re: widen the OB scan to seven encodings; run inconclusive, probe lacked a witness

ob_by_hud.py now scans seven readings of the same bytes and keeps a separate
candidate set for each, as byte offsets: u32 big and little endian, u16 both
endiannesses at both alignments, and u8. The big-endian u32 reading had been
refuted, so widening rather than assuming is the point.

u32le is much the tightest at 154 candidates against u32be's 4452. That is a
hint about the encoding rather than a result, since a rarer bit pattern narrows
faster regardless of meaning.

The run is inconclusive. The HUD read 4 at every sample, so there was no second
value to collapse the sets against, and from t=136 the candidate counts are
byte-identical across five samples in all seven encodings, which is what a
frozen guest looks like -- nothing in 32 MB changed at all. The probe had no
stall witness, so the run cannot prove it either way. One is added now.

Worth stating plainly: this is the fourth probe written without a witness and
the third whose flat output could not be distinguished from a freeze. Each time
the fix gets applied to that one script. The durable fix is the shared probe
harness already noted in this file, and the lesson recurring four times is
itself the argument for building it.

What the hunt needs is unchanged: two HUD readings at different values in
non-stalled samples. The counter moves on kills, which lands back on the combat
limit, though the earlier 4 to 11 observation shows it does move.

docs/re/BACKLOG.md

@@ -620,6 +620,16 @@ search cannot find a *schedule*.
   (template strip covers only **0 1 2 4 8**; most samples read `00?`/`???`), or a
   counter that can rise. **Next: widen the scan to u16/u8 and LE, and beyond the
   heap** — one function, no combat cost. Also extend `ob_digits.png`.
+* 🔴 **(2026-08-24) Widened to 7 encodings; run inconclusive.** `ob_by_hud.py`
+  now keeps a candidate set per encoding (`u32be/le`, `u16be/le` at both
+  alignments, `u8`) as byte offsets. **`u32le` is tightest at 154** vs u32be's
+  4452 — a hint, not a result. 🔴 HUD read **4 at every sample**, so nothing
+  collapsed; and counts were **byte-identical across five samples in all seven
+  encodings** from t=136 s, which is what a freeze looks like. The probe had **no
+  witness** — now added. ⚠️ **Fourth probe written without one, third flat run
+  that cannot be told from a freeze**; the recurring fix is the shared harness
+  noted earlier. Needs two HUD readings at *different* values in non-stalled
+  samples.
 * ~~🚧 BLOCKER: t=210/240 unreachable in one turn~~ — **superseded, see above**;
   it rested on an untested assumption that a turn is one shell call. 595 s shell cap − ~220 s boot (a ~190 s title movie that cannot be
   tapped through) − ~25 s startup = **~350 s observation ≈ 193 game-seconds**.

docs/re/remaining-ob-hunt.md

@@ -215,3 +215,46 @@ kill-driven route it costs no combat.
 
 Also worth extending `ob_digits.png` with the missing digits — most samples were
 unreadable, which is why only two data points survived a 480 s run.
+
+---
+
+# Widened to every plausible encoding (2026-08-24)
+
+`ob_by_hud.py` now scans seven readings of the same bytes and keeps a separate
+candidate set for each, expressed as byte offsets:
+
+`u32be`, `u32le`, `u16be`, `u16be@1`, `u16le`, `u16le@1`, `u8`.
+
+First run, HUD showing **4** throughout:
+
+```
+t=  7s   u32be:6285  u32le:414  u16be:10668  u16be@1:2718  u16le:656  u16le@1:9012  u8:25559
+t=110s   u32be:4472  u32le:156  u16be: 8794  u16be@1:1386  u16le:286  u16le@1:8124  u8:20381
+t=136s   u32be:4452  u32le:154  u16be: 8773  u16be@1:1380  u16le:284  u16le@1:8119  u8:20346
+t=162s .. t=267s   identical to t=136 in every encoding
+```
+
+**`u32le` is much the tightest at 154**, an order of magnitude below `u32be`.
+That is a hint about the encoding, not a result — a rarer bit pattern narrows
+faster regardless of meaning.
+
+## 🔴 Inconclusive, and probably a frozen guest
+
+The HUD read **4** at every sample, so there was no second value to collapse the
+sets against. Worse, from t = 136 s the candidate counts are **byte-identical
+across five samples in all seven encodings**, which is what a frozen guest looks
+like — nothing in 32 MB changed at all.
+
+`ob_by_hud.py` had **no stall witness**, so the run cannot prove it. Added now.
+
+**This is the fourth probe written without a witness and the third whose flat
+output could not be told from a freeze.** Each time the fix is applied to that
+one script. The durable fix is the shared harness already noted in this file, and
+the fact that the lesson keeps recurring is itself the argument for building it.
+
+## What is needed
+
+Two HUD readings at **different** values, in non-stalled samples. The counter
+changes on kills, so this lands back on the combat limit — unless a phase change
+or another event moves it. The earlier 4 → 11 observation shows it does move,
+which is what makes the approach worth continuing.

tools/re-capture/ob_by_hud.py

@@ -24,7 +24,30 @@ def region(fd):
     out, pos = bytearray(), lo
     while pos < hi:
         n = min(1 << 24, hi - pos); out += os.pread(fd, n, pos); pos += n
-    return np.frombuffer(bytes(out), dtype='>u4'), lo
+    return bytes(out), lo
+
+# The big-endian u32 reading was refuted: no word held 4 then 11. Widen rather
+# than assume -- the counter may be narrower, little-endian, or unaligned. Each
+# encoding keeps its own candidate set, expressed as BYTE OFFSETS so the answer
+# is directly usable whichever one wins.
+ENCODINGS = [
+    ('u32be', '>u4', 4, 0), ('u32le', '<u4', 4, 0),
+    ('u16be', '>u2', 2, 0), ('u16be@1', '>u2', 2, 1),
+    ('u16le', '<u2', 2, 0), ('u16le@1', '<u2', 2, 1),
+    ('u8',    'u1',  1, 0),
+]
+
+def matches(buf, value):
+    """byte offsets whose value equals `value`, per encoding"""
+    out = {}
+    for name, dt, w, skew in ENCODINGS:
+        if value > (1 << (8 * w)) - 1:
+            continue
+        n = (len(buf) - skew) // w * w
+        a = np.frombuffer(buf[skew:skew + n], dtype=dt)
+        idx = np.nonzero(a == value)[0]
+        out[name] = idx.astype(np.int64) * w + skew
+    return out
 
 def hud_value():
     subprocess.run(['screenshot', SHOT], capture_output=True, timeout=60)
@@ -40,6 +63,19 @@ def main():
     secs  = int(sys.argv[1]) if len(sys.argv) > 1 else 240
     every = int(sys.argv[2]) if len(sys.argv) > 2 else 25
     w = gworld.World(); fd = w.fd
+    # Witness. This is the FOURTH probe written without one, and the third to
+    # produce a run whose flat output could not be distinguished from a freeze.
+    # The recurring fix is a shared harness; until that exists, carry it.
+    b0, _ = region(fd); time.sleep(3.0); b1, _ = region(fd)
+    a0 = np.frombuffer(b0, dtype='>u4').astype(np.int64)
+    a1 = np.frombuffer(b1, dtype='>u4').astype(np.int64)
+    d = a1 - a0
+    rates = collections.Counter((d[(d > 15) & (d < 600)] // 3).tolist())
+    band = [r for r in rates if 8 <= r <= 40]
+    pick = max(band, key=lambda r: rates[r]) if band else None
+    ticks = np.nonzero(d // 3 == pick)[0][:32] if pick is not None else np.array([], dtype=int)
+    print('tick witnesses: %d at %s/s' % (len(ticks), pick), flush=True)
+    last_t = a1[ticks] if len(ticks) else None
     cand = None; base = None; seen = []
     t0 = time.time()
     while time.time() - t0 < secs:
@@ -48,23 +84,39 @@ def main():
         if v is None:
             print('  t=%4ds  HUD unreadable (%s)' % (el, note), flush=True)
         else:
-            r, base = region(fd)
+            buf, base = region(fd)
-            hit = set(np.nonzero(r == v)[0].tolist())
+            hit = matches(buf, v)
-            cand = hit if cand is None else (cand & hit)
+            if cand is None:
+                cand = hit
+            else:
+                cand = {k: np.intersect1d(cand[k], hit[k], assume_unique=True)
+                        for k in cand if k in hit}
             seen.append(v)
-            print('  t=%4ds  HUD=%-4d words==%d: %-8d -> candidates %d  (%s)'
+            st = ''
-                  % (el, v, v, len(hit), len(cand), note), flush=True)
+            if len(ticks):
-            if len(cand) <= 40 and len(set(seen)) >= 2:
+                now_t = np.frombuffer(buf, dtype='>u4').astype(np.int64)[ticks]
+                if int((now_t > last_t).sum()) == 0: st = '  *** GUEST STALLED ***'
+                last_t = now_t
+            print('  t=%4ds  HUD=%-4d  %s%s  (%s)'
+                  % (el, v, '  '.join('%s:%d' % (k, len(cand[k])) for k in cand), st, note),
+                  flush=True)
+            live = {k: c for k, c in cand.items() if len(c)}
+            if live and len(set(seen)) >= 2 and min(len(c) for c in live.values()) <= 40:
                 break
         time.sleep(every)
     print('\nHUD values seen: %s' % sorted(set(seen)))
     if cand and base is not None:
-        print('candidates: %d' % len(cand))
+        for k in sorted(cand, key=lambda k: len(cand[k])):
-        for i in sorted(cand)[:20]:
+            c = cand[k]
-            va = gmem.primary_va(base + i * 4)
+            print('  %-8s %d candidate(s)' % (k, len(c)))
-            print('   va %s' % (('%#010x' % va) if va else '?'))
+            for off in c[:8]:
+                va = gmem.primary_va(base + int(off))
+                print('      va %s' % (('%#010x' % va) if va else '?'))
+        if all(len(c) == 0 for c in cand.values()):
+            print('\nEVERY encoding eliminated -- the counter is not in this region '
+                  'in any of them, or a HUD reading is wrong.')
     else:
-        print('no candidates (HUD never read, or value never matched a u32)')
+        print('no candidates (HUD never read)')
     return 0
 
 if __name__ == '__main__':