run 26

docs/discovery-notes.md

@@ -5554,3 +5554,88 @@ Best current `EC` model:
   - without `A0`, `EC` falls back toward heartbeat-family behavior
 - This is closer to a controlled state map than where we started, but it is
   still not a stable multi-turn "conversation" ladder.
+
+### HE23: Does `A0` Also Context `E8` And `E9`?
+
+The `EC` branch now strongly suggests that leading `A0` is part of branch
+selection context rather than just a neutral primer. These tests ask whether the
+same is true for `E8` and `E9`.
+
+#### HE23a: `E8` Without Leading `A0`
+
+```powershell
+python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 E8 40 30 C2" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-e8-no-a0.txt
+```
+
+#### HE23b: `E9` Without Leading `A0`
+
+```powershell
+python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 E9 40 30 C3" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-e9-no-a0.txt
+```
+
+#### HE23c: `A0` Then `E8` Control Repeat
+
+This is just the known-good comparison run so the no-`A0` tests can be read
+against a fresh control captured in the same phase of work.
+
+```powershell
+python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 E8 40 30 C2" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-e8-with-a0-control.txt
+```
+
+#### HE23d: `A0` Then `E9` Control Repeat
+
+```powershell
+python scripts/serial_sequence_probe.py --port COM5 --prompt --frame "00 00 A0 00 80 7A" --frame "00 00 E9 40 30 C3" --frame "00 00 00 00 80 DA" --frame "00 00 00 00 80 DA" --repeat 2 --frame-interval 0.20 --read-after-frame 0.30 --read-after-group 0.8 --log captures/rcp-heartbeat-e9-with-a0-control.txt
+```
+
+What would count as a hit:
+
+- `E8` or `E9` without `A0` still producing their known branches:
+  - `07 80 7A 50 26 D1`
+  - `07 80 7A 28 D3 5C`
+- `E8` or `E9` without `A0` collapsing into heartbeat-family transients like
+  `07 80 40 40 30 ED` or `07 80 C0 40 30 6D`
+- a clear difference between no-`A0` and with-`A0` controls that would promote
+  `A0` from "possible primer" to "general selector context"
+
+### 2026-05-13 `A0` Context Test For `E8` / `E9`
+
+Captures:
+
+- `captures/rcp-heartbeat-e8-no-a0.txt`
+- `captures/rcp-heartbeat-e9-no-a0.txt`
+- `captures/rcp-heartbeat-e8-with-a0-control.txt`
+- `captures/rcp-heartbeat-e9-with-a0-control.txt`
+
+Observed outcomes:
+
+| Setup | Group-1 result |
+| --- | --- |
+| `E8` without `A0` | later heartbeat provoked `07 80 40 40 30 ED` |
+| `E9` without `A0` | later heartbeat provoked `07 80 40 40 30 ED` |
+| `A0 -> E8` control | `07 80 7A 50 26 D1` |
+| `A0 -> E9` control | heartbeat only in this repeat |
+
+Interpretation:
+
+- `E8` now clearly matches the `EC` pattern in one important way:
+  without leading `A0`, it falls back into heartbeat-family behavior instead of
+  opening its `7A` branch.
+- `E9` without `A0` also fell back to the same heartbeat-family transient
+  `07 80 40 40 30 ED`.
+- So `A0` is now more plausibly a broader selector-context opener, not just an
+  `EC`-specific quirk.
+- However, `A0` is **not** sufficient to make every selector deterministic:
+  the `A0 -> E9` control repeat produced only heartbeat in this run, even though
+  earlier `A0 -> E9` work did produce `07 80 7A 28 D3 5C`.
+
+Best current read:
+
+- `A0` looks necessary or at least strongly supportive for entering the `E8`
+  and `EC` family spaces.
+- `E9` remains more timing- or state-sensitive than `E8`; `A0` helps define the
+  context, but does not guarantee the `7A 28 D3` branch on every clean repeat.
+- This strengthens the state-map model:
+  - no `A0` -> selector values tend to collapse into heartbeat-family
+    transients
+  - with `A0` -> some selector values can open structured family spaces

docs/pt2-protocol-summary.md

@@ -285,6 +285,11 @@ Current caution:
   `A0`, `EC` fell back to `07 80 C0 40 30 6D`.
 - Exact and host-shaped handling of `07 80 FB 50 26 50` did not yield a stable
   next-stage response.
+- Follow-up context tests suggest `A0` matters beyond `EC`:
+  without `A0`, both `E8` and `E9` fell back to `07 80 40 40 30 ED` instead of
+  opening their known `7A` branches.
+- At the same time, `A0` is not sufficient to make every branch deterministic:
+  a later `A0 -> E9` control repeat returned heartbeat only.
 
 ## What We Know
 
@@ -322,8 +327,8 @@ Current caution:
 - Overall, the evidence is getting stronger for selector-like host entries that
   open related response families, but weaker for a fully reproducible multi-turn
   "conversation" state.
-- At least on the `EC` branch, `A0` now looks more like part of the selection
-  context than a purely generic primer.
+- `A0` now looks more like part of the selection context than a purely generic
+  primer, but its effect is branch-dependent and not fully deterministic.
 
 ## What We Do Not Know
 

docs/pt2-state-map.md

@@ -126,15 +126,25 @@ different downstream response families.
 ### S4: `A0`-Contexted Selector Space
 
 Current best evidence says `A0` is not always just a neutral primer. On at
-least the `EC` branch, it appears to be part of the context that opens the
-family space.
+least the `EC` and `E8` branches, it appears to be part of the context that
+opens the family space.
 
 Observed:
 
+- `A0 + E8` can produce `7A 50 26` family output
 - `A0 + EC` tends to produce `7B`-family output
+- `E8` without `A0` fell back to `07 80 40 40 30 ED`
+- `E9` without `A0` also fell back to `07 80 40 40 30 ED`
 - `EC` without leading `A0` falls back to `07 80 C0 40 30 6D`
 
-Confidence: medium-high
+Nuance:
+
+- `A0` is not sufficient to make every selector deterministic.
+- A later control repeat of `A0 + E9` produced heartbeat only, even though
+  earlier `A0 + E9` runs produced `07 80 7A 28 D3 5C`.
+
+Confidence: high that `A0` is contextual; medium on the exact scope of that
+context.
 
 ## `Ex` Selector Neighborhood
 
@@ -198,12 +208,16 @@ What we know:
 
 - reproducible in group 1
 - exact and host-shaped replies to `7A 28 D3 5C` did not advance state
+- without `A0`, `E9` fell back to `07 80 40 40 30 ED`
+- with `A0`, a later control repeat returned only heartbeat, so this branch is
+  less deterministic than `E8`
 
 Read:
 
 - `E9` looks like a sibling selector to `E8`
 - payload differences may represent page/class selection rather than random
   drift
+- `E9` appears more timing- or state-sensitive than `E8`
 
 Confidence: medium