A0L, A1 → B1R, B0 → A1L, B1 → B0R. A simultaneous read, write, state change and move counts as one transition. Moving right off the edge terminates the run, and that final transition is counted. Input zero is excluded. There is no time cap, tape bound or counter width. Both divisor sums range over positive divisors; the signed Moebius arithmetic is in the integers.
Motivation
Our foothold
D5/S0/Computability/BinaryMachine/Machine3626.result supplies an actual
first-boundary-crossing trace for every n > 0 whose natural transition
count, cast to the integers, equals the source's signed Moebius/totient
sum. Operational execution is defined independently of the arithmetic
formula. The literal source, published arithmetic prerequisites and
primary machine-number decoder are cited in
Library/Computability/bala2026machine3626.md.
Gap
Missing bridges
The external conjecture identifies an operational transition count with an arithmetic divisor transform. Its exact all-positive-input event is registered in issue 25808, with the unchanged prospective declaration. The formal bridge relates a restoring binary block to the arithmetic transform for all positive inputs.
Route
Proposed approach
A block of k zero bits followed by one and arbitrary higher cells admits a restoring trace. The recursive step composes two traversals of the smaller block and three literal transitions, giving T(0) = 1 and T(k+1) = 2 T(k) + 3, hence T(k) = 4·2^k − 3. Binary recursion finds this first one for every positive input. Independently, the published multiplicative Moebius and totient facts reduce the divisor transform to the same value: the odd factor contributes one and the power of two contributes 4·2^k − 3. Uniqueness of the initial zero block matches the two values, yielding the original signed sum and a genuine operational trace.
Falsifier
What would falsify this route
Source fidelity would fail if the rule-number decoder, initial head,
implicit blanks or counted boundary-crossing convention differed from the
source. Under the stated semantics, a positive input whose actual first
crossing count differs from the displayed divisor sum would contradict
D5/S0/Computability/BinaryMachine/Machine3626.result. Agreement on a
finite set of inputs alone cannot establish this universal claim.
Evidence
Evidence to collect
The formal result is
D5/S0/Computability/BinaryMachine/Machine3626.result.
Its operational prerequisites are
D5/S0/Computability/BinaryMachine/Machine3626.block_first_crossing and
D5/S0/Computability/BinaryMachine/Machine3626.positive_input_first_crossing.
The source-equivalence registrations are in
Reg/D5/S0/Computability/BinaryMachine/Machine3626.lean; the Blueprint
mirrors the source formula, integer casts and operational conventions.
The primary rule decoder is WolframInstitute/TuringMachine commit
d23b7bcb741c711ea7b87a76391da1ac150c3bba, from_number in
TuringMachine/Libs/ndtm_search/src/models.rs, and initialization/counting
are given by OneSidedTuringMachineFunctionNative in
TuringMachine/Kernel/Functions.wl. The implementation's maxSteps
parameter is not a hypothesis of the theorem.
Triage
Scope assessment
theorem; one source-qualified universal question with the aliases
A393210 and A069517. The source author is Peter Bala; the source tier is 1.
What the settlement shows
- Proved (
D5/S0/Computability/BinaryMachine/Machine3626.result): every positive binary input has an actual first-crossing trace whose count equals the original signed Moebius/totient divisor sum. This proves the displayed A393210 conjecture and the equivalent relation in A069517. - Proved (
D5/S0/Computability/BinaryMachine/Machine3626.block_first_crossing): arbitrary higher Boolean cells can replace the canonical higher binary digits once the initial block is k zeros followed by one. The same exact count 4·2^k − 3 applies without assumptions on those higher cells; the restoring trace is the decisive mechanism. - Proved (
D5/S0/Computability/BinaryMachine/Machine3626.positive_input_first_crossing): every positive input has such a block. The count is an equality, including the final crossing transition, rather than an upper bound. - Proved (
D5/S0/Computability/BinaryMachine/Machine3626.result): source arguments using precisely the registered conjectural equality may use this theorem under the same initialization and counting convention. The arithmetic multiplicativity and prime-power prerequisites remain reused published results. The citation to A296306 supplies context; no additional identity for that sequence is asserted here. - Open: input zero, other machines, other boundary or counting conventions, and any further identities or classifications require separate statements and proofs. They are outside this resolution.
ASSUMED-UNVERIFIED
Unverified assumptions
The bounded known-answer screen covered the exact OEIS identifiers, repository issues and pull requests, arXiv, Crossref and a Turing-machine inventory. It found no exact prior answer in the inspected scope; some broader provider searches were unavailable or incomplete. This is not an exhaustive absence or priority claim. The theorem resolves the stated source-qualified event without asserting global novelty.