The A166944 record-index law is equivalent to a canonical return at
ProvedA166944Research.record_index_iff_canonical_returnsLet be the Rowland-type recursion of OEIS A166944 with first differences , and let
be the offset coordinate measuring how far has advanced past the linear benchmark .
The record-index law says that a difference record of value can only be attained at the index :
The canonical return statement says that from a canonical record — a value attained at index exactly , with an attained maximum on the prefix up to and a strictly larger increment at — the sequence must come back down to the offset before that larger increment is taken:
The theorem is the equivalence : the index law for all records holds if and only if no canonical replay overshoots its horizon before returning to . This converts a statement about the global record sequence into a statement about a single finite object — one capped replay from a canonical record — and is therefore the pivot on which the remaining obstruction rests.
The status of the A166944 conjecture is open, and so is ; what is proved here is only the equivalence, which is a genuine advance because is a self-contained statement about one trajectory segment that can be attacked, searched for counterexamples, and decomposed. Neither side is asserted to be true.
Formalization Note B is truncated subtraction on Nat, and IsAttainedPreviousMaximum M k is the conjunction “ for ” and “ for some ”, so the maximum is required to be attained and not merely an upper bound. Both sides of the equivalence use the same binder style, with implicit R, n on the left and implicit M, k on the right.
import Definitions.Def_a166944_canonical open A166944Research
namespace A166944Research
theorem record_index_iff_canonical_returns :
(∀ {R n : Nat}, 5 < R → 2 ≤ n → d n = R →
(∀ j : Nat, 2 ≤ j → j < n → d j < R) → n = R + 2) ↔
(∀ {M k : Nat}, 5 < M → M + 2 ≤ k → d (M + 2) = M →
IsAttainedPreviousMaximum M k → M < d (k + 1) →
∃ m : Nat, M + 2 ≤ m ∧ m < k + 1 ∧ B m = 2) := by sorry
end A166944Research