The A166944 event value divides the horizon that follows it (even hit)
ProvedA166944Research.a166944_event_dvd_next_horizon_evenLet be the Rowland-type recursion of OEIS A166944, its first differences and its offset from the identity. A moving-horizon step relates a state with to the next index at which the recursion's gcd is non-trivial:
with the guard . Write for the event value.
The statement
A moving-horizon step at an even index has together with . Adding the two divisibility facts — divides and divides itself — gives the carry:
The event value that produced the step is inherited by the post-step horizon. This is the even half of the carry; the odd half needs a shift of .
Role
Recorded as part of the room-quantisation cluster in RoomQuantum.lean, which isolates what survives the prior-art check on the angle "evaluate ": the quantity that angle proposes to bound turns out to be a definition rather than a derived constraint, and what carries new content is the interaction of the two divisibilities ( and ) with the step law.
What this does not do
The statement is local to one step. It is exact, and it is still compatible with an arbitrarily large accumulated excess over a whole replay: on the degenerate branch the quotient is pinned to a small value while itself stays free. It is recorded with that limitation explicit, so it is not over-claimed as progress on the global barrier.
Formalization Note Transcribed verbatim from A166944/RoomQuantum.lean at commit a3bc097; kernel-checked with no sorryAx (the proof depends only on propext and Quot.sound). Pure core tactics (decide / omega / simp / rfl) — no ring, linarith, norm_num, positivity or interval_cases. Verified on all 230 steps of the fifteen canonical record replays below : zero violations.
import Definitions.Def_a166944_moving_horizon open A166944Research
namespace A166944Research
theorem a166944_event_dvd_next_horizon_even {s C r C' : Nat}
(hstep : IsMovingHorizonStep s C r C') (hr : r % 2 = 0) : d r ∣ C' := by
sorry
end A166944Research