@@ -26,7 +26,7 @@ The protocol uses no words or alphabet, but four meaning-free local symbols. For
 See [`docs/results-ecp6.md`](docs/results-ecp6.md) for the conclusion, [`docs/protocol-specification-ecp6.md`](docs/protocol-specification-ecp6.md) for the wire format and [`evidence/ecp6/report.md`](evidence/ecp6/report.md) for the compact confirmatory evidence.
 
 ECP-7 development now tests how much of that result depends on the explicit
-one-factor-per-slot architecture. Its first twenty-four sealed batches are valid
+one-factor-per-slot architecture. Its first twenty-five sealed batches are valid
 development results but none passes the full gate. Batch 15 established the
 strong position-aware base: at 30,000 optimization steps it reaches 83.46% train
 exactness, 82.59% validation and 83.37% translator validation while using
@@ -51,7 +51,11 @@ sender-only warmup. Validation remained 83.50%; total fragmented errors fell,
 but the shared-error pool immediately regrew and injectivity still failed.
 Batch 24 made the second phase receiver-only. It established new-best means of
 84.69% train, 84.48% validation and 84.55% translator validation, but the
-worst-link and injectivity gates still failed. See
+worst-link and injectivity gates still failed.
+Batch 25 extended that catch-up to 45,000 steps and improved again to 85.26%
+train, 85.24% validation and 85.13% translator validation. Receiver performance
+now nearly reaches the mathematical ceiling of the non-injective sender
+codebooks, so another horizon-only extension is not justified. See
 [`docs/development-log-ecp7.md`](docs/development-log-ecp7.md).
 
 ## Build on this work
@@ -116,6 +120,7 @@ Do not translate those artifacts in place. New experiment configurations and all
 - [`config/ecp7-b22-development.yaml`](config/ecp7-b22-development.yaml) — Batch 21 replay with receiver-parameter gradients disabled only on the replay branch.
 - [`config/ecp7-b23-development.yaml`](config/ecp7-b23-development.yaml) — Batch 22 sender-only replay through step 20,000 followed by restored joint replay gradients.
 - [`config/ecp7-b24-development.yaml`](config/ecp7-b24-development.yaml) — Batch 23 routing with sender messages detached for receiver-only replay after step 20,000.
+- [`config/ecp7-b25-development.yaml`](config/ecp7-b25-development.yaml) — Batch 24 receiver-only catch-up with an isolated 45,000-step maximum horizon.
 - [`config/ecp7-positive-control-development.yaml`](config/ecp7-positive-control-development.yaml) — ECP-6 positive control on the ECP-7 split.
 - [`config/ecp0.yaml`](config/ecp0.yaml) — machine-readable configuration from step 1.
 - [`config/ecp1.yaml`](config/ecp1.yaml) — frozen configuration of the population trial.
 
@@ -0,0 +1,12 @@
+extends: ecp7-b24-development.yaml
+
+experiment:
+  title: ECP-7 extended receiver-only catch-up horizon
+  arm: extended_receiver_only_catchup_horizon
+  objective: >-
+    Test whether the late receiver-only catch-up phase continues improving
+    reconstruction or reaches a non-injective plateau when given enough time
+    for the registered early-stopping patience to operate.
+
+training:
+  max_steps: 45000
@@ -41,7 +41,7 @@ python3.12 -m venv .venv
 .venv/bin/ecp6 --config config/ecp6.yaml validate
 ```
 
-Expected baseline: 121 passing tests and split sizes `14336/1024/1024`.
+Expected baseline: 122 passing tests and split sizes `14336/1024/1024`.
 
 ## 4. Reproduce ECP-6
 
@@ -71,7 +71,7 @@ This reruns a known experiment. It must not be presented as a new independent co
 
 ## 5. Continue ECP-7 scientifically
 
-Do not modify ECP-0 through ECP-6 or the completed ECP-7 Batch 1 through Batch 24
+Do not modify ECP-0 through ECP-6 or the completed ECP-7 Batch 1 through Batch 25
 configs. ECP7-B1-I collapsed to 130–139 hard messages. ECP7-B2-I improved its
 soft objective but collapsed further to 85–104 hard messages. ECP7-B3-I applied
 that loss to straight-through hard messages and improved to 585–972 messages,
@@ -144,11 +144,16 @@ ECP7-B24-I made the post-warmup replay phase receiver-only. It set new ECP-7
 bests of 84.69% mean train, 84.48% mean validation and 84.55% translator
 validation, but worst-link validation remained 82.23%, sender injectivity
 failed and the confirmatory ECP-7 test remains sealed.
+ECP7-B25-I extended only that catch-up horizon to 45,000 steps. It improved to
+85.26% mean train, 85.24% mean validation, 82.71% worst-link validation and
+85.13% translator validation. The worst train link exactly matches its 82.81%
+unique-code ceiling, so sender injectivity remains the structural blocker and
+the confirmatory ECP-7 test remains sealed.
 
 Continue with this sequence:
 
 1. Read `docs/research-design-ecp7.md` and `docs/development-log-ecp7.md`.
-2. Define exactly one ECP7-B25 intervention and its failure criterion.
+2. Define exactly one ECP7-B26 intervention and its failure criterion.
 3. Register the variant and immutable configuration hashes before training.
 4. Add tests for every new invariant.
 5. Use only `smoke` and `develop`; keep the ECP-7 test split sealed.
@@ -161,26 +166,28 @@ Continue with this sequence:
 
 ## 6. Recommended next experiment
 
-Batch 24 confirms that receiver-only replay catch-up improves all mean metrics
-but retains cross-link fragmentation and non-injective sender codes. Its best
-checkpoint occurs at step 29,400, only 600 steps before the fixed horizon, so
-the registered patience could not test a late plateau. A clean ECP7-B25
-progression is:
-
-- inherit the complete Batch 24 implementation and configuration;
-- change only `training.max_steps` from `30,000` to `45,000`;
-- keep the 5,000-step early-stopping patience and minimum selection step
-  unchanged, so the run may stop before 45,000 if a genuine plateau occurs;
+Batch 25 confirms that extra receiver-only catch-up time improves decoder fit
+but leaves the sender collision structure almost unchanged. The observed worst
+train link exactly reaches its sender's unique-code ceiling, so another horizon
+extension cannot solve the gate. A clean ECP7-B26 coordinate-descent progression
+is:
+
+- inherit the complete Batch 25 implementation and 45,000-step ceiling;
+- keep sender-only replay through step 20,000 and receiver-only replay through
+  step 30,000 exactly as in Batch 25;
+- route the additional replay branch sender-only again for steps 30,001–35,000,
+  then restore receiver-only replay from step 35,001;
+- reuse the registered 5,000-step sender-separation duration;
 - keep ordinary base-task updates to both senders and receivers unchanged;
 - keep weight `0.25`, the 64-meaning replay batch, 200-step refresh,
   seed-derived sampler, predicate and temperature unchanged;
 - continue to mine and replay training meanings only;
 - measure shared and any-link pool sizes, train exactness, injectivity,
   validation and new-reader induction, and rerun the frozen ECP-6 control.
 
-Do not combine the horizon change with another predicate, coefficient, graded
+Do not combine the routing cycle with another predicate, coefficient, graded
 sampling, collision pressure, a new architecture, factor-specific weights,
-optimizer changes, a second horizon change, variable-length messages or
+optimizer changes, a horizon change, variable-length messages or
 negotiation.
 
 ## 7. Definition of done for any contribution
 
@@ -32,6 +32,7 @@ this log.
 | ECP7-B22-I | Intervention | B21 replay gradients routed to senders but not receiver parameters | Position-aware MLP | valid negative; shared errors fall but link-specific fragmentation rises |
 | ECP7-B23-I | Intervention | B22 sender-only replay warmup followed by restored joint replay gradients | Position-aware MLP | valid negative; joint phase recreates shared errors without improving accuracy |
 | ECP7-B24-I | Intervention | B23 sender-only warmup followed by receiver-only replay catch-up | Position-aware MLP | valid negative; new-best means but link-specific errors and injectivity fail |
+| ECP7-B25-I | Intervention | B24 receiver-only catch-up with a 45,000-step maximum horizon | Position-aware MLP | valid negative; new-best means reach the non-injective code-cap |
 
 ## Batch 1 preregistration
 
@@ -1982,3 +1983,108 @@ A future ECP7-B25 may inherit Batch 24 unchanged and extend only `max_steps` to
 45,000. This isolates whether the receiver-only phase is still improving or
 merely approaching a non-injective plateau. No coefficient, route, predicate,
 architecture, optimizer or data setting may change with that horizon test.
+
+## Batch 25 preregistration
+
+Preregistered: July 18, 2026 at 17:26:20 UTC<br>
+Seed: `11`<br>
+Maximum population steps: `45,000`<br>
+Minimum selection steps: `15,000`<br>
+Early-stopping patience: `5,000` steps<br>
+Hard-meaning predicate: all `16` current population links have an incorrect factor<br>
+Replay sender-message gradients: **enabled through step 20,000; disabled from step 20,001**<br>
+Replay receiver-parameter gradients: **disabled through step 20,000; enabled from step 20,001**<br>
+Ordinary-task sender and receiver gradients: **enabled throughout**<br>
+Replay weight: `0` through step `15,000`, linear to `0.25` at step `20,000`, then hold<br>
+Replay batch size: `64` meanings, uniform with replacement<br>
+Training-only pool refresh interval: `200` steps<br>
+Replay sampler seed offset: `+1211`<br>
+Raw configuration SHA-256: `3c125aa0e21b29327b332a77f0a367633267804523eb941d5dad9eb622255387`<br>
+Effective configuration SHA-256: `00c9f13c5919f98f5b81c14c8ea2d852a1cd4be55ecbfcaf0d61abed6b9acc91`<br>
+Split-SHA-256: `4947058c75ab07cb43a87eb82776b12cb2a7e2eeba7114de110d3b852cbc64cd`<br>
+Test unsealed: **no**
+
+ECP7-B25-I inherits Batch 24 and changes only `training.max_steps` from
+`30,000` to `45,000`. The registered 5,000-step early-stopping patience remains
+active, so 45,000 is a ceiling rather than a required duration. All replay
+routes, timing, weights, data and selection criteria remain unchanged. The
+positive control is rerun, and no alternative horizon or accompanying
+intervention is admitted into this batch.
+
+## Batch 25 results
+
+Positive-control run: `runs/ecp7-batch25-control-development/20260718T172746Z-ecp7-development`<br>
+Intervention run: `runs/ecp7-batch25-intervention-development/20260718T172746Z-ecp7-development`<br>
+Test unsealed: **no**
+
+| Metric | Positive control | ECP7-B25-I |
+|---|---:|---:|
+| Population train, mean | 100% | **85.2631%** |
+| Population train, worst link | 100% | 82.8125% |
+| Population validation, mean | 100% | **85.2417%** |
+| Population validation, worst link | 100% | 82.7148% |
+| Universal translator, validation | 100% | **85.1318%** |
+| Exact sender-message agreement | 100% | 92.16% |
+| Unique messages per sender | 15,360 | 12,720–13,440 |
+| Collision meanings per sender | 0 | 1,920–2,640 |
+| Message entropy | 13.91 bits | 13.54–13.64 bits |
+| Development gate | pass | **fail** |
+
+The positive control passed every gate at 100%. ECP7-B25-I improves every
+primary B24 accuracy and sender agreement, setting new ECP-7 development bests
+for mean and worst-link train, mean and worst-link validation, and translator
+validation. It still fails both train thresholds and sender injectivity. The
+joint gate fails and confirmatory access remains unauthorized.
+
+All 151 optimization records through step 30,000 match Batch 24 exactly. The
+extended receiver-only phase raises mean validation from 84.45% at step 30,000
+to the selected 85.24% at step 43,800. Training reaches the 45,000-step ceiling
+with 85.09% validation, so the registered patience still does not terminate the
+run before the cap. Improvement after step 30,000 is real but slow and noisy.
+
+The all-link pool contains 1,710 meanings at step 30,000, 1,691 at step 40,000,
+1,677 at selection and 1,659 at the final step. The selected diagnostic finds
+1,677 meanings failed by all links and 2,967 failed by at least one link, with
+11.39 mean failed links per erroneous meaning. Relative to B24, the longer
+phase removes 175 any-link failures but only 13 shared failures.
+
+The selected codebooks explain the plateau. Their unique-message counts are
+13,440, 12,960, 12,720 and 13,440. Even a perfect receiver can therefore decode
+at most 87.50%, 84.38%, 82.81% and 87.50% of each sender's 15,360 training
+meanings. The mean code-cap is 85.55%; observed mean train is already 85.26%,
+and the observed worst link exactly equals the lowest 82.81% code-cap. Longer
+receiver-only optimization cannot satisfy the train gates without first
+changing sender collisions.
+
+Compared with Batch 24, B25 improves mean train by 0.58 percentage points,
+worst-link train by 0.25 points, mean validation by 0.76 points, worst-link
+validation by 0.49 points, translator validation by 0.59 points and sender
+agreement by 0.66 points. Receiver induction has caught up to the existing
+codes, but the remaining performance gap is structurally sender-side.
+
+Selected sender codebooks contain 13,440, 12,960, 12,720 and 13,440 unique
+messages, with 1,920, 2,400, 2,640 and 1,920 collision meanings. Population
+validation factor accuracies were `[99.99%, 100%, 98.44%, 86.82%]`;
+universal-translator factor accuracies were
+`[100%, 100%, 98.44%, 86.69%]`. The stable size collisions and sender-specific
+texture caps remain the limiting structure.
+
+Both sealed analyses report 65 matching artifact hashes, 16,384 validation-only
+episodes, no confirmatory-test keys, valid local alphabets, and exactly 14
+declared bits for every logged message.
+
+## Batch 25 decision
+
+ECP7-B25-I is a valid negative result and the strongest ECP-7 development
+variant so far. The extended catch-up phase improves receiver fit but reaches
+the mathematical ceiling imposed by non-injective sender codebooks. The
+confirmatory split remains sealed, and another horizon-only extension is not
+justified.
+
+A future ECP7-B26 may test one deterministic coordinate-descent cycle while
+keeping Batch 25's 45,000-step ceiling: sender-only replay through step 20,000,
+receiver-only through 30,000, a second sender-only separation pulse through
+35,000, and receiver-only catch-up afterward. The second sender pulse reuses the
+registered 5,000-step duration and directly tests whether saturated receivers
+provide a better fixed target for breaking remaining sender collisions. No
+other setting may change.
@@ -1053,3 +1053,41 @@ sealed-test requirement fails. No alternative switch step, gradient route,
 predicate, coefficient, replay batch, refresh, architecture, base loss,
 optimizer, duration, data, translator or seed may be tried inside Batch 24.
 Failure is recorded without confirmatory test access.
+
+## Batch 25 preregistration — Extended receiver-only catch-up horizon
+
+Status: **preregistered for sealed development**<br>
+Preregistered: July 18, 2026 at 17:26:20 UTC<br>
+Configuration: `config/ecp7-b25-development.yaml`<br>
+Raw configuration SHA-256: `3c125aa0e21b29327b332a77f0a367633267804523eb941d5dad9eb622255387`<br>
+Effective configuration SHA-256: `00c9f13c5919f98f5b81c14c8ea2d852a1cd4be55ecbfcaf0d61abed6b9acc91`
+
+Batch 24 established the strongest mean ECP-7 performance, selected its best
+checkpoint at step 29,400, and reached the 30,000-step ceiling before the
+5,000-step patience could assess a late plateau. Batch 25 isolates whether that
+receiver-only catch-up phase is still improving or has already converged to a
+non-injective solution.
+
+ECP7-B25-I inherits the complete Batch 24 implementation, configuration,
+gradient routes, hard-meaning predicate, replay mechanism, architecture,
+ordinary task batch, utilization objective, optimizer, data, temperature
+schedule, coefficient, replay batch, sampler, refresh, warmup, selection,
+translator and thresholds. The sole change is:
+
+- `training.max_steps` increases from 30,000 to 45,000.
+
+The 5,000-step early-stopping patience and 15,000-step minimum remain unchanged.
+The run may therefore finish before step 45,000. The complete optimization
+trajectory through step 30,000 must match Batch 24 exactly.
+
+The batch contains exactly two seed-11 runs:
+
+1. the unchanged ECP-6 positive control;
+2. ECP7-B25-I with the extended receiver-only catch-up ceiling.
+
+All existing validity and development gates remain unchanged. The intervention
+fails if any train, validation, translator, injectivity, channel-integrity or
+sealed-test requirement fails. No alternative horizon, gradient route,
+predicate, coefficient, replay batch, refresh, architecture, base loss,
+optimizer, data, translator or seed may be tried inside Batch 25. Failure is
+recorded without confirmatory test access.
@@ -136,6 +136,13 @@ validation. The selected error pool contains 1,690 shared failures and 3,142
 any-link failures, so worst-link accuracy and injectivity still fail. Exact
 identities and hashes are in `batch24-manifest.json`.
 
+Batch 25 extended only the receiver-only catch-up ceiling to 45,000 steps. It
+set new-best means of 85.26% train, 85.24% validation and 85.13% translator
+validation, but the observed worst train link exactly reached its sender's
+82.81% unique-code ceiling. The remaining failure is sender injectivity, not
+receiver training time. Exact identities and hashes are in
+`batch25-manifest.json`.
+
 Reproduce the development arms with:
 
 ```bash