@@ -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 eighteen sealed batches are valid
+one-factor-per-slot architecture. Its first nineteen sealed batches are valid
 development results but none passes the full gate. Batch 15 remains 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
@@ -36,7 +36,9 @@ added late worst-factor pressure and regressed to 76.46% validation. Batch 17
 replayed globally mined training collisions and modestly improved code use, but
 still regressed to 77.09% validation. Batch 18 reduced replay to task-loss scale,
 recovering 80.71% validation and a new-best 84.06% translator score, but train
-exactness and injectivity still failed. The ECP-6 positive controls remain
+exactness and injectivity still failed. Batch 19 bounded replay to a pulse and
+reached 82.04% validation plus a new-best 80.57% worst-link validation, but
+again failed train exactness and injectivity. 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).
 
@@ -96,6 +98,7 @@ Do not translate those artifacts in place. New experiment configurations and all
 - [`config/ecp7-b16-development.yaml`](config/ecp7-b16-development.yaml) — late normalized factor-minimax pressure on the Batch 15 base.
 - [`config/ecp7-b17-development.yaml`](config/ecp7-b17-development.yaml) — late training-only global collision-pair replay on the Batch 15 base.
 - [`config/ecp7-b18-development.yaml`](config/ecp7-b18-development.yaml) — the Batch 17 replay mechanism with final weight reduced to 0.1.
+- [`config/ecp7-b19-development.yaml`](config/ecp7-b19-development.yaml) — the Batch 18 replay mechanism decayed back to zero after step 20,000.
 - [`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 +1,17 @@
+extends: ecp7-b18-development.yaml
+
+experiment:
+  title: ECP-7 bounded global collision-replay pulse
+  arm: bounded_global_collision_replay_pulse
+  objective: >-
+    Test whether decaying the Batch 18 replay coefficient after its early
+    collision-reduction phase preserves code-use gains while restoring an
+    unperturbed final semantic-refinement phase.
+
+training:
+  global_collision_replay:
+    weight_decay:
+      enabled: true
+      start_step: 20000
+      end_step: 25000
+      final_weight: 0.0
@@ -41,7 +41,7 @@ python3.12 -m venv .venv
 .venv/bin/ecp6 --config config/ecp6.yaml validate
 ```
 
-Expected baseline: 102 passing tests and split sizes `14336/1024/1024`.
+Expected baseline: 105 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 18
+Do not modify ECP-0 through ECP-6 or the completed ECP-7 Batch 1 through Batch 19
 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,
@@ -114,11 +114,16 @@ ECP7-B18-I reduced replay weight to `0.1` and recovered 84.08% train, 80.71%
 validation and a new-best 84.06% translator validation. Validation and
 translator gates pass, but train and injectivity do not. Batch 15 retains the
 best validation result, and the confirmatory ECP-7 test remains sealed.
+ECP7-B19-I decayed replay back to zero from steps 20,000–25,000. It reached
+83.74% train, 82.04% validation, a new-best 80.57% worst-link validation and
+83.50% translator validation. The bounded pulse improved cross-link balance but
+still failed train thresholds and injectivity. Batch 15 retains the best mean
+validation result, 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-B19 intervention and its failure criterion.
+2. Define exactly one ECP7-B20 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.
@@ -131,26 +136,27 @@ Continue with this sequence:
 
 ## 6. Recommended next experiment
 
-Batch 18 shows that replay at task-loss scale avoids B17's early collapse and
-improves train and translator scores, but validation peaks at step 22,400 and
-then declines while replay remains active. The clean next question is whether a
-bounded replay pulse can preserve early collision reduction and then return to
-unperturbed semantic refinement. A clean ECP7-B19 progression is:
-
-- keep the same world and bit budget;
-- keep the complete B15 architecture, horizon and base losses;
-- keep the complete B17 training-only collision mining, replay batch,
-  temperature and refresh mechanism;
-- keep replay weight `0` through step 15,000, warm only to `0.1` at step 20,000,
-  then decay linearly to `0` at step 25,000 and hold;
-- measure injectivity, validation composition and new-reader induction;
-- rerun the ECP-6 architecture as the frozen positive control.
-
-Continue to use training meanings only for pair mining. Do not change the
-mining source, pair sampler, architecture, base utilization, optimizer, data,
-temperature, duration, validation cadence or development thresholds, and do
-not combine the schedule change with factor reweighting, variable-length
-messages or negotiation.
+Batches 17–19 show that collision-pair replay changes code use but does not
+stably separate the remaining collisions. Batch 19 improves validation only
+after replay is removed, while known-meaning train exactness remains near 84%.
+The clean next question is whether the residual error is concentrated in a
+sparse set of hard training meanings that uniform task batches do not repair.
+A clean ECP7-B20 progression is:
+
+- keep the same world, bit budget, Batch 15 architecture and 30,000-step horizon;
+- do not enable collision replay or factor reweighting;
+- periodically evaluate only the training codebook and deterministically mine
+  meanings reconstructed incorrectly by the current population;
+- add ordinary reconstruction updates on one fixed, registered number of mined
+  meanings after step 15,000, using a separate seed-derived sampler;
+- keep the ordinary random task batch and all base objectives unchanged;
+- measure train exactness, injectivity, validation composition and new-reader
+  induction, and rerun the frozen ECP-6 positive control.
+
+Preregister the mining score, refresh interval, replay batch size, coefficient
+and schedule before training. Do not combine hard-meaning replay with collision
+pressure, a new architecture, factor-specific weights, optimizer changes,
+longer training, variable-length messages or negotiation.
 
 ## 7. Definition of done for any contribution
 
 
@@ -26,6 +26,7 @@ this log.
 | ECP7-B16-I | Intervention | B15 with late normalized factor minimax | Position-aware MLP | completed; gate failed |
 | ECP7-B17-I | Intervention | B15 with late global collision-pair replay | Position-aware MLP | completed; gate failed |
 | ECP7-B18-I | Intervention | B17 replay with final weight 0.1 | Position-aware MLP | completed; gate failed |
+| ECP7-B19-I | Intervention | B18 with replay decayed to zero after step 20,000 | Position-aware MLP | valid negative; validation 82.04%, train and injectivity fail |
 
 ## Batch 1 preregistration
 
@@ -1353,3 +1354,103 @@ it at zero thereafter. This tests whether a bounded replay pulse can preserve
 the early collision reduction while returning the final phase to pure B15 task
 and utilization refinement. Mining, pair batch, refresh, architecture, base
 losses, optimizer, duration, data, translator and thresholds must not change.
+
+## Batch 19 preregistration
+
+Preregistered: July 18, 2026 at 15:15:36 UTC<br>
+Seed: `11`<br>
+Maximum population steps: `30,000`<br>
+Minimum selection steps: `15,000`<br>
+Early-stopping patience: `5,000` steps<br>
+Replay weight: `0` through step `15,000`, linear to `0.1` at step `20,000`, linear to `0` at step `25,000`, then hold<br>
+Replay pairs per sender and update: `32`<br>
+Training-codebook refresh interval: `200` steps<br>
+Relaxed replay temperature: `1.0`<br>
+Raw configuration SHA-256: `dab2cd0dadbadb47a10e2c45a720611dc91d140d3783f0ca1a537e4e8a8549c0`<br>
+Effective configuration SHA-256: `cbd144a5fce451315a7d5fb322a11f2f29c32514fc22fbe06f46285aef677d13`<br>
+Split-SHA-256: `4947058c75ab07cb43a87eb82776b12cb2a7e2eeba7114de110d3b852cbc64cd`<br>
+Test unsealed: **no**
+
+ECP7-B19-I is identical to Batch 18 through step 20,000. It then changes only
+the replay coefficient, decaying from `0.1` to zero by step 25,000. Mining
+continues for diagnostics, but the final 5,000 steps receive no replay gradient
+or sampler draws. This directly tests whether the early code-use gain can be
+retained while the final phase returns to the unchanged B15 objectives. The
+positive control is rerun, and no alternative pulse or accompanying
+intervention is admitted into this batch.
+
+## Batch 19 results
+
+Positive-control run: `runs/ecp7-batch19-control-development/20260718T151853Z-ecp7-development`<br>
+Intervention run: `runs/ecp7-batch19-intervention-development/20260718T151853Z-ecp7-development`<br>
+Test unsealed: **no**
+
+| Metric | Positive control | ECP7-B19-I |
+|---|---:|---:|
+| Population train, mean | 100% | **83.7385%** |
+| Population train, worst link | 100% | 82.7148% |
+| Population validation, mean | 100% | **82.0435%** |
+| Population validation, worst link | 100% | **80.5664%** |
+| Universal translator, validation | 100% | **83.4961%** |
+| Exact sender-message agreement | 100% | 91.81% |
+| Unique messages per sender | 15,360 | 12,709–13,200 |
+| Collision meanings per sender | 0 | 2,160–2,651 |
+| Message entropy | 13.91 bits | 13.54–13.61 bits |
+| Development gate | pass | **fail** |
+
+The positive control again passed every gate at 100%. ECP7-B19-I passed the
+validation and universal-translator thresholds, and its 80.57% worst-link
+validation is the strongest ECP-7 worst-link result so far. It still failed
+both known-meaning train thresholds and every sender remained non-injective.
+The joint gate therefore fails and confirmatory access remains unauthorized.
+
+All 101 logged trajectory records through step 20,000 match Batch 18 exactly.
+Replay then decayed as registered: weight was `0.06` at step 22,000, `0.02` at
+step 24,000 and zero from step 25,000 onward. At zero weight the logged replay
+loss is also zero and the sampler consumes no draws; deterministic full-codebook
+mining remains visible only through diagnostic collision-pair counts.
+
+Validation rose from 81.59% at the end of the decay to the selected 82.04% at
+step 26,200, with replay fully inactive. It later fluctuated and finished at
+81.20% at step 30,000. The selected checkpoint had task loss `0.06103`,
+utilization loss `-0.72858`, and mined collision-pair counts of 2,674, 2,645,
+2,921 and 2,352. The run reached the registered maximum horizon rather than an
+early stop.
+
+Compared with Batch 18, mean validation improved by 1.33 percentage points and
+sender agreement improved from 90.13% to 91.81%. Train exactness fell by 0.34
+points, translator validation fell by 0.56 points, and three of four senders
+had more collision meanings. Compared with Batch 15, the pulse slightly reduced
+collisions for three senders and produced a stronger worst validation link, but
+mean validation remained 0.54 points lower and train exactness remained lower.
+The selected sender codebooks contain 12,855, 12,886, 12,709 and 13,200 unique
+messages, with 2,505, 2,474, 2,651 and 2,160 collision meanings.
+
+Population validation factor accuracies were
+`[97.93%, 99.73%, 98.44%, 85.30%]`; universal-translator factor accuracies were
+`[100%, 100%, 98.44%, 85.06%]`. The bounded pulse therefore improves population
+color generalization and cross-link balance after replay is removed, but does
+not resolve the persistent texture errors or hard-code collisions.
+
+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 19 decision
+
+ECP7-B19-I is a valid negative result under the joint gate. A bounded replay
+pulse is less disruptive to validation than persistent replay, and its
+replay-free tail improves the worst population link, but the collision benefit
+does not persist and the train bottleneck remains essentially unchanged. Batch
+15 retains the strongest mean validation, while Batch 18 retains the strongest
+train and translator values. The confirmatory split remains sealed.
+
+A future ECP7-B20 should stop tuning collision-replay schedules. A clean next
+causal test is deterministic global hard-meaning task replay on the unchanged
+Batch 15 base: mine only training meanings that the current population
+reconstructs incorrectly, then add ordinary sender-receiver reconstruction
+updates on a registered bounded sample. This directly tests whether the
+remaining train error is a sparse hard-example problem without adding a
+meaning-to-slot binding or a code-distance surrogate. Collision replay, factor
+reweighting, architecture, optimizer, horizon, data, translator and thresholds
+must remain unchanged or absent as appropriate.
@@ -796,3 +796,45 @@ sealed-test requirement fails. No alternative coefficient, schedule, mining
 source, pair sampler, refresh, temperature, architecture, duration, optimizer,
 data, translator or seed may be tried inside Batch 18. Failure is recorded
 without confirmatory test access.
+
+## Batch 19 preregistration — Bounded collision-replay pulse
+
+Status: **preregistered for sealed development**<br>
+Preregistered: July 18, 2026 at 15:15:36 UTC<br>
+Configuration: `config/ecp7-b19-development.yaml`<br>
+Raw configuration SHA-256: `dab2cd0dadbadb47a10e2c45a720611dc91d140d3783f0ca1a537e4e8a8549c0`<br>
+Effective configuration SHA-256: `cbd144a5fce451315a7d5fb322a11f2f29c32514fc22fbe06f46285aef677d13`
+
+Batch 18 confirmed that replay at task-loss scale avoids B17's early collapse,
+but validation peaked at step 22,400 and then regressed while weight `0.1`
+remained active. Batch 19 tests whether replay is useful as a bounded codebook
+diversification pulse rather than a permanent late objective.
+
+ECP7-B19-I inherits the complete Batch 18 architecture, training-only mining,
+seed-derived pair sampler, 32-pair sender batches, 200-step refresh interval,
+relaxed temperature, base losses, optimizer, data, duration, selection,
+translator and thresholds. It is identical to B18 through step 20,000. The sole
+change is a coefficient decay:
+
+- replay weight remains `0` through step `15,000`;
+- replay weight rises linearly to `0.1` at step `20,000`;
+- replay weight falls linearly to `0` over steps `20,000–25,000`;
+- replay weight remains `0` through step `30,000`.
+
+Training-codebook collision mining continues at evaluation boundaries after
+step 25,000 for diagnostics, but contributes no gradient and consumes no replay
+sampler draws. The final optimization phase therefore contains only the B15
+task and utilization objectives. The complete trajectory through step 20,000
+must match B18 exactly.
+
+The batch contains exactly two seed-11 runs:
+
+1. the unchanged ECP-6 positive control;
+2. ECP7-B19-I with one bounded global collision-replay pulse.
+
+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 decay, endpoint, mining, pair
+sampler, architecture, base loss, optimizer, duration, data, translator or seed
+may be tried inside Batch 19. Failure is recorded without confirmatory test
+access.
@@ -98,6 +98,12 @@ exactness, 80.71% validation and a new-best 84.06% translator validation, but
 still failed train thresholds and injectivity. Exact identities and hashes are
 in `batch18-manifest.json`.
 
+Batch 19 decayed that replay weight back to zero after step 20,000. It reached
+83.74% train exactness, 82.04% validation, a new-best 80.57% worst-link
+validation and 83.50% translator validation. The bounded pulse improved
+cross-link balance but still failed train thresholds and injectivity. Exact
+identities and hashes are in `batch19-manifest.json`.
+
 Reproduce the development arms with:
 
 ```bash
 
@@ -0,0 +1,49 @@
+{
+  "batch": "ECP7-B19",
+  "confirmatory_test_unsealed": false,
+  "config_effective_sha256": "cbd144a5fce451315a7d5fb322a11f2f29c32514fc22fbe06f46285aef677d13",
+  "config_raw_sha256": "dab2cd0dadbadb47a10e2c45a720611dc91d140d3783f0ca1a537e4e8a8549c0",
+  "control": {
+    "artifact_hashes": {
+      "effective_config.yaml": "4bfb976293f117d3ae35f790e4fd6371628be95d91febb334bc275dc5d6b66d4",
+      "posthoc-analysis.json": "d38450c328b0528137d7873144338300f6542c54f28ed1a096329a5404bb13ac",
+      "seed-11/metrics.json": "3fa1aa30d7795153e24622d5d4f519b5f86b178fa79bb2b7a39c75c875016927",
+      "summary.json": "cfea6b314e74d04a0236e7eaf61ac70f9945f3e15513d343f19267b4093d7f46"
+    },
+    "development_gate_passed": true,
+    "population_train_mean_exact_match": 1.0,
+    "population_validation_mean_exact_match": 1.0,
+    "run_id": "20260718T151853Z-ecp7-development",
+    "universal_translator_validation_exact_match": 1.0,
+    "unique_messages_per_sender": [15360, 15360, 15360, 15360]
+  },
+  "intervention": {
+    "artifact_hashes": {
+      "effective_config.yaml": "05bfddf2185783338a4c9fc626b06f7dcb335d883380b0d5f14c5117eb897e76",
+      "posthoc-analysis.json": "e6954381e530120522873543d0e89a454d191e2e05e878fa1ee7b81c04bf661e",
+      "seed-11/metrics.json": "26d40d118482038fca8bb73930a646babd7c1ae8c7a41178254cd5d8d2528b15",
+      "summary.json": "965b1ca352583378eedfdacb72651d8656bda495f5307b952540ad592e2fc668"
+    },
+    "best_code_utilization_loss": -0.7285780310630798,
+    "best_global_collision_replay_loss": 0.0,
+    "best_global_collision_replay_pair_counts": [2674, 2645, 2921, 2352],
+    "best_global_collision_replay_weight": 0.0,
+    "best_step": 26200,
+    "collision_meanings_per_sender": [2505, 2474, 2651, 2160],
+    "development_gate_passed": false,
+    "exact_sender_message_agreement": 0.9180881076388889,
+    "final_step": 30000,
+    "message_entropy_bits_per_sender": [13.56302604477656, 13.567062503109893, 13.538167215820996, 13.60794791977656],
+    "population_train_mean_exact_match": 0.8373849051339286,
+    "population_train_worst_link_exact_match": 0.8271484375,
+    "population_validation_factor_accuracy": [0.97930908203125, 0.99725341796875, 0.984375, 0.85302734375],
+    "population_validation_mean_exact_match": 0.8204345703125,
+    "population_validation_worst_link_exact_match": 0.8056640625,
+    "run_id": "20260718T151853Z-ecp7-development",
+    "universal_translator_validation_exact_match": 0.8349609375,
+    "unique_messages_per_sender": [12855, 12886, 12709, 13200]
+  },
+  "seed": 11,
+  "split_artifact_sha256": "676195997ee0c77d5c9565bae045dbb938b432a2e4e74b666233b0a41f924718",
+  "split_sha256": "4947058c75ab07cb43a87eb82776b12cb2a7e2eeba7114de110d3b852cbc64cd"
+}