@@ -26,15 +26,12 @@ 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 fourteen sealed batches are valid
-development results but none passes the full gate. Batch 10 remains strongest,
-extending the position-aware design to 15,000 optimization steps. It reaches
-82.08% train exactness, 72.98% validation and 75.15% translator validation while
-using 12,358–12,906 hard messages per sender. The translator threshold now passes,
-but the registered train, validation and injectivity thresholds do not. Batch
-11's utilization-weight decay regressed all principal measures, and Batch 12's
-deeper decoder regressed further. Batch 13's deeper sender collapsed almost
-completely, and Batch 14's late learning-rate decay also regressed. The ECP-6
+one-factor-per-slot architecture. Its first fifteen sealed batches are valid
+development results but none passes the full gate. Batch 15 is now strongest:
+the position-aware design at 30,000 optimization steps reaches 83.46% train
+exactness, 82.59% validation and 83.37% translator validation while using
+12,585–13,200 hard messages per sender. Validation and translator thresholds pass
+together, but the registered train and injectivity thresholds do not. The ECP-6
 positive controls remain perfect and the ECP-7 confirmatory test remains sealed. See
 [`docs/development-log-ecp7.md`](docs/development-log-ecp7.md).
 
@@ -90,6 +87,7 @@ Do not translate those artifacts in place. New experiment configurations and all
 - [`config/ecp7-b12-development.yaml`](config/ecp7-b12-development.yaml) — one additional generic shared decoder layer on the Batch 10 base.
 - [`config/ecp7-b13-development.yaml`](config/ecp7-b13-development.yaml) — one additional generic shared sender layer on the Batch 10 base.
 - [`config/ecp7-b14-development.yaml`](config/ecp7-b14-development.yaml) — late learning-rate decay on the Batch 10 base.
+- [`config/ecp7-b15-development.yaml`](config/ecp7-b15-development.yaml) — the Batch 10 design with a 30,000-step population 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 +1,14 @@
+extends: ecp7-b10-development.yaml
+
+experiment:
+  title: ECP-7 position-aware decoder with a 30,000-step horizon
+  arm: position_aware_decoder_30000_step_optimization
+  objective: >-
+    Test whether continued constant-rate optimization beyond the strongest
+    15,000-step checkpoint can cross the remaining development gates without
+    changing the architecture, losses, or temperature schedule.
+
+training:
+  max_steps: 30000
+  minimum_steps: 15000
+  patience_steps: 5000
@@ -41,7 +41,7 @@ python3.12 -m venv .venv
 .venv/bin/ecp6 --config config/ecp6.yaml validate
 ```
 
-Expected baseline: 88 passing tests and split sizes `14336/1024/1024`.
+Expected baseline: 89 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 14
+Do not modify ECP-0 through ECP-6 or the completed ECP-7 Batch 1 through Batch 15
 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,
@@ -98,15 +98,17 @@ to 79.09% train, 72.22% validation and 72.41% translator validation. B10 remains
 the strongest base. ECP7-B12-I added one shared decoder layer and regressed
 further to 71.53% train, 47.98% validation and 63.04% translator validation.
 ECP7-B13-I added the symmetric sender layer and collapsed to 1.32% train, 0.32%
-validation and 0.61% translator validation. The confirmatory ECP-7 test is still
-sealed. ECP7-B14-I decayed late learning rate and regressed to 79.13% train,
-65.49% validation and 71.02% translator validation. B10 remains the strongest
-base.
+validation and 0.61% translator validation. ECP7-B14-I decayed late learning
+rate and regressed to 79.13% train, 65.49% validation and 71.02% translator
+validation. ECP7-B15-I extended constant-rate training to 30,000 steps and is
+now strongest at 83.46% train, 82.59% validation and 83.37% translator
+validation. Its validation and translator gates pass, but train and injectivity
+do not. The confirmatory ECP-7 test is still sealed.
 
 Continue with this sequence:
 
 1. Read `docs/research-design-ecp7.md` and `docs/development-log-ecp7.md`.
-2. Define exactly one ECP7-B15 intervention and its failure criterion.
+2. Define exactly one ECP7-B16 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.
@@ -119,25 +121,23 @@ Continue with this sequence:
 
 ## 6. Recommended next experiment
 
-Batch 10 is the strongest weak-structure result. Batch 11 ruled out reduced late
-utilization pressure, while Batches 12 and 13 ruled out extra generic decoder
-and sender depth. Batch 14 showed that reducing late learning rate is also
-harmful, while B10 selected its final 15,000-step checkpoint. The most
-informative next step is one final duration-only extension. A clean ECP7-B15
-progression is:
+Batch 15 is the strongest weak-structure result. It passes validation and
+translator gates, with residual known-meaning error concentrated in size after
+the high-entropy code is established. Earlier Batch 6 minimax pressure started
+from step zero on a much weaker protocol. The most informative next step is one
+late worst-factor schedule. A clean ECP7-B16 progression is:
 
 - keep the same world and bit budget;
-- return to the complete B10 training and loss schedule;
-- keep constant learning rate `0.001` and the temperature endpoint `0.8`;
-- extend only maximum population steps to 30,000, minimum selection to 15,000
-  and patience to 5,000;
+- keep the complete B15 architecture, horizon and base losses;
+- keep normalized factor-minimax weight `0` through step 15,000;
+- warm only that weight linearly to `1.0` over steps 15,000–20,000, then hold;
 - measure injectivity, validation composition and new-reader induction;
 - rerun the ECP-6 architecture as the frozen positive control.
 
-Do not change architecture, loss weights, optimizer, data, temperature,
-validation cadence or development thresholds, and do not combine the longer
-horizon with variable-length messages or negotiation. That would make its
-effect uninterpretable.
+Do not change architecture, utilization, optimizer, data, temperature, duration,
+validation cadence or development thresholds, and do not combine the late
+minimax schedule with variable-length messages or negotiation. That would make
+its effect uninterpretable.
 
 ## 7. Definition of done for any contribution
 
 
@@ -992,3 +992,79 @@ updates is harmful. A future ECP7-B15 may therefore keep the exact B10
 architecture, losses, constant learning rate and temperature endpoint, then
 extend only the maximum population horizon to 30,000 steps with minimum
 selection step 15,000 and patience 5,000. No other setting may change.
+
+## Batch 15 preregistration
+
+Date: July 18, 2026<br>
+Seed: `11`<br>
+Maximum population steps: `30,000`<br>
+Minimum selection steps: `15,000`<br>
+Early-stopping patience: `5,000` steps<br>
+Learning rate: constant `0.001`<br>
+Temperature schedule steps: `5,000`, then hold at `0.8`<br>
+Split-SHA-256: `4947058c75ab07cb43a87eb82776b12cb2a7e2eeba7114de110d3b852cbc64cd`<br>
+Test unsealed: **no**
+
+ECP7-B15-I keeps every Batch 10 architecture, loss, optimizer, data and
+translator setting. It extends only population optimization and its selection
+window. The unchanged positive control is rerun, and no alternative horizon or
+accompanying intervention is admitted into this batch.
+
+## Batch 15 results
+
+Positive-control run: `runs/ecp7-batch15-control-development/20260718T134437Z-ecp7-development`<br>
+Intervention run: `runs/ecp7-batch15-intervention-development/20260718T134437Z-ecp7-development`<br>
+Test unsealed: **no**
+
+| Metric | Positive control | ECP7-B15-I |
+|---|---:|---:|
+| Population train, mean | 100% | 83.4612% |
+| Population train, worst link | 100% | 81.9336% |
+| Population validation, mean | 100% | **82.5867%** |
+| Population validation, worst link | 100% | 79.7852% |
+| Universal translator, validation | 100% | **83.3740%** |
+| Exact sender-message agreement | 100% | 92.33% |
+| Unique messages per sender | 15,360 | 12,585–13,200 |
+| Collision meanings per sender | 0 | 2,160–2,775 |
+| Message entropy | 13.91 bits | 13.52–13.61 bits |
+| Development gate | pass | **fail** |
+
+The positive control passed every gate at 100%. ECP7-B15-I passed the registered
+mean validation and universal-translator thresholds for the first time in one
+batch. It still failed known-meaning train exactness and sender injectivity, so
+the joint gate fails and confirmatory access remains unauthorized.
+
+Relative to B10, the longer constant-rate horizon raised validation from
+72.9797% to 82.5867% and translator validation from 75.1465% to 83.3740%.
+Population train exactness increased only from 82.0836% to 83.4612%, indicating
+that residual hard-message collisions, not held-out composition, are now the
+principal gate failure.
+
+Population validation factor accuracies were
+`[98.86%, 99.75%, 98.44%, 85.05%]`; universal-translator factor accuracies were
+`[100%, 100%, 98.44%, 84.94%]`. Color and shape became essentially perfect.
+Size is now the dominant residual factor, with the fixed texture error also
+preventing exact reconstruction.
+
+The selected checkpoint was step 28,800 with task loss `0.0675` and utilization
+loss `-0.7399`. Validation first crossed 80% between steps 24,000 and 26,000 and
+remained above it through the final step. Every sender used at least 12,585
+messages, and no collision bucket contained more than three meanings.
+
+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 15 decision
+
+ECP7-B15-I is a negative result under the joint gate but becomes the new
+strongest weak-structure base. It passes composition and new-reader induction;
+only known-meaning exactness and injectivity remain below their registered
+thresholds. The confirmatory split remains sealed.
+
+The residual error is concentrated in size after the shared code is already
+established. A future ECP7-B16 may keep the complete B15 design and introduce
+only late normalized worst-factor emphasis: factor-minimax weight `0` through
+step 15,000, linearly warm to `1.0` over 5,000 steps, then hold. This revisits the
+Batch 6 objective only after the high-entropy B15 protocol exists and must not
+change architecture, utilization, duration, data, translator or thresholds.
@@ -626,3 +626,40 @@ All existing validity and development gates remain unchanged. Failure is
 recorded without confirmatory test access. No alternative final rate, decay
 shape, decay boundary, architecture, loss, duration, optimizer or seed may be
 tried inside Batch 14.
+
+## Batch 15 preregistration — 30,000-step optimization horizon
+
+Status: **preregistered for sealed development**<br>
+Preregistered: July 18, 2026<br>
+Configuration: `config/ecp7-b15-development.yaml`<br>
+Raw configuration SHA-256: `2b4ed0979cb1f6a6aac0409e2624bf982fd020306492706f4aa0e6b3a8c9d15d`<br>
+Effective configuration SHA-256: `d60b8a9aebe75514560b35790a3de9674990724fff0a9c52283c91ff429a487b`
+
+Batch 10 selected its final 15,000-step checkpoint. Batch 14 showed that
+decaying late learning rate harms validation, while Batches 11–13 rejected loss
+removal and added capacity. Batch 15 therefore tests one remaining optimization
+hypothesis: the constant-rate B10 trajectory needs more time.
+
+ECP7-B15-I keeps the complete B10 architecture, sender, receiver, translator,
+losses, AdamW settings, data, temperature schedule, evaluation cadence and
+thresholds. It changes only the population optimization horizon and associated
+selection window:
+
+- maximum population steps increase from 15,000 to 30,000;
+- minimum selection steps increase from 5,000 to 15,000;
+- early-stopping patience increases from 3,000 to 5,000 steps;
+- learning rate remains constant at `0.001`;
+- temperature still reaches `0.8` at step 5,000 and holds thereafter.
+
+Translator training remains unchanged. Smoke runs are mechanical only and may
+not authorize tuning.
+
+The batch contains exactly two seed-11 runs:
+
+1. the unchanged ECP-6 positive control;
+2. ECP7-B15-I with the extended B10 population optimization horizon.
+
+All existing validity and development gates remain unchanged. Failure is
+recorded without confirmatory test access. No alternative horizon, patience,
+learning rate, architecture, loss, temperature, optimizer or seed may be tried
+inside Batch 15.
@@ -78,6 +78,11 @@ exactness, 65.49% validation and 71.02% translator validation. Constant late
 learning rate is therefore retained. Exact identities and hashes are in
 `batch14-manifest.json`.
 
+Batch 15 extended the constant-rate horizon to 30,000 steps. It reached 83.46%
+train exactness, 82.59% validation and 83.37% translator validation. Validation
+and translator gates passed together, but train exactness and injectivity still
+failed. Exact identities and hashes are in `batch15-manifest.json`.
+
 Reproduce the development arms with:
 
 ```bash
 
@@ -0,0 +1,43 @@
+{
+  "batch": "ECP7-B15",
+  "confirmatory_test_unsealed": false,
+  "control": {
+    "artifact_hashes": {
+      "effective_config.yaml": "4bfb976293f117d3ae35f790e4fd6371628be95d91febb334bc275dc5d6b66d4",
+      "posthoc-analysis.json": "751c592fb1921d6f6fb5d68bd7b7b351523766db3f4aa6be7f1412c3130937a4",
+      "seed-11/metrics.json": "708b3c84a8c4a6b086b34a55f6d388c637fbf65c076555c7a9912526f5684f47",
+      "summary.json": "8cb14faab55e787d62dcaabd9ac2215fa100b21dc58d4bfc6b91505348f4c5c0"
+    },
+    "development_gate_passed": true,
+    "population_train_mean_exact_match": 1.0,
+    "population_validation_mean_exact_match": 1.0,
+    "run_id": "20260718T134437Z-ecp7-development",
+    "universal_translator_validation_exact_match": 1.0,
+    "unique_messages_per_sender": [15360, 15360, 15360, 15360]
+  },
+  "intervention": {
+    "artifact_hashes": {
+      "effective_config.yaml": "218926e9313ba3dfb7ae3fc52fd3fa8a2f954627253f6d00fbf76109d01aeca5",
+      "posthoc-analysis.json": "b471cb9c8e69fa5bfaf5bf20eff060924911a31969b8ae0a321e6c126b9a4c6b",
+      "seed-11/metrics.json": "baeb1840fab41d08655ec4cff90e77d98b04f45ea33cb2140589c79e7f1dd841",
+      "summary.json": "4565ca20b499167f99ed2463185d097c4e6fbe563611682141630300036a1ff0"
+    },
+    "best_code_utilization_loss": -0.7398611903190613,
+    "best_step": 28800,
+    "collision_meanings_per_sender": [2520, 2490, 2775, 2160],
+    "development_gate_passed": false,
+    "exact_sender_message_agreement": 0.9232530381944444,
+    "message_entropy_bits_per_sender": [13.56107291977656, 13.56497916977656, 13.516074677555256, 13.60794791977656],
+    "population_train_mean_exact_match": 0.8346121651785714,
+    "population_train_worst_link_exact_match": 0.8193359375,
+    "population_validation_factor_accuracy": [0.9886474609375, 0.99749755859375, 0.984375, 0.8504638671875],
+    "population_validation_mean_exact_match": 0.82586669921875,
+    "population_validation_worst_link_exact_match": 0.7978515625,
+    "run_id": "20260718T134437Z-ecp7-development",
+    "universal_translator_validation_exact_match": 0.833740234375,
+    "unique_messages_per_sender": [12840, 12870, 12585, 13200]
+  },
+  "seed": 11,
+  "split_artifact_sha256": "676195997ee0c77d5c9565bae045dbb938b432a2e4e74b666233b0a41f924718",
+  "split_sha256": "4947058c75ab07cb43a87eb82776b12cb2a7e2eeba7114de110d3b852cbc64cd"
+}
@@ -113,3 +113,8 @@ def ecp7_b13_config():
 @pytest.fixture(scope="session")
 def ecp7_b14_config():
     return load_config(PROJECT_ROOT / "config" / "ecp7-b14-development.yaml")
+
+
+@pytest.fixture(scope="session")
+def ecp7_b15_config():
+    return load_config(PROJECT_ROOT / "config" / "ecp7-b15-development.yaml")
@@ -511,3 +511,21 @@ def test_learning_rate_decay_must_be_bounded(ecp7_b14_config):
     ] = 0.001
     with pytest.raises(ConfigError, match="below its initial value"):
         validate_config(invalid_final)
+
+
+def test_ecp7_b15_changes_only_the_population_optimization_horizon(
+    ecp7_b10_config, ecp7_b15_config
+):
+    assert ecp7_b15_config["world"] == ecp7_b10_config["world"]
+    assert ecp7_b15_config["dataset"] == ecp7_b10_config["dataset"]
+    assert ecp7_b15_config["channel"] == ecp7_b10_config["channel"]
+    assert ecp7_b15_config["agents"] == ecp7_b10_config["agents"]
+    b10_training = deepcopy(ecp7_b10_config["training"])
+    b15_training = deepcopy(ecp7_b15_config["training"])
+    assert b10_training.pop("max_steps") == 15000
+    assert b15_training.pop("max_steps") == 30000
+    assert b10_training.pop("minimum_steps") == 5000
+    assert b15_training.pop("minimum_steps") == 15000
+    assert b10_training.pop("patience_steps") == 3000
+    assert b15_training.pop("patience_steps") == 5000
+    assert b15_training == b10_training