re: freezes are stochastic, not eliminated; calibration vectorised

Correcting the previous iteration's wording. It claimed on n=1 that disabling
the periodic rescan "removes the freeze". With more runs that is too strong: the
configuration is now clean at 210, 240 and 300 seconds and frozen at 60 on a
fourth run.

The tally across configurations: the heavy probe froze at 27, 45, 83, 183 and
255 seconds; the cheap probe with a 90 s rescan froze at 183; the cheap probe
with no rescan is clean in three runs of four past 200 s. That is a real and
large improvement in survival, but it is a change in probability rather than a
fix, and the documentation now says so.

The witness handled the frozen run correctly -- losses stop at t=45 and stalls
are flagged from t=60 onward, with no contradiction between them.

Separately, the candidate search was the last Python loop over eight million
words and most of the remaining startup cost. Replaced with two numpy vector
operations; 7250 candidates found and startup is no longer the bottleneck. Boot
now dominates: about 190 s of title movie plus 35 s to flight against a 595 s
turn cap leaves roughly 350 s of observation, so boot is the only remaining
lever if longer windows are needed.

Practical rule recorded: do not treat a single frozen run as evidence. Run,
check the witness, discard the frozen ones and keep the clean ones -- about
three in four are usable now, which is workable where it previously was not.

docs/re/BACKLOG.md

@@ -454,9 +454,12 @@ search cannot find a *schedule*.
   90 s, and the new calibration added **two more 32 MB reads** — the 45 s freeze
   came right after it. 🔴 **The trim BROKE the witness** (17 candidates, `0/17` on
   every sample of a run with 13 losses) — reverted; two 32 MB reads once at
-  startup is the price of a working witness. ✅ **Instead, disabling the periodic
-  rescan removed the freeze**: first fully clean probed run, **zero stalls over
-  210 s** with 8 losses, ended by the turn timeout not a freeze (n=1). The rescan
+  startup is the price of a working witness. ✅🟡 **Instead, disabling the periodic
+  rescan LARGELY removes the freeze** — corrected from "removes" after more runs:
+  clean at 210/240/300 s but **frozen at 60 s** on a fourth. Tally — heavy probe
+  froze at 27/45/83/183/255 s; cheap+rescan at 183 s; cheap, no rescan: 3 of 4
+  clean past 200 s. A large probability improvement, **not a fix**. Practical
+  rule: run, check the witness, discard frozen runs (~3 in 4 usable). The rescan
   existed only to catch newly-allocated craft, which the roster work showed never
   happens.
 * ✅ **Run 16: the first TRUSTWORTHY negative.** Validated witness, no stall on

docs/re/guest-stalls.md

@@ -376,3 +376,40 @@ clean no-probe control it points clearly at recurring heavy reads rather than at
 memory reading as such.
 
 The run ended on the turn's timeout, not a freeze.
+
+---
+
+# Freezes are stochastic, not eliminated (2026-08-24)
+
+The previous entry claimed disabling the periodic rescan "removes the freeze" on
+n = 1. With more runs that is too strong. The tally:
+
+| configuration | freeze onset |
+|---|---|
+| heavy probe | 27, 45, 83, 183, 255 s |
+| cheap probe + 90 s rescan | 183 s |
+| **cheap probe, no rescan** | clean 210 s, clean 240 s, clean 300 s, **frozen at 60 s** |
+
+Three of four runs now survive past 200 s where essentially none did before, so
+removing the rescan is a **real and large improvement** — but it is a change in
+probability, not a fix, and the wording is corrected accordingly.
+
+The witness handled the frozen run correctly: losses stop at t = 45 s, stalls are
+flagged from t = 60 s onward, no contradiction.
+
+## ✅ Calibration is no longer the bottleneck
+
+The candidate search was the last Python loop over 8 million words. Replaced with
+two numpy vector operations (`AGENT.md` notes numpy is installed, and its absence
+used to look like a logic bug). 7250 candidates found, and the startup cost is
+now small enough that **boot dominates**: ~190 s of title movie plus ~35 s to
+flight, against a 595 s turn cap, leaves roughly 350 s of observation.
+
+Boot is now the binding constraint on how much game time one turn can see, and
+it is the only remaining lever worth pulling if longer windows are needed.
+
+## Practical consequence
+
+Do not treat a single frozen run as evidence of anything. Run, check the witness,
+discard the frozen ones, and keep the clean runs — about three in four are usable
+now, which is workable where it previously was not.