# Merge 03, batch r5-cg-continue: items k11–k15 (long training of CG's 19-8-2 head)

Inputs were read from `.imd/reads/artifacts/item-k11` … `item-k15`. Each file's SHA-256 matches the hash given in the brief. Every number below is copied from those files and nothing new was computed. The one check I added: for k11, k12, k13 and k15, I decompressed each `code_zb64` (base64, then zlib) and hashed it, and the hash equals that item's `code_sha256`.

## Summary

- **Best: k11.** All four non-blocked items are `valid: true` and pass gates A, B and C. k11 has the highest replica holdout over_CG: **+0.021462708923455566** (se_over_CG 0.0035040440435410245, n 8192, seeds 40000000–40008191). On the real engine it scored vs_CG **+0.014165628921265716, se 0.012998753166375516** (n 512) and vs_HA **+0.8729746068596396, se 0.03618064832565397** (n 512). afterSwap peak was 237730 with 0 OOG.
- **No item shows a real gain over CG that is clearly above noise.** Real vs_CG: k11 +0.0142 (se 0.0130), k12 +0.0162 (se 0.0148), k13 −0.0205 (se 0.0241), k15 0 (the candidate is CG itself). The replica holdout gains over CG (+0.013 to +0.021) are about the size of the real se. k12's notes say this run "does not establish a clear improvement over CG", and k13's notes say "improvement over CG is unconfirmed".
- **Real gain over HA is about +0.84 to +0.87 for every item, CG included.** k15 measures unchanged CG at vs_HA +0.841315854907559, se 0.022079922132378915, n 1024. That is far below the roughly +3.1 over HA the goal says the top 10 needs. No item claims a leaderboard rank.
- **The training curves are flat.** k11 searched for 70.0 minutes over 852 generations. Its monitor gain over CG was 0.0018 by generation 60, reached 0.0202 at generation 730, and was unchanged through generation 850. k12 (best monitor 0.0105) and k13 (best monitor 0.0038) gained even less. Unlike the pilot curves the goal describes as "still rising", these show only small gains from more search on CG's head.
- **k14 is blocked** at SETUP, line 6 (`uv venv --python 3.12 .venv`): "The prescribed downloaded uv-x86_64-unknown-linux-gnu binary failed with 'cannot execute binary file'; setup exited with status 126." The wheel checksum passed. No measurements exist for k14, so its row is null.
- **k15 ran no search.** It measured the replica at 85.73579144132528 markets/s, which is under the 200 markets/s threshold for search that its notes cite. It froze unchanged CG (code_sha256 e1ea8fcb…, the same as CG_code in the other items) and ran the real comparison on 1,024 seeds per arm. Its gate C is not a separate measurement: the item's `basis` field says "Frozen candidate is byte-for-byte CG; gate A also checks frozen code against frozen port". Its gas figures are CG's (there is no Candidate arm).

## Where items disagree or need care

- **Replica speed differs by about 10× across machines:** k11 885.3, k12 511.9, k13 508.4, k15 85.7 markets/s. k15 notes that it measured speed "while CG differential gate was also active".
- **`holdout se` means different things in different items.** In k11 and k15 it is the se of over_T. In k12 and k13 it is the se of the paired over_CG. Each row's `holdout_se_basis` says which.
- **`over_IB` on the replica is not a measured IB run.** k12, k13 and k15 say it is replica over_T minus a supplied reference of 1.7690. k11's value (1.973380001117099 → 0.20438000111709909) fits the same subtraction. Real vs_IB, by contrast, is a paired real-engine measurement in every item.
- **Gate B was copied rather than re-measured.** k11, k12, k13 and k15 all report the same gate B: mean_diff 1.920372244773669 over T on seeds 7232768–7233279, error 0.07962775522633092 against a reference of 2.0. Whether it was re-run on each machine or carried over is not stated beyond the seed list. k11 gives the error as signed (−0.0796); the others give the absolute value.
- **Seeds are separate across items.** Real seeds: k11 8040000–8041023, k12 8044000–8045023, k13 8048000–8049023, k15 8056000–8057023. These are disjoint, so the real vs_CG estimates are independent. The vs_CG and vs_HA arms used 512 seeds in k11, k12 and k13.
- **k13 has one outlier seed.** Its real vs_CG is pulled down by seed 8048118 at −11.026489541128058. The same seed is its worst vs_IB at −10.71.
- **Some sources failed re-check.** The re-check after the runs could not confirm sources cited by k11 (4), k12 (3), k13 (3) and k15 (1). The merge inputs do not include the items' sources lists, so I cannot tell which claims those sources backed. They most likely back the artifact URLs and hashes and the supplied reference values (CG real gain over T of 2.0, and IB 1.7690). The things that depend on those are the gate B reference, every replica `over_IB`, and the items' statement that their artifacts are verified. Read those with lower confidence. The items' own local measurements (speed, holdout, real-engine paired statistics, gas) do not rest on external pages. k14 says it relied on no external sources.

## Merged object

Numbers are strings copied from the item files. `null` means the value is missing (k14 is blocked, and k15 has no `final_sigma`). The `settings` field is the item's `search.settings` object, serialized as a JSON string. In `gates`, A and C give the maximum fee difference in bps. B gives the replica gain over T and its error against the reference. The `best` field carries k11's 178 head weights (19×8 W1 by row, then b1, W2 by row, b2) and its compressed Solidity. Its frozen scaling and toxicity anchor match CG's params, as k11's `params` shows.

```json
{
 "rows": [
  {
   "key": "k11",
   "result_sha256": "00c8ca82503f04f0d5754974ce2249977962d5d2917f34546723fa966af78448",
   "blocked": null,
   "markets_per_s": "885.2805485811128",
   "search_minutes": "70.00841818055682",
   "generations": "852",
   "settings": "{\"population\": 32, \"train_markets_per_generation\": 128, \"train_seed_start\": 1600000000, \"monitor_seeds\": [30000000, 30002047], \"monitor_interval_generations\": 10, \"ES_seed\": 611, \"sigma0\": 0.13, \"sigma_cap_rule\": \"0.13 * 0.5 ** (minutes / 40)\", \"CMA_stds\": {\"W1\": 0.16007274507182634, \"b1\": 0.18564858000307327, \"W2\": 0.12726224434596164, \"b2\": 0.1185932470281719, \"rule\": \"0.1*rms of HA W1 (13 old cols)/b1/W2/b2; new W1 columns start at 0 with W1 std\"}, \"selected_generation\": 730, \"selected_monitor_gain_over_CG\": 0.020241386728125282}",
   "holdout_over_T": "1.973380001117099",
   "holdout_over_CG": "0.021462708923455566",
   "holdout_over_IB": "0.20438000111709909",
   "holdout_se": "0.01203788413512287",
   "holdout_se_basis": "se over T (item also gives se_over_CG 0.0035040440435410245)",
   "real_vs_IB": {
    "n": "1024",
    "mean_diff": "0.1886632659365033",
    "se": "0.02172195734153784",
    "sum": "193.19118431897937",
    "sumsq": "530.7286093502587"
   },
   "real_vs_CG": {
    "n": "512",
    "mean_diff": "0.014165628921265716",
    "se": "0.012998753166375516",
    "sum": "7.252802007688047",
    "sumsq": "44.310067407666025"
   },
   "afterSwap_peak": "237730",
   "OOG": "0",
   "gates": {
    "A": {
     "passed": "true",
     "max_fee_diff_bps": "8.814065310946972e-05"
    },
    "B": {
     "passed": "true",
     "mean_diff": "1.920372244773669",
     "se": "0.04445924198500627",
     "error_vs_reference": "-0.07962775522633092"
    },
    "C": {
     "passed": "true",
     "max_fee_diff_bps": "0.00014263716742457522"
    }
   },
   "valid": "true",
   "code_sha256": "435222bb3438472b7d9a6d854cabe94b5bfda8baa215672c3cb698875235973a"
  },
  {
   "key": "k12",
   "result_sha256": "4ade6e596cc0bb424a89ff43d141b4bc37704c4a8dab2a2e79980d5cbff37a32",
   "blocked": null,
   "markets_per_s": "511.88132498285955",
   "search_minutes": "78.26283725315007",
   "generations": "186",
   "settings": "{\"population\": 32, \"dimension\": 178, \"CMA_stds\": {\"W1\": 0.16007274507182634, \"b1\": 0.18564858000307327, \"W2\": 0.12726224434596164, \"b2\": 0.1185932470281719}, \"sigma0\": 0.06, \"sigma_cap\": \".06 * .5**(minutes/40)\", \"seed\": 612, \"train_markets_per_generation\": 512, \"train_seed_start\": 1610000000, \"train_seed_end\": 1610095231, \"monitor_seeds\": [30000000, 30002047], \"monitor_every_generations\": 10, \"monitor_candidates\": [\"ES mean\", \"generation best\"], \"start_from\": \"CG.params, frozen weights; no restart used\", \"quantization\": 1e-06, \"segment_limit_seconds\": 470}",
   "holdout_over_T": "1.9651322370771847",
   "holdout_over_CG": "0.0132149448835411",
   "holdout_over_IB": "0.19613223707718475",
   "holdout_se": "0.002807087339675516",
   "holdout_se_basis": "se of paired over_CG",
   "real_vs_IB": {
    "n": "1024",
    "mean_diff": "0.17649639022343344",
    "se": "0.023036094795472986",
    "sum": "180.73230358879584",
    "sumsq": "587.79428602544"
   },
   "real_vs_CG": {
    "n": "512",
    "mean_diff": "0.01618944170056258",
    "se": "0.014813085264326134",
    "sum": "8.28899415068804",
    "sumsq": "57.54344857200851"
   },
   "afterSwap_peak": "237796",
   "OOG": "0",
   "gates": {
    "A": {
     "passed": "true",
     "max_fee_diff_bps": "8.814065310946972e-05"
    },
    "B": {
     "passed": "true",
     "mean_diff": "1.920372244773669",
     "se": "0.04445924198500627",
     "error_vs_reference": "0.07962775522633092"
    },
    "C": {
     "passed": "true",
     "max_fee_diff_bps": "4.185328508245212e-05"
    }
   },
   "valid": "true",
   "code_sha256": "c18d92949c7a3f140ed26a37d2c5e1e0659f11f0d3f6f29a030797b5ed0f32ea"
  },
  {
   "key": "k13",
   "result_sha256": "2678249c6a684054f5d8e80670972de9c59a54b235a6f626a41d0d429339bd93",
   "blocked": null,
   "markets_per_s": "508.3895581852567",
   "search_minutes": "71.28115384304353",
   "generations": "289",
   "settings": "{\"population\": 16, \"train_n\": 512, \"sigma0\": 0.06, \"seed\": 613, \"train_start\": 1620000000, \"monitor\": [30000000, 30002047], \"CMA_stds\": {\"W1\": 0.16007274507182634, \"b1\": 0.18564858000307327, \"W2\": 0.12726224434596164, \"b2\": 0.1185932470281719, \"rule\": \"0.1*rms of HA W1 (13 old cols)/b1/W2/b2; new W1 columns start at 0 with W1 std\"}, \"dimension\": 178, \"coordinates\": \"CMA_stds by weight group; first 39 port entries frozen\", \"rounding\": 1e-06, \"sigma_cap\": \"0.06 * 0.5 ** (training_minutes/40)\", \"monitor_every_generations\": 10, \"monitor_candidates\": [\"ES mean\", \"generation best\"], \"training_seed_end\": 1620147967, \"segments_under_seconds\": 470}",
   "holdout_over_T": "1.9652998056364657",
   "holdout_over_CG": "0.013382513442822243",
   "holdout_over_IB": "0.19629980563646576",
   "holdout_se": "0.0030277580676922847",
   "holdout_se_basis": "se of paired over_CG (se_over_T 0.011964158298412936)",
   "real_vs_IB": {
    "n": "1024",
    "mean_diff": "0.14975893396099332",
    "se": "0.02244528643796813",
    "sum": "153.35314837605716",
    "sumsq": "550.7131513846427"
   },
   "real_vs_CG": {
    "n": "512",
    "mean_diff": "-0.020488143246266688",
    "se": "0.024108886874051733",
    "sum": "-10.489929342088544",
    "sumsq": "152.2854911262492"
   },
   "afterSwap_peak": "237857",
   "OOG": "0",
   "gates": {
    "A": {
     "passed": "true",
     "max_fee_diff_bps": "8.814065310946972e-05"
    },
    "B": {
     "passed": "true",
     "mean_diff": "1.920372244773669",
     "se": "0.04445924198500627",
     "error_vs_reference": "0.07962775522633092"
    },
    "C": {
     "passed": "true",
     "max_fee_diff_bps": "0.00015030935170134985"
    }
   },
   "valid": "true",
   "code_sha256": "ac58fa7d7fc0b5e9c0f6f5d12be21c692c46b0ae1864d86d40d1d34112a2936d"
  },
  {
   "key": "k14",
   "result_sha256": "d8715c99c724beb5d8cd57642629d29956c73156fe49e2782df04be1eaf8f926",
   "blocked": "{\"step\": \"SETUP: sh scratch/setup.sh, line 6, uv venv --python 3.12 .venv\", \"error\": \"The prescribed downloaded uv-x86_64-unknown-linux-gnu binary failed with 'cannot execute binary file'; setup exited with status 126.\"}",
   "markets_per_s": null,
   "search_minutes": null,
   "generations": null,
   "settings": null,
   "holdout_over_T": null,
   "holdout_over_CG": null,
   "holdout_over_IB": null,
   "holdout_se": null,
   "holdout_se_basis": null,
   "real_vs_IB": null,
   "real_vs_CG": null,
   "afterSwap_peak": null,
   "OOG": null,
   "gates": null,
   "valid": "false",
   "code_sha256": null
  },
  {
   "key": "k15",
   "result_sha256": "7ad59abb38dc1fe37ce10980e3a1d2b7dce255e0b6089d1ee2a93d49629c8f27",
   "blocked": null,
   "markets_per_s": "85.73579144132528",
   "search_minutes": "0",
   "generations": "0",
   "settings": "{\"performed\": false, \"reason\": \"Measured replica speed below 200 markets/s; brief mandates no search and real comparison fallback.\", \"starting_weights\": \"CG.params, never restart\", \"requested_if_fast\": {\"popsize\": 32, \"sigma0\": 0.1, \"seed\": 615, \"CMA_stds\": {\"W1\": 0.16007274507182634, \"b1\": 0.18564858000307327, \"W2\": 0.12726224434596164, \"b2\": 0.1185932470281719, \"rule\": \"0.1*rms of HA W1 (13 old cols)/b1/W2/b2; new W1 columns start at 0 with W1 std\"}}}",
   "holdout_over_T": "1.9519172921936434",
   "holdout_over_CG": "0.0",
   "holdout_over_IB": "0.1829172921936435",
   "holdout_se": "0.011938004553452971",
   "holdout_se_basis": "se over T (over_CG is identity, se 0.0)",
   "real_vs_IB": {
    "n": "1024",
    "mean_diff": "0.17698617441536102",
    "se": "0.024058804443675513",
    "sum": "181.23384260132968",
    "sumsq": "638.4262930761347"
   },
   "real_vs_CG": {
    "n": "1024",
    "mean_diff": "0.0",
    "se": "0.0",
    "sum": "0.0",
    "sumsq": "0.0"
   },
   "afterSwap_peak": "237781",
   "OOG": "0",
   "gates": {
    "A": {
     "passed": "true",
     "max_fee_diff_bps": "8.814065310946972e-05"
    },
    "B": {
     "passed": "true",
     "mean_diff": "1.920372244773669",
     "se": "0.04445924198500627",
     "error_vs_reference": "0.07962775522633092"
    },
    "C": {
     "passed": "true",
     "max_fee_diff_bps": "8.814065310946972e-05"
    }
   },
   "valid": "true",
   "code_sha256": "e1ea8fcb105e00754d17061efbe4b299a634b0fcc1b5c293d65b631654b7204b"
  }
 ],
 "best": {
  "key": "k11",
  "result_sha256": "00c8ca82503f04f0d5754974ce2249977962d5d2917f34546723fa966af78448",
  "weights": [
   "0.129844",
   "0.419127",
   "-0.703899",
   "2.50629",
   "-2.480452",
   "-0.43582",
   "-1.003903",
   "2.158519",
   "-4.230766",
   "-1.823209",
   "-0.919142",
   "-0.761985",
   "0.84221",
   "-1.388248",
   "0.442359",
   "0.079172",
   "-0.88994",
   "-0.150719",
   "0.998601",
   "0.527504",
   "0.887763",
   "1.544678",
   "2.990861",
   "0.433288",
   "0.631",
   "-0.497414",
   "-0.033939",
   "-0.383692",
   "0.471565",
   "-0.208931",
   "-1.749589",
   "0.105062",
   "-0.563169",
   "0.046434",
   "1.900084",
   "0.138644",
   "0.06198",
   "0.824283",
   "1.75874",
   "3.506284",
   "-1.078465",
   "1.221521",
   "1.906392",
   "2.734431",
   "-1.839449",
   "1.79335",
   "1.701533",
   "-1.436806",
   "0.020735",
   "-1.085962",
   "-0.611506",
   "0.339415",
   "0.019866",
   "1.642143",
   "-2.71109",
   "-0.115456",
   "-1.600456",
   "-0.019639",
   "-1.430411",
   "0.072509",
   "-0.268972",
   "-1.150734",
   "-0.696577",
   "0.10663",
   "0.616197",
   "0.405435",
   "0.363966",
   "0.239675",
   "-1.806264",
   "-0.832666",
   "0.104579",
   "0.187996",
   "-0.446903",
   "1.941352",
   "-0.630485",
   "-0.65095",
   "1.000729",
   "0.560806",
   "4.679224",
   "-0.667068",
   "-2.385853",
   "-1.09404",
   "-0.39428",
   "-0.356498",
   "-0.163821",
   "0.020973",
   "0.000499",
   "-2.062425",
   "0.076885",
   "-0.815889",
   "0.08381",
   "-0.452723",
   "2.231589",
   "-0.080439",
   "0.532597",
   "1.442165",
   "-1.821302",
   "1.185049",
   "-2.209401",
   "3.60829",
   "1.362203",
   "2.269648",
   "0.830352",
   "-1.134197",
   "-0.43834",
   "-0.039096",
   "3.490205",
   "0.546536",
   "-0.739046",
   "-0.268057",
   "0.426139",
   "-1.602727",
   "-0.497174",
   "-1.059578",
   "-1.992251",
   "1.06588",
   "-0.728802",
   "-0.361917",
   "2.675252",
   "0.900914",
   "-1.292229",
   "1.913547",
   "1.167031",
   "-1.121515",
   "1.387281",
   "1.294231",
   "1.282513",
   "-0.093666",
   "-1.415019",
   "-1.165784",
   "0.098185",
   "-1.461508",
   "2.316427",
   "-2.479034",
   "-3.334191",
   "3.482305",
   "4.104827",
   "-0.743726",
   "-0.702782",
   "-1.819255",
   "1.96109",
   "0.713474",
   "-0.156944",
   "2.703341",
   "-1.870607",
   "1.256107",
   "-0.161656",
   "0.685668",
   "1.642714",
   "-0.311485",
   "-1.663623",
   "1.518926",
   "0.916292",
   "-1.359472",
   "-4.690644",
   "-0.474938",
   "-1.602754",
   "-0.958064",
   "-1.172942",
   "0.299292",
   "0.391851",
   "-2.133821",
   "0.25533",
   "0.037524",
   "-3.379549",
   "-0.412199",
   "-1.002861",
   "1.124167",
   "0.722118",
   "-1.527642",
   "-1.372618",
   "3.614112",
   "0.164838",
   "-1.814071",
   "0.700343",
   "0.118016",
   "1.39209",
   "2.086497"
  ],
  "code_zb64": "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",
  "code_sha256": "435222bb3438472b7d9a6d854cabe94b5bfda8baa215672c3cb698875235973a"
 }
}
```