{
  "schemaVersion": 1,
  "kind": "odin-rnd.amendment",
  "experiment": "EXP 020",
  "amendment": "05",
  "name": "EXP 020 amendment 05. It is not a runner amendment of runner/PROTOCOL-S2.md, whose own amendments 01 to 04 are a different list; each is named in full below.",
  "title": "S2's run record: the pause, all 13 discards and the report's hashes",
  "authoredOn": "2026-11-08",
  "amends": {
    "preregistration": "experiments/mongodb-home-ground/preregistration.json",
    "preregistrationSha256": "128087dbba8b09c97fcbf2ab4a873b1c4b9c656b27ad07291ba4693e26d83fec",
    "frozenCommit": "0af2c4861c489531916884ceb6894ee771d0ec71",
    "after": {
      "amendment": "04",
      "record": "experiments/mongodb-home-ground/amendment-04/amendment.json",
      "sha256": "4caae68a2cb93f4f6f86f6290b656f3c566305e2e92deecee63cd23498a3e067"
    }
  },
  "why": "S2 is complete, and its report was computed under the published rules. How it got there is written only in the private run log: a pause of 13 days and 6 hours, a stack-stop incident, four more discarded cells than amendment 01 lists, and the report's own refusal of two cells. A reader of the verdict needs that record and the hashes of what the report read, so this amendment states them. It records facts and changes no rule.",
  "timing": "The report was computed once all 60 counted cells existed, on 2026-11-08, after amendment 04 was merged (odin-rnd #147, 2026-10-28), as amendment 04 requires. The runner owner drafted this record at private commit 9d24bd858d040ebc7c8e8e3a1a3422d3a7519e52 on 2026-11-08 at 15:02Z. It is published after the report was computed, because it records how the run went; it does not change how the report decides.",
  "changes": [
    {
      "id": "1",
      "what": "The pause of 13 days and 6 hours",
      "class": "run record: no rule change",
      "rule": "The last counted cell before the pause ended on 2026-10-25 at 19:58:54, and discards were applied at 21:33:41, with 39 counted cells kept. The plan was then held by the shared pause flag. Its 5 holders had all died, and the flag was stale from 2026-10-27 11:43. The flag was removed on 2026-11-08 at 01:54:03 and the plan resumed at 01:55:00. No counted cell ran between 2026-10-25 19:58:54 and 2026-11-08 01:55:00, 13 days, 5 hours and 56 minutes. The second pass re-measured exactly the missing cells, and the plan completed on 2026-11-08 at 09:22:22 with 60 of 60 counted cells.",
      "sameHost": "The runner owner reports the same host and the same Docker data disk as the kept cells. One pg_test_fsync run on 2026-11-08 shows fdatasync at 568.9 ops/s and open_datasync at 565.3 ops/s (one 8 kB write); the runner owner's 10-25 figures were 539 and 560. Only the 11-08 probe is a committed file."
    },
    {
      "id": "2",
      "what": "The stack-stop incident, with no discard",
      "class": "run record: no rule change",
      "rule": "On 2026-11-08 at 01:59:30, a wrong process check in a helper script that runs after the plan stopped the database stack in the middle of a cell. That cell and the 21 after it each failed in 0 to 1 s ('client not running') and wrote no file. No counted cell was affected. The script was fixed and the plan restarted at 02:02:10."
    },
    {
      "id": "3",
      "what": "All 13 discarded cell runs, kept in the runner's discarded/ directory and all re-run",
      "class": "run record: the disturbance rule of amendment 01, applied",
      "rule": "A counted cell that ran while other heavy work ran on the same host is discarded and re-run; nothing is pooled. Amendment 01 lists the first nine. This lists all 13.",
      "cells": [
        {
          "cell": "mongo-x1-rep0-seed1",
          "when": "2026-10-25",
          "why": "filming overlapped it; it ran before PROTOCOL-S2 runner amendments 01 to 04"
        },
        {
          "cell": "pgoutput-x1-rep0-seed1",
          "when": "2026-10-25",
          "why": "filming overlapped it; it ran before the runner amendments, and it also lost 1 event to the runner cursor bug that runner amendment 01 fixes"
        },
        {
          "cell": "outbox-x10-rep4-seed1",
          "when": "2026-10-25",
          "why": "the 13:07 to 14:18 full test suite with browsers; moved at 21:33:41 (amendment 01 records that these five were first kept)"
        },
        {
          "cell": "mongo-x10-rep4-seed1",
          "when": "2026-10-25",
          "why": "the same test suite"
        },
        {
          "cell": "mongo-x1-rep0-seed2",
          "when": "2026-10-25",
          "why": "the same test suite"
        },
        {
          "cell": "pgoutput-x1-rep0-seed2",
          "when": "2026-10-25",
          "why": "the same test suite"
        },
        {
          "cell": "outbox-x1-rep0-seed2",
          "when": "2026-10-25",
          "why": "the same test suite"
        },
        {
          "cell": "pgoutput-x1-rep4-seed2",
          "when": "2026-10-25",
          "why": "headless-browser work, 17:02 to 17:09; moved at 19:57:18"
        },
        {
          "cell": "mongo-x1-rep4-seed2",
          "when": "2026-10-25",
          "why": "headless-browser work, 19:34 to 19:51; moved at 19:59:09"
        },
        {
          "cell": "mongo-x10-rep2-seed2",
          "when": "2026-11-08",
          "why": "headless-browser console walks, window 05:58 to 06:30:48; moved at 06:10:13"
        },
        {
          "cell": "pgoutput-x10-rep2-seed2",
          "when": "2026-11-08",
          "why": "headless-browser console walks, window 06:39:48 to 07:04:03; moved at 07:03:28"
        },
        {
          "cell": "mongo-x10-rep3-seed2",
          "when": "2026-11-08",
          "why": "refused by the report's own overlap check at 08:55:25: the run was 07:03:56 to 07:17:35, and window 14 ends at 07:04:03 by a 20 s poll, so an overlap of up to 7 s cannot be ruled out"
        },
        {
          "cell": "outbox-x1-rep3-seed2",
          "when": "2026-10-25",
          "why": "refused by the report's own overlap check on 2026-11-08 at 08:55:25: the run was 16:49:27 to 17:02:45, and window 9 starts at 17:02:45; bounds are inclusive"
        }
      ],
      "disclosed": "The last two were refused by the report, not flagged in the disturbance log: the log names 11 discards, the discarded/ directory holds 13 files. The two are explained only in the run log, and the runner owner's drafter caught the gap. Times are the host's local time, CET (UTC+01:00)."
    },
    {
      "id": "4",
      "what": "The disturbance-window refusal rule, as it ran (amendment 04 item 8)",
      "class": "run record: amendment 04 item 8, applied",
      "rule": "disturbance-windows.json transcribes every row of the private disturbance log at private commit 576463d31edb53082c26764ad75ea788e91a86aa (sha256 e3bad331f58b12292a0c163c0e051f553a741d5e54fcc01f06de99b3ec4172ea) as 14 windows, 9 heavy and 5 light, at UTC+01:00, with inclusive bounds. The report writes nothing while any counted cell's run, from its t0 to its last event, overlaps a heavy window or cannot be placed in time. It lists by name the counted cells that overlap a light window. In the final report no counted cell overlaps a heavy window, and three overlap a light one: mongo-x1-rep2-seed2 (window 7), outbox-x10-rep3-seed1 (window 3) and pgoutput-x10-rep4-seed1 (window 3). All 60 run intervals came from the cells' own t0, none from the run log.",
      "judgement": "Calling the light rows (single-core unit tests, no browser, no database) not heavy is a judgement. It is recorded so a reader can disagree."
    },
    {
      "id": "5",
      "what": "The report's outputs, the files it read, and the 60 counted cells, by sha256",
      "class": "run record: hashes of private files",
      "rule": "The results were committed at private commit 2230e9cb28c922f06f7fc10b1acf77e16bb77cbb and stored byte for byte at 0838f7e232679811729aa891a2eb4ab5202ec58c. REPORT.json and REPORT.md are the report's outputs. RESULT.json is a later copy of REPORT.json's verdicts for the lab's fit map, added at 94125bedcf6ce50f24fa1c56a1869b2d128bccf4; it is not something the report read. The inputs are the frozen rule texts it applies, the report code, and the disturbance files it reads. plan.log and the fsync probe are run records the report does not read; they are listed because this record cites them. Paths are relative to the experiment's private directory. The 60 counted cells are listed with the sha256 the report records for each (REPORT.json, cell_sha256).",
      "outputs": [
        {
          "file": "runner/results/s2/REPORT.json",
          "sha256": "c6fd5b40a692e523f74234d5bbc3c51d7a2d0785f5d52130623c1a6ce991c1e7",
          "note": "equals RESULT.json's report_sha256"
        },
        {
          "file": "runner/results/s2/REPORT.md",
          "sha256": "8b939319e3780f25232cd3e8c16293615021e021eb624d090540d55459d6daa4"
        },
        {
          "file": "RESULT.json",
          "sha256": "9fc056a8859e5b15e4c2fa747fe863ea3baa35732c5548fac2744fa0cb56c07b"
        }
      ],
      "inputs": [
        {
          "file": "PREREG.md",
          "sha256": "83f35a8c2885dbf8c3e748a67cf0d4a574e2540ac9eb95b67bef2e9d9b5b1ca1",
          "note": "the frozen PREREG the pre-registration cites"
        },
        {
          "file": "runner/PROTOCOL-S2.md",
          "sha256": "c50a7d3020f0564c4c00b941324b6b1b723de5f1e07872018996edc93ec6554d",
          "note": "the private source of the protocol amendment 01 publishes with substitutions"
        },
        {
          "file": "runner/results/s2/disturbance-windows.json",
          "sha256": "e07ab72c217e4bbbe9037cd54d23e1e77b10da7c99642e846329b5522b4de762"
        },
        {
          "file": "runner/results/s2/DISTURBANCES.md",
          "sha256": "e3bad331f58b12292a0c163c0e051f553a741d5e54fcc01f06de99b3ec4172ea"
        },
        {
          "file": "runner/s2_report.py",
          "sha256": "6cfe7873c350290656e83b79dbe4138095a210888dd26712a1d5f8dc4d1a2d03",
          "note": "the report code amendment 04 pins"
        }
      ],
      "runRecords": [
        {
          "file": "runner/results/s2/plan.log",
          "sha256": "c3b2f58f152836e6a35ac4a62ae9b7a595b2cb15c21cf9e6604e06f98cdb58ad"
        },
        {
          "file": "runner/results/s2/fsync-probe-2026-11-08.txt",
          "sha256": "2ab18773946a5e0e2e06c850206e98d4776b7ca12aab83a60c0b0a0a5cc8ae5f"
        }
      ],
      "cells": [
        {
          "file": "mongo-x1-rep0-seed1.json",
          "sha256": "f7f38bf8017a71e69dfe972981f00472b9524d344f083791395183a73637186f"
        },
        {
          "file": "mongo-x1-rep0-seed2.json",
          "sha256": "c4aecfb3ecfacac77d3cb65dd7f7977678595431633a92db7d39b2f6f6630a5c"
        },
        {
          "file": "mongo-x1-rep1-seed1.json",
          "sha256": "68f94fed299e933894c9c9d8d8c9e75360cdbd11d14e1190f596da20c5bfcffd"
        },
        {
          "file": "mongo-x1-rep1-seed2.json",
          "sha256": "b11690ebb733af7f69f44e49ecafa78ed13af71cc86584ba3c60f46fd00a999f"
        },
        {
          "file": "mongo-x1-rep2-seed1.json",
          "sha256": "9f1c7fbfe711c0f5f976db6047a865bb6942606148b82c288e0fc8c53356437e"
        },
        {
          "file": "mongo-x1-rep2-seed2.json",
          "sha256": "358c4b1eaf7a4321d860e780740cf7f6678b0ad343b1a9ce1598d537bc783818"
        },
        {
          "file": "mongo-x1-rep3-seed1.json",
          "sha256": "047a5af73152f69f11326ec9f14519bf69082aff8ff1a3cc6303468a4dc78a10"
        },
        {
          "file": "mongo-x1-rep3-seed2.json",
          "sha256": "3cc4d82e18c8e774503363c593f182d057a618a1b1d8b56aee023a435f675388"
        },
        {
          "file": "mongo-x1-rep4-seed1.json",
          "sha256": "6aa5d513efb9fd2c3ba0f80ccd0049cdaf24f153c213162a04ea9f3813fc6a29"
        },
        {
          "file": "mongo-x1-rep4-seed2.json",
          "sha256": "fe3ecab4716515d2ad9d0575a794fc064083ee65a831bc09886f8993c9815b9a"
        },
        {
          "file": "mongo-x10-rep0-seed1.json",
          "sha256": "e99084f68cdcfea1c1e1eede2533cef3278c3995648fd15a1d057873bdd41e4d"
        },
        {
          "file": "mongo-x10-rep0-seed2.json",
          "sha256": "e3357d8f18be1e8cabc7f0f1a54d9ec43c794ee246f2a01befcce1c68c0dc1ec"
        },
        {
          "file": "mongo-x10-rep1-seed1.json",
          "sha256": "a5cff18f06f18782fbf8980e0e0ed2a18b18e022fbbe348a482585c7c4d71781"
        },
        {
          "file": "mongo-x10-rep1-seed2.json",
          "sha256": "6b8e562eb3c856909f70507e573eb0bf8e54cbf9cc85483f09c8d864e40fed95"
        },
        {
          "file": "mongo-x10-rep2-seed1.json",
          "sha256": "5e6972a0d9e19dbd04e0f9aa0520fb3c8c961f271d529ae44d0aaef94f2ee4a8"
        },
        {
          "file": "mongo-x10-rep2-seed2.json",
          "sha256": "155a46f5a98e47fceff8634e466a4a0984050c3227e8cf4d504cb94d16868bcf"
        },
        {
          "file": "mongo-x10-rep3-seed1.json",
          "sha256": "59e33fb7f303b343851cd4df1335bd603a5f012346be719e9d3e3d23f72c6ec3"
        },
        {
          "file": "mongo-x10-rep3-seed2.json",
          "sha256": "a52a309a59284034ecd55746b7fba966205e791487ab6b670ad35508d2495041"
        },
        {
          "file": "mongo-x10-rep4-seed1.json",
          "sha256": "3d6c8bb282ec218544b328a228b8ef6d936e4a575016b442144f33ed2769dcae"
        },
        {
          "file": "mongo-x10-rep4-seed2.json",
          "sha256": "cf9b59b955df46cf384d7dca6f9ddaa7e9cf22cb42640a0cecad9f013f033a49"
        },
        {
          "file": "outbox-x1-rep0-seed1.json",
          "sha256": "79c7ee3d31ae0c90db1e0e8b772f851b1edf35da16cf6ea23093b54654442adf"
        },
        {
          "file": "outbox-x1-rep0-seed2.json",
          "sha256": "d55e8c083304074525e60ed93f2bf0871b6875bde79cbb4792494ed9f359377f"
        },
        {
          "file": "outbox-x1-rep1-seed1.json",
          "sha256": "7fb21419452bd6e5a47dc0b7d664f068427e8d316ba8e3e1070eb67c316baf9f"
        },
        {
          "file": "outbox-x1-rep1-seed2.json",
          "sha256": "be47337fc135ef09824c9f60d9f8dc05292a58ab6079c69b4c1cbba88baa4192"
        },
        {
          "file": "outbox-x1-rep2-seed1.json",
          "sha256": "9f88f014c26c70ccfcd6dc828a4f0f20b0cf14c07a02a88ae093e6a024b54118"
        },
        {
          "file": "outbox-x1-rep2-seed2.json",
          "sha256": "4c63ed5e90a495b62cacc0c83020e6afe40365673475ac3d94c757a2ba01bc9e"
        },
        {
          "file": "outbox-x1-rep3-seed1.json",
          "sha256": "3d9a350ccf7ce6e8247d98640171dd9cad2d671e1c55bbbdba52fc01034fba79"
        },
        {
          "file": "outbox-x1-rep3-seed2.json",
          "sha256": "cd6fc10bd437c3a62a3d80a5a538634db546cd9dc62a7b8de0280437b17ee915"
        },
        {
          "file": "outbox-x1-rep4-seed1.json",
          "sha256": "896e41b9db7a2a3f8646e12c4689ce27bc408a17edb9fe9405158681b9bef672"
        },
        {
          "file": "outbox-x1-rep4-seed2.json",
          "sha256": "f944bc32976c6c6392e2cc1def8b1c3a330928fbe3f39caa0e3507de48ac2c6e"
        },
        {
          "file": "outbox-x10-rep0-seed1.json",
          "sha256": "04fecc57dc72083713fa6fec4cdc7647b71b5827664c5d4de25304efe518777a"
        },
        {
          "file": "outbox-x10-rep0-seed2.json",
          "sha256": "cdcb826007bf2709cad8f39bd92f165b7ea2b232659a4186a49be3008e7926f6"
        },
        {
          "file": "outbox-x10-rep1-seed1.json",
          "sha256": "c42e5a7915a34681384454419367caee75f4347d8503af93b58191aeeccb9a56"
        },
        {
          "file": "outbox-x10-rep1-seed2.json",
          "sha256": "842e083f7d9d6e55f18bee65728218d2540d51ca67004d0c47e969db598f79ba"
        },
        {
          "file": "outbox-x10-rep2-seed1.json",
          "sha256": "8971cab115f0549d0038a256d6124d5a0fa6668e5ed23d867f0683e87553035e"
        },
        {
          "file": "outbox-x10-rep2-seed2.json",
          "sha256": "2e5226c0784c4fbbbbde81337248c566923f72d99a834ced58c6796302910abd"
        },
        {
          "file": "outbox-x10-rep3-seed1.json",
          "sha256": "007a1ffdb4d75422f5577ad5cdb6d778a16db303734683cc89f6ba4e1d407085"
        },
        {
          "file": "outbox-x10-rep3-seed2.json",
          "sha256": "cbaf380a0efdc8175970690e57532c8f3e89da687c3e2eff9bcee47c64832a99"
        },
        {
          "file": "outbox-x10-rep4-seed1.json",
          "sha256": "aa729ab0cfa740f7adc3ae33209620c59da4d59188bab132b9af294dbade2f05"
        },
        {
          "file": "outbox-x10-rep4-seed2.json",
          "sha256": "3e98bfb4fc546217ecf0d347718608202f74dedd72c7581736a4b3479d018e0e"
        },
        {
          "file": "pgoutput-x1-rep0-seed1.json",
          "sha256": "9ca17ab5d8d67b3066a5f2b4b0ddeb1c17bbdc6ab3ca3011d5deb7a9220fa4b6"
        },
        {
          "file": "pgoutput-x1-rep0-seed2.json",
          "sha256": "bad10a1329adad6e5086204c642b68506ef6c1b7b29ad858f0f172cb59a2b206"
        },
        {
          "file": "pgoutput-x1-rep1-seed1.json",
          "sha256": "7c9f8e5c1958fc04e23ed0206ffdea06e1cf7062c48d0cb29689b9b9c0ac5fab"
        },
        {
          "file": "pgoutput-x1-rep1-seed2.json",
          "sha256": "2cc6524c80661bc45041709b95958dbd194e9e90fa5b1ad7183d0c6d1fa2d337"
        },
        {
          "file": "pgoutput-x1-rep2-seed1.json",
          "sha256": "2b23dfdd9acf81c7911cdbee4038e0d888875662c90c30a57703cd2e998aafd6"
        },
        {
          "file": "pgoutput-x1-rep2-seed2.json",
          "sha256": "901703f2dc1bcce85ce794dd7447d49a24142e0aa3b84ab4bb9dde2f02cf28f1"
        },
        {
          "file": "pgoutput-x1-rep3-seed1.json",
          "sha256": "9af3a40876e9edfaddc479d6cf610b0b2e65c3209103c86cb2e6c6cd588c6062"
        },
        {
          "file": "pgoutput-x1-rep3-seed2.json",
          "sha256": "f5ffeae7b5002b793fe489d09b71fcd6e3387d513f1bab6d9e02b151e39b830f"
        },
        {
          "file": "pgoutput-x1-rep4-seed1.json",
          "sha256": "5c7615939add314f1da0475ec6a3eafb69bc8bf21b5a51f50ca7bc501dfa5949"
        },
        {
          "file": "pgoutput-x1-rep4-seed2.json",
          "sha256": "b66623f8c0e7e9a3f55a1fad1e3b7bf9803baa5180984209fd378da20c53e389"
        },
        {
          "file": "pgoutput-x10-rep0-seed1.json",
          "sha256": "a77b36fcbf060cb528559097a4af4b098208c4278f0701dce795345107eb6b99"
        },
        {
          "file": "pgoutput-x10-rep0-seed2.json",
          "sha256": "9f81e618cd821f319e0af5e3b006f1397e743bdfb37903c8f030c11805fc2cf2"
        },
        {
          "file": "pgoutput-x10-rep1-seed1.json",
          "sha256": "c088134193d5e45160733384a6027e98125d0cc07aebbf36a7ac1cb011260846"
        },
        {
          "file": "pgoutput-x10-rep1-seed2.json",
          "sha256": "4951ecaf550af1763199ee12587b1bda34e64ea828482f6b1fd8ff43072f46db"
        },
        {
          "file": "pgoutput-x10-rep2-seed1.json",
          "sha256": "0128bbd5ed7121851595701ea28d6851c72413daee6a62f0022a680d79ea5ec9"
        },
        {
          "file": "pgoutput-x10-rep2-seed2.json",
          "sha256": "1f6221329bd59dd16ad2230903eb5de64f3958ab7deb1a98b9c3f59a127db7ec"
        },
        {
          "file": "pgoutput-x10-rep3-seed1.json",
          "sha256": "d07f1ebf14018f2915405aeaeeeece6c741864d27d7c025a03a7f2035d2ed50b"
        },
        {
          "file": "pgoutput-x10-rep3-seed2.json",
          "sha256": "1a6313f60570d53bf86dd8e1edceb148c3494bd3ecf8ac110d22adec102982f4"
        },
        {
          "file": "pgoutput-x10-rep4-seed1.json",
          "sha256": "180aa78d63b3ef5e83115a8244971077c90d3f20efba3de5a8aa28f17b3c76bc"
        },
        {
          "file": "pgoutput-x10-rep4-seed2.json",
          "sha256": "19749e439bc1a9e9ca6a95579874a327f39f6f73de8138ec6a5207bbd7b96b02"
        }
      ]
    },
    {
      "id": "6",
      "what": "The verdict, as the report states it",
      "class": "run record: the report's verdict, quoted, not recomputed here",
      "rule": "S2 is postgres-fits at 1x and at 10x, with pgoutput chosen as the PostgreSQL variant at both scales. At 1x PostgreSQL's lag p95 is at least 20 % lower in 5 of 5 seed-1 pairs, at 10x in 4 of 5, and at both scales the seed-2 reproduction agrees (median paired difference, MongoDB minus PostgreSQL, +4.8 ms and +8.4 ms). Every one of the 60 counted cells passed the gate, 0 lost and 0 duplicated. Secondary, never deciding: against the zero-component outbox variant alone, MongoDB fits at both scales (5 of 5 pairs, reproduced: -43.6 ms and -40.0 ms). The components count that decides is 4 / 5 / 0 (MongoDB / pgoutput / outbox); latency decided, so it was not reached. S1, S3 and S4 have not been run.",
      "source": "REPORT.json: verdicts.x1, verdicts.x10; scales.<x1|x10>.primary.wins and primary.repro.median_paired_diff_ms; scales.<x1|x10>.secondary.2_selection.chosen; scales.<x1|x10>.secondary.3_single_variant.outbox."
    }
  ],
  "documents": [],
  "sources": [
    {
      "what": "The runner owner's draft of this record, with the S2 result page",
      "commit": "9d24bd858d040ebc7c8e8e3a1a3422d3a7519e52",
      "file": "S2-RESULT-PAGE-DRAFT.md",
      "sha256": "239172cc053225f31f7265091c187b71e37df0f6a1062a097c4dd60782148027"
    }
  ],
  "verifiedHere": "The publisher fetched the ten listed files in change 5 from the private repository at commit 9d24bd85 and recomputed each sha256: all ten match. The 60 cell hashes are copied from REPORT.json; the publisher did not fetch the cell files. The publisher read change 6's numbers from REPORT.json (verdicts, the chosen variant, the wins, the reproduction medians, the outbox-alone line, 60 cell hashes, 14 / 9 / 5 windows with no heavy overlap and the three light overlaps) and the 13 file names in discarded/: all match this record. The publisher did not rerun the report.",
  "limits": [
    "The report, its inputs and the 60 cell files are in a private repository and are cited here by commit and sha256, not published. A reader can check a hash but cannot open the files, so the verdict in change 6 is the report's, quoted, not something this repository lets a reader recompute.",
    "The pause, the incident and the discard times are the runner owner's account and the private run log's; they are not independently verifiable from this repository.",
    "That the 11-08 cells ran on the same host and disk as the 10-25 cells is the runner owner's statement. The 10-25 fsync figures are the runner owner's numbers; only the 11-08 probe is a committed file.",
    "Two discards were refused by the report's overlap check and are missing from the disturbance log; they are explained only in the run log.",
    "Three counted cells are kept under a light overlap; calling those windows light is a judgement.",
    "The variant is chosen on seed 1 and compared on the same seed-1 data, which leans slightly toward PostgreSQL (amendments 02 and 03, and both advocates' reviews). Seed 2's lag values are out of sample; its gate outcomes are not.",
    "One machine and one OS stack. The pre-registered second-OS replication has not been run."
  ],
  "unchanged": "No rule changes. The pre-registration and amendments 01 to 04 stand and stay byte-identical: one PostgreSQL variant per scale, chosen once on seed 1; the 20 % margin, the 4-of-5 wins and the reproduction rule; the components count 4 / 5 / 0 with the literal counts beside it; the disturbance refusal of amendment 04 item 8. S1, S3 and S4 are not affected.",
  "notBefore": "Effective when merged to main of odin-labs-ai/odin-rnd. It records a computed report and holds nothing back."
}
