{
  "title": "The Simplifier · The Record",
  "url": "https://thesimplifier7.github.io/record/",
  "source": "https://github.com/TheSimplifier7/record/blob/main/index.html",
  "history": "https://github.com/TheSimplifier7/record/commits/main",
  "standard_from": "2026-07-03",
  "last_updated": "2026-10-10",
  "generated": "2026-10-10T16:42:08.164Z",
  "tally": {
    "resolved": 35,
    "held": 26,
    "wrong": 9,
    "never_reached": 5,
    "open": 0,
    "hit": 2,
    "note": "Never reached counts for neither side. A hit is a held row that also went through its room, counted once, as held. Standing calls are not in these figures."
  },
  "grammar": {
    "from": "2026-09-28",
    "named": 6,
    "slate": 1,
    "copper": 5,
    "graded": 6,
    "held": 4,
    "hit": 2,
    "wrong": 2,
    "expected": 4,
    "note": "Expected is the sum of the odds each graded row was printed with on the day it was named: what the rows would have held if the read added nothing. See note 34."
  },
  "rows": [
    {
      "id": "lrow-0",
      "anchor": "r-2026-10-05-gold-4170-41",
      "url": "https://thesimplifier7.github.io/record/#r-2026-10-05-gold-4170-41",
      "named": "2026-10-05",
      "type": "Map",
      "metal": "Gold",
      "level": "4,170.41",
      "book": "Turn",
      "call": "Gold stays below 4,170.41, the high of the box it opened this week in, frozen Monday afternoon and read off the 8h chart at 22:06 CEST. Room 4,124.48, the box low. Odds at naming 61 in 100, off the data window. Named on the week's own box before any 8h close through it.",
      "kill": "a weekly close above 4,170.41.",
      "result": "WRONG. The weekly close printed 4,194.085, 23.675 above the line of 4,170.41 and 69.605 above the room, 4,124.48. Close read off the founder's 10 minute TVC chart, market shut, 23:00 CEST 9 October. See note 46.",
      "grade": "miss",
      "grade_word": "wrong",
      "grammar": 1,
      "side": "copper",
      "line": "4,170.41",
      "room": "4,124.48",
      "odds": 61,
      "odds_from": "data window",
      "line_from": "box high",
      "box": [
        "4,124.48",
        "4,170.41"
      ],
      "last_range": null,
      "read_at": "22:06 CEST",
      "close": "4,194.085",
      "hit": false,
      "pre_standard": false,
      "graded_on_close": "2026-10-09",
      "graded_at": "2026-10-09T23:00:00+02:00",
      "written": "2026-10-10",
      "commit": "https://github.com/TheSimplifier7/record/commit/62dd841"
    },
    {
      "id": "lrow-1",
      "anchor": "r-2026-10-05-silver-62-036",
      "url": "https://thesimplifier7.github.io/record/#r-2026-10-05-silver-62-036",
      "named": "2026-10-05",
      "type": "Map",
      "metal": "Silver",
      "level": "62.036",
      "book": "Turn",
      "call": "Silver stays below 62.036, the high of the box it opened this week in, frozen Monday afternoon and read off the 8h chart at 22:05 CEST. Room 60.326, the box low. Odds at naming 60 in 100, off the data window. Named on the week's own box before any 8h close through it.",
      "kill": "a weekly close above 62.036.",
      "result": "HELD. The weekly close printed 60.7790, 1.257 under the line of 62.036 and 0.453 above the room, 60.326. Close read off the founder's 10 minute TVC chart, market shut, 23:00 CEST 9 October. See note 46.",
      "grade": "pass",
      "grade_word": "held",
      "grammar": 1,
      "side": "copper",
      "line": "62.036",
      "room": "60.326",
      "odds": 60,
      "odds_from": "data window",
      "line_from": "box high",
      "box": [
        "60.326",
        "62.036"
      ],
      "last_range": null,
      "read_at": "22:05 CEST",
      "close": "60.7790",
      "hit": false,
      "pre_standard": false,
      "graded_on_close": "2026-10-09",
      "graded_at": "2026-10-09T23:00:00+02:00",
      "written": "2026-10-10",
      "commit": "https://github.com/TheSimplifier7/record/commit/62dd841"
    },
    {
      "id": "lrow-2",
      "anchor": "r-2026-10-05-platinum-1699-00",
      "url": "https://thesimplifier7.github.io/record/#r-2026-10-05-platinum-1699-00",
      "named": "2026-10-05",
      "type": "Map",
      "metal": "Platinum",
      "level": "1,699.00",
      "book": "Turn",
      "call": "Platinum holds above 1,699.00, the low of the box it opened this week in, frozen Monday afternoon and read off the 8h chart at 22:06 CEST. Room 1,740.40, the box high. Odds at naming 59 in 100, off the data window. Named on the week's own box before any 8h close through it.",
      "kill": "a weekly close below 1,699.00.",
      "result": "WRONG. The weekly close printed 1,686.90, 12.10 under the line of 1,699.00 and 53.50 under the room, 1,740.40. Close read off the founder's 8h TVC chart, market shut, 23:14 CEST 9 October. See note 46.",
      "grade": "miss",
      "grade_word": "wrong",
      "grammar": 1,
      "side": "slate",
      "line": "1,699.00",
      "room": "1,740.40",
      "odds": 59,
      "odds_from": "data window",
      "line_from": "box low",
      "box": [
        "1,699.00",
        "1,740.40"
      ],
      "last_range": null,
      "read_at": "22:06 CEST",
      "close": "1,686.90",
      "hit": false,
      "pre_standard": false,
      "graded_on_close": "2026-10-09",
      "graded_at": "2026-10-09T23:00:00+02:00",
      "written": "2026-10-10",
      "commit": "https://github.com/TheSimplifier7/record/commit/62dd841"
    },
    {
      "id": "lrow-3",
      "anchor": "r-2026-09-28-gold-4280-50",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-28-gold-4280-50",
      "named": "2026-09-28",
      "type": "Map",
      "metal": "Gold",
      "level": "4,280.50",
      "book": "Turn",
      "call": "Gold stays below 4,280.50, the high of the box it opened this week in, frozen Monday afternoon and read off the 8h chart at 18:03 CEST. Room 4,135.74, the box low. Odds at naming 78 in 100, off the data window. Named on the week's own box before any 8h close through it.",
      "kill": "a weekly close above 4,280.50.",
      "result": "HELD. The weekly close printed 4,142.710, 137.79 under the line of 4,280.50 and 6.97 above the room, 4,135.74. Close read off the founder's 8h TVC chart, market shut, 23:00 CEST 2 October. See note 44.",
      "grade": "pass",
      "grade_word": "held",
      "grammar": 1,
      "side": "copper",
      "line": "4,280.50",
      "room": "4,135.74",
      "odds": 78,
      "odds_from": "data window",
      "line_from": "box high",
      "box": [
        "4,135.74",
        "4,280.50"
      ],
      "last_range": null,
      "read_at": "18:03 CEST",
      "close": "4,142.710",
      "hit": false,
      "pre_standard": false,
      "graded_on_close": "2026-10-02",
      "graded_at": "2026-10-02T23:00:00+02:00",
      "written": "2026-10-02",
      "commit": "https://github.com/TheSimplifier7/record/commit/eaee555"
    },
    {
      "id": "lrow-4",
      "anchor": "r-2026-09-28-silver-64-212",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-28-silver-64-212",
      "named": "2026-09-28",
      "type": "Map",
      "metal": "Silver",
      "level": "64.212",
      "book": "Turn",
      "call": "Silver stays below 64.212, the high of the box it opened this week in, frozen Monday afternoon and read off the 8h chart at 18:03 CEST. Room 60.932, the box low. Odds at naming 71 in 100, off the data window. Named on the week's own box before any 8h close through it.",
      "kill": "a weekly close above 64.212.",
      "result": "HIT. The weekly close printed 60.4470, 3.765 under the line of 64.212 and 0.485 under the room, 60.932. A hit is counted once, as held. Close read off the founder's 8h TVC chart, market shut, 23:00 CEST 2 October. See note 44.",
      "grade": "pass",
      "grade_word": "held, hit the room",
      "grammar": 1,
      "side": "copper",
      "line": "64.212",
      "room": "60.932",
      "odds": 71,
      "odds_from": "data window",
      "line_from": "box high",
      "box": [
        "60.932",
        "64.212"
      ],
      "last_range": null,
      "read_at": "18:03 CEST",
      "close": "60.4470",
      "hit": true,
      "pre_standard": false,
      "graded_on_close": "2026-10-02",
      "graded_at": "2026-10-02T23:00:00+02:00",
      "written": "2026-10-02",
      "commit": "https://github.com/TheSimplifier7/record/commit/eaee555"
    },
    {
      "id": "lrow-5",
      "anchor": "r-2026-09-28-platinum-1775-80",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-28-platinum-1775-80",
      "named": "2026-09-28",
      "type": "Map",
      "metal": "Platinum",
      "level": "1,775.80",
      "book": "Turn",
      "call": "Platinum stays below 1,775.80, the high of the box it opened this week in, frozen Monday afternoon and read off the 8h chart at 18:03 CEST. Room 1,720.72, the box low. Odds at naming 71 in 100, off the data window. Named on the week's own box before any 8h close through it.",
      "kill": "a weekly close above 1,775.80.",
      "result": "HIT. The weekly close printed 1,699.70, 76.10 under the line of 1,775.80 and 21.02 under the room, 1,720.72. A hit is counted once, as held. Close read off the founder's 8h TVC chart, market shut, 23:00 CEST 2 October. See note 44.",
      "grade": "pass",
      "grade_word": "held, hit the room",
      "grammar": 1,
      "side": "copper",
      "line": "1,775.80",
      "room": "1,720.72",
      "odds": 71,
      "odds_from": "data window",
      "line_from": "box high",
      "box": [
        "1,720.72",
        "1,775.80"
      ],
      "last_range": null,
      "read_at": "18:03 CEST",
      "close": "1,699.70",
      "hit": true,
      "pre_standard": false,
      "graded_on_close": "2026-10-02",
      "graded_at": "2026-10-02T23:00:00+02:00",
      "written": "2026-10-02",
      "commit": "https://github.com/TheSimplifier7/record/commit/eaee555"
    },
    {
      "id": "lrow-6",
      "anchor": "r-2026-09-20-gold-4337",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-20-gold-4337",
      "named": "2026-09-20",
      "type": "Map",
      "metal": "Gold",
      "level": "4,337",
      "book": "Turn",
      "call": "Gold holds above 4,337, the top of the box it opened the week of 14 September in, read off the 8h chart on 20 September and published as 4,337. The Fed wick of 16 September went under the bottom of that box without a close; gold then left it through the top and finished the week 40 above it at 4,377.290. First row on the opening box, the object the record moves to tonight by the founder's ruling (note 29). Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 4,337.",
      "result": "The kill fired. The weekly close printed 4,284.130 against a floor of 4,337, 52.87 below the level as published. MISS. Gold printed under the line on Monday, closed back over it on the 8h chart on Tuesday night at 4,357.68, and went back under on Wednesday; the founder's 8h export of 23 September carries those bars. The line is the edge of the box as published and does not move, so the close is measured against the number as published and nothing else. Close read off the TVC chart, market shut, 23:07 CEST 25 September. See note 42.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-09-25",
      "graded_at": "2026-09-25T23:00:00+02:00",
      "written": "2026-09-25",
      "commit": "https://github.com/TheSimplifier7/record/commit/22e78b5"
    },
    {
      "id": "lrow-7",
      "anchor": "r-2026-09-20-silver-64-08",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-20-silver-64-08",
      "named": "2026-09-20",
      "type": "Map",
      "metal": "Silver",
      "level": "64.08",
      "book": "Turn",
      "call": "Silver holds above 64.08, the top of the box it opened the week of 14 September in, read off the 8h chart on 20 September and published as 64.08. Silver left that box through the top and finished the week 2.14 above it at 66.2210. First row for silver since 23 August. The retired door of 66.80 sits 2.72 above this line, and that is said here rather than left for a reader to find (note 29). Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 64.08.",
      "result": "Held. The weekly close printed 64.2560, 0.176 above the level as published. PASS. The pass came at arm's length: silver traded under the line during the week, 63.8480 on the founder's chart at 08:21 CEST on Friday, and the weekly close came back over it by 0.176. The kill is a weekly close below 64.08 and none printed. Close read off the TVC chart, market shut, 23:08 CEST 25 September. See note 42.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-09-25",
      "graded_at": "2026-09-25T23:00:00+02:00",
      "written": "2026-09-25",
      "commit": "https://github.com/TheSimplifier7/record/commit/22e78b5"
    },
    {
      "id": "lrow-8",
      "anchor": "r-2026-09-20-platinum-1746",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-20-platinum-1746",
      "named": "2026-09-20",
      "type": "Map",
      "metal": "Platinum",
      "level": "1,746",
      "book": "Turn",
      "call": "Platinum holds above 1,746, the bottom of the box it opened the week of 14 September in, read off the 8h chart on 20 September and published as 1,746. Platinum is the one metal that finished the week inside its box, 1,800.40 against a top of 1,803, so the row names the edge that would take it out the bottom and states the top so the reader has both. Fourth row for platinum on this page. Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 1,746.",
      "result": "Held. The weekly close printed 1,777.80, 31.80 above the level as published. PASS. Platinum went through the line on Wednesday, a low of 1,740.486 on the founder's 8h export of 23 September, 5.51 under it, and the weekly close came back above it. The kill is a weekly close below 1,746 and none printed. Close read off the TVC chart, market shut, 23:08 CEST 25 September. See note 42.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-09-25",
      "graded_at": "2026-09-25T23:00:00+02:00",
      "written": "2026-09-25",
      "commit": "https://github.com/TheSimplifier7/record/commit/22e78b5"
    },
    {
      "id": "lrow-9",
      "anchor": "r-2026-09-06-gold-4357",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-06-gold-4357",
      "named": "2026-09-06",
      "type": "Map",
      "metal": "Gold",
      "level": "4,357",
      "book": "Turn",
      "call": "Gold holds above the floor read off the 6 September chart, labelled 4,357.307 and published as 4,357. The line moved up from 4,350.628 in a week and the floor moved with it; it would not have moved the other way. Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 4,357.",
      "result": "The kill fired. The weekly close printed 4,347.760 with the 464 swing slow mean read at 4,357.307 tonight, the same read as the label the level was taken from on 6 September. The close sat 9.24 below the level as named and 9.547 below the line as it reads tonight; the row is graded on the number as published, which is 4,357. MISS. The floor had held on the 4 September close by 80.52 and it was moved up 7 dollars for this week because the line moved up; the close then printed under both the old floor and the new one, so the move did not decide the grade. First weekly close under gold's line since the 7 August close that turned its clock. Close and line read off the TVC chart, market shut, 03:09 CEST 12 September. See note 26.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-09-11",
      "graded_at": "2026-09-11T23:00:00+02:00",
      "written": "2026-09-12",
      "commit": "https://github.com/TheSimplifier7/record/commit/10cf76a"
    },
    {
      "id": "lrow-10",
      "anchor": "r-2026-09-06-platinum-1797-40",
      "url": "https://thesimplifier7.github.io/record/#r-2026-09-06-platinum-1797-40",
      "named": "2026-09-06",
      "type": "Map",
      "metal": "Platinum",
      "level": "1,797.40",
      "book": "Turn",
      "call": "Platinum holds above 1,797.40, the floor named on 30 August and held on the 4 September close. The 464 swing slow mean reads 1,797.35 on the 6 September chart, five cents lower; a floor does not follow a line down, so the published number stands for a second week. Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 1,797.40.",
      "result": "The kill fired. The weekly close printed 1,791.50 with the 464 swing slow mean read at 1,797.35 tonight, unchanged from the 6 September read and five cents under the number as published. The close sat 5.90 below the level as named and 5.85 below the line as it reads tonight; the row is graded on the number as published, which is 1,797.40, and the five cents did not decide anything. MISS. Third row for platinum on this page and its first wrong one, after two held. First weekly close under platinum's line since the 19 August close that turned its clock. Close and line read off the TVC chart, market shut, 03:10 CEST 12 September. See note 26.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-09-11",
      "graded_at": "2026-09-11T23:00:00+02:00",
      "written": "2026-09-12",
      "commit": "https://github.com/TheSimplifier7/record/commit/10cf76a"
    },
    {
      "id": "lrow-11",
      "anchor": "r-2026-08-30-gold-4350",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-30-gold-4350",
      "named": "2026-08-30",
      "type": "Map",
      "metal": "Gold",
      "level": "4,350",
      "book": "Turn",
      "call": "Gold holds above the floor read off the 30 August chart, labelled 4,350.628 and published as 4,350. Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 4,350.",
      "result": "The kill never fired. The weekly close printed 4,430.520 with the 464 swing slow mean read at 4,350.628 tonight, the same read as the label the level was taken from on 30 August. The close sat 80.52 above the level as named and 79.89 above the line as it reads tonight; the row is graded on the number as published, which is 4,350. PASS. Close and line read off the TVC chart, market shut, 03:04 CEST 5 September. See note 17.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-09-04",
      "graded_at": "2026-09-04T23:00:00+02:00",
      "written": "2026-09-05",
      "commit": "https://github.com/TheSimplifier7/record/commit/3b563f4"
    },
    {
      "id": "lrow-12",
      "anchor": "r-2026-08-30-platinum-1797-40",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-30-platinum-1797-40",
      "named": "2026-08-30",
      "type": "Map",
      "metal": "Platinum",
      "level": "1,797.40",
      "book": "Turn",
      "call": "Platinum holds above the floor read off the 30 August chart, 1,797.40. Second row for the newest metal on this page, same method, same rule, one week after its first grade. Under the ruling in note 11 this is a claim about a price level and not a claim about the swing clock, so it keeps a price kill and it is graded on the close alone. Named before the week opened.",
      "kill": "a weekly close below 1,797.40.",
      "result": "The kill never fired. The weekly close printed 1,819.60 with the 464 swing slow mean read at 1,797.40 tonight, which is the same figure as the level named on 30 August, so this row quotes one number twice for the same reason the 23 August platinum row did. The close sat 22.20 above the line. PASS. Second row for platinum, second pass. Close and line read off the TVC chart, market shut, 03:04 CEST 5 September. See note 17.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-09-04",
      "graded_at": "2026-09-04T23:00:00+02:00",
      "written": "2026-09-05",
      "commit": "https://github.com/TheSimplifier7/record/commit/3b563f4"
    },
    {
      "id": "lrow-13",
      "anchor": "r-2026-08-23-gold-4308",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-23-gold-4308",
      "named": "2026-08-23",
      "type": "Map",
      "metal": "Gold",
      "level": "4,308",
      "book": "Turn",
      "call": "Gold holds above the rising 464 swing slow mean, read 4,308 tonight. The swing clock stays bullish while the close stays above the line. Named before the week opened.",
      "kill": "a weekly close back below the 464 swing slow mean, 4,308 at naming.",
      "result": "The kill never fired. The weekly close printed 4,453.670 with the 464 swing slow mean read at 4,308.037 tonight (4,308 at naming), 145.63 above the line. PASS. Close and line read off the TVC chart, market shut, 02:02 CEST 29 August. See note 11.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-08-28",
      "graded_at": "2026-08-28T23:00:00+02:00",
      "written": "2026-08-29",
      "commit": "https://github.com/TheSimplifier7/record/commit/2e4ddff"
    },
    {
      "id": "lrow-14",
      "anchor": "r-2026-08-23-silver-66-66",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-23-silver-66-66",
      "named": "2026-08-23",
      "type": "Map",
      "metal": "Silver",
      "level": "66.66",
      "book": "Turn",
      "call": "Silver holds above the 464 swing slow mean it closed through on 21 August, read 66.66 tonight. The door is now the floor and the swing clock stays bullish while the close stays above the line. Named before the week opened.",
      "kill": "a weekly close back below the 464 swing slow mean, 66.66 at naming.",
      "result": "The kill fired. The weekly close printed 66.3700 with the 464 swing slow mean read at 66.6600 tonight (66.66 at naming), 0.29 below the line. The door that became the floor on 21 August gave way at its first weekly test. MISS. The row stays, and the swing clock in silver is no longer bullish until a weekly close re-earns the line; the standing entry records the end under note 02, and the restatement is logged in note 11. Close and line read off the TVC chart, market shut, 02:04 CEST 29 August.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-08-28",
      "graded_at": "2026-08-28T23:00:00+02:00",
      "written": "2026-08-29",
      "commit": "https://github.com/TheSimplifier7/record/commit/2e4ddff"
    },
    {
      "id": "lrow-15",
      "anchor": "r-2026-08-23-platinum-1790-84",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-23-platinum-1790-84",
      "named": "2026-08-23",
      "type": "Map",
      "metal": "Platinum",
      "level": "1,790.84",
      "book": "Turn",
      "call": "Platinum, on this page for the first time, same method, same rule. Price has climbed back above the 464 swing slow mean, read 1,790.84 tonight. The swing clock is bullish while the close stays above the line. Named before the week opened.",
      "kill": "a weekly close back below the 464 swing slow mean, 1,790.84 at naming.",
      "result": "The kill never fired. The weekly close printed 1,821.80 with the 464 swing slow mean read at 1,790.84 tonight (1,790.84 at naming), 30.96 above the line. PASS. First row, first grade, same method, same rule. Close and line read off the TVC chart, market shut, 29 August. See note 11.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-08-28",
      "graded_at": "2026-08-28T23:00:00+02:00",
      "written": "2026-08-29",
      "commit": "https://github.com/TheSimplifier7/record/commit/2e4ddff"
    },
    {
      "id": "lrow-16",
      "anchor": "r-2026-08-16-gold-4281",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-16-gold-4281",
      "named": "2026-08-16",
      "type": "Map",
      "metal": "Gold",
      "level": "4,281",
      "book": null,
      "call": "Gold holds above the rising 464 swing slow mean, read 4,281 tonight. The swing clock stays bullish while the close stays above the line. Named before the week opened.",
      "kill": "a weekly close back below the 464 swing slow mean, 4,281 at naming.",
      "result": "The kill never fired. The weekly close printed 4,602.660 with the 464 swing slow mean read at 4,308.037 tonight (4,281 at naming), 294.62 above the line. PASS. Close and line read off the TVC chart at 23:00 CEST, 21 August.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-08-21",
      "graded_at": "2026-08-21T23:00:00+02:00",
      "written": "2026-08-21",
      "commit": "https://github.com/TheSimplifier7/record/commit/a0e26cb"
    },
    {
      "id": "lrow-17",
      "anchor": "r-2026-08-16-silver-66-60",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-16-silver-66-60",
      "named": "2026-08-16",
      "type": "Map",
      "metal": "Silver",
      "level": "66.60",
      "book": null,
      "call": "Silver stays below the falling 464 swing slow mean, read 66.60 tonight. The door overhead holds and the turn stays unconfirmed until a close says otherwise. Named before the week opened.",
      "kill": "a weekly close above the 464 swing slow mean, 66.60 at naming.",
      "result": "The close said otherwise. Silver closed the week at 68.9460 through the line, which read 66.6600 tonight (66.60 at naming). The turn is confirmed, and with it the stance published before the week on X, 10 August: silver needs to clear the line, bullish unconfirmed. PASS. The kill as staged on 16 August named a close above the line, which inverted that stance; the page graded the same conditional no test on 14 August when the line was touched and not closed through, so the claim was always the conditional. The wording stands as posted and the inversion is corrected in note 10.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-08-21",
      "graded_at": "2026-08-21T23:00:00+02:00",
      "written": "2026-08-21",
      "commit": "https://github.com/TheSimplifier7/record/commit/a0e26cb"
    },
    {
      "id": "lrow-18",
      "anchor": "r-2026-08-14-gold-4268",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-14-gold-4268",
      "named": "2026-08-14",
      "type": "Verdict",
      "metal": "Gold",
      "level": "4,268",
      "book": null,
      "call": "Gold holds above the floor it reclaimed on the 7 August close. Named before the week opened.",
      "kill": "a weekly close back below 4,268.",
      "result": "The kill never fired and price never came back to test it. Gold closed the week at 4,376.160, more than a hundred dollars clear of the floor.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-08-14",
      "graded_at": "2026-08-14T23:00:00+02:00",
      "written": "2026-08-16",
      "commit": "https://github.com/TheSimplifier7/record/commit/38fdc34"
    },
    {
      "id": "lrow-19",
      "anchor": "r-2026-08-14-silver-66-86",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-14-silver-66-86",
      "named": "2026-08-14",
      "type": "Verdict",
      "metal": "Silver",
      "level": "66.86",
      "book": null,
      "call": "Silver clears the door overhead on a weekly close or the turn stays unconfirmed. Named before the week opened.",
      "kill": "a weekly close above 66.86.",
      "result": "Silver climbed to the line and touched it during the week. It never closed through. The week printed 64.7080, 2.15 underneath. A wick through a level is weather and only the close is tape, so the door was never reached. NEVER REACHED. Nothing is banked on either side.",
      "grade": "notest",
      "grade_word": "never reached",
      "pre_standard": false,
      "graded_on_close": "2026-08-14",
      "graded_at": "2026-08-14T23:00:00+02:00",
      "written": "2026-08-16",
      "commit": "https://github.com/TheSimplifier7/record/commit/38fdc34"
    },
    {
      "id": "lrow-20",
      "anchor": "r-2026-08-07-gold-4098-275",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-07-gold-4098-275",
      "named": "2026-08-07",
      "type": "Verdict",
      "metal": "Gold",
      "level": "4,098.275",
      "book": null,
      "call": "Ceiling, standing since 21 July. The lockout holds until the close says otherwise.",
      "kill": "a weekly close back above 4,098.275 grades the break a trap and puts gold back inside the house.",
      "result": "The kill fired. Gold closed the week at 4,342. The ceiling is gone. WRONG. The prior weekly grades for the same level remain on the page as the record of those closes; this row grades the final resolution of the standing claim.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-08-07",
      "graded_at": "2026-08-07T23:00:00+02:00",
      "written": "2026-08-08",
      "commit": "https://github.com/TheSimplifier7/record/commit/5d671f6"
    },
    {
      "id": "lrow-21",
      "anchor": "r-2026-08-07-silver-61-02",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-07-silver-61-02",
      "named": "2026-08-07",
      "type": "Verdict",
      "metal": "Silver",
      "level": "61.02",
      "book": null,
      "call": "The reclaim signal named 7 July and the ceiling claims that followed it.",
      "kill": "a weekly close back above 61.022 grades the break a trap.",
      "result": "The kill fired. Silver closed the week at 63.45. One close, and the reclaim signal and every restatement of the same level are graded the same way under the 7 August ruling. WRONG. See note 08.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-08-07",
      "graded_at": "2026-08-07T23:00:00+02:00",
      "written": "2026-08-08",
      "commit": "https://github.com/TheSimplifier7/record/commit/5d671f6"
    },
    {
      "id": "lrow-22",
      "anchor": "r-2026-08-07-gold-3942-430",
      "url": "https://thesimplifier7.github.io/record/#r-2026-08-07-gold-3942-430",
      "named": "2026-08-07",
      "type": "Verdict",
      "metal": "Gold",
      "level": "3,942.430",
      "book": null,
      "call": "The June floor beneath, the first thing that would give way.",
      "kill": "a weekly close beneath 3,942.430 and the last floor under this move is gone.",
      "result": "Price went the other way. The weekly close printed 4,342. NEVER REACHED. Untested lines are not wins.",
      "grade": "notest",
      "grade_word": "never reached",
      "pre_standard": false,
      "graded_on_close": "2026-08-07",
      "graded_at": "2026-08-07T23:00:00+02:00",
      "written": "2026-08-08",
      "commit": "https://github.com/TheSimplifier7/record/commit/5d671f6"
    },
    {
      "id": "lrow-23",
      "anchor": "r-2026-07-27-gold-4098",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-27-gold-4098",
      "named": "2026-07-27",
      "type": "Map",
      "metal": "Gold",
      "level": "4,098",
      "book": null,
      "call": "The door overhead stands and the lockout holds until a close says otherwise.",
      "kill": "a weekly close back above 4,098.275 grades the break a trap and puts gold back inside the house.",
      "result": "A standing claim, and it held, and it was tested harder than this page first said. Gold traded above the door on four of the five days that week and closed above it on the Thursday. It could not do it again at the weekly close. The week printed 4,041.18, fifty-seven dollars under, and the kill never fired. Third week of the lockout, and price finished the week further from the line than it started it. Corrected 3 August 2026, see note 07. The claim continued and was finally resolved on 7 August, see the Verdict row of that date.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-31",
      "graded_at": "2026-07-31T23:00:00+02:00",
      "written": "2026-07-31",
      "commit": "https://github.com/TheSimplifier7/record/commit/c03276a"
    },
    {
      "id": "lrow-24",
      "anchor": "r-2026-07-27-gold-3942",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-27-gold-3942",
      "named": "2026-07-27",
      "type": "Map",
      "metal": "Gold",
      "level": "3,942",
      "book": null,
      "call": "The June floor beneath is the first thing that gives way.",
      "kill": "a weekly close beneath 3,942.430 and the last floor under this move is gone.",
      "result": "Never approached, and this row first overstated by how much. The ninety-nine dollars published here was the distance from the closing price; the week's low came roughly half that distance from the line. The weekly close printed 4,041.18. Stated as no test, counted for neither side. Under the shape rule this could be argued a standing pass; the harsher grade is taken and the claim is restated for next week rather than banked. See note 05.",
      "grade": "notest",
      "grade_word": "never reached",
      "pre_standard": false,
      "graded_on_close": "2026-07-31",
      "graded_at": "2026-07-31T23:00:00+02:00",
      "written": "2026-07-31",
      "commit": "https://github.com/TheSimplifier7/record/commit/c03276a"
    },
    {
      "id": "lrow-25",
      "anchor": "r-2026-07-27-silver-61-02",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-27-silver-61-02",
      "named": "2026-07-27",
      "type": "Map",
      "metal": "Silver",
      "level": "61.02",
      "book": null,
      "call": "The ceiling stands and the lockout holds until the close says otherwise.",
      "kill": "a weekly close back above 61.022 grades the break a trap and puts the recovery on two legs.",
      "result": "A standing claim, and it held. Silver came within about one and a half percent of the line on the Monday, and the Thursday close was about three percent under. The weekly close printed 57.564, nearly six percent underneath, measured against the line. The kill never fired. Fourth week of the lockout, and the gap to gold's own line is four times as wide. Corrected 3 August 2026, see note 07.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-31",
      "graded_at": "2026-07-31T23:00:00+02:00",
      "written": "2026-07-31",
      "commit": "https://github.com/TheSimplifier7/record/commit/c03276a"
    },
    {
      "id": "lrow-26",
      "anchor": "r-2026-07-21-gold-4098",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-21-gold-4098",
      "named": "2026-07-21",
      "type": "Map",
      "metal": "Gold",
      "level": "4,098",
      "book": null,
      "call": "The failed floor at 4,098.275 works overhead; the second break stands.",
      "kill": "a weekly close back above 4,098.275 grades the break a trap and puts gold back inside the house.",
      "result": "Gold walked through the doorway in daylight midweek and could not stay. The weekly close printed 4,052.60, forty-six dollars beneath the door. The kill never fired; the lockout held its second week.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-24",
      "graded_at": "2026-07-24T23:00:00+02:00",
      "written": "2026-07-25",
      "commit": "https://github.com/TheSimplifier7/record/commit/3d124fa"
    },
    {
      "id": "lrow-27",
      "anchor": "r-2026-07-21-gold-3810",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-21-gold-3810",
      "named": "2026-07-21",
      "type": "Map",
      "metal": "Gold",
      "level": "3,810",
      "book": null,
      "call": "The catch at 3,810, the weekly slow mean rising since May 2019, is the room's floor.",
      "kill": "a weekly close beneath 3,810 and this claim dies, the bull's oldest line under review.",
      "result": "Never approached this week; untested lines are not wins and this one is not graded. The claim is restated and stays open, the catch rising toward the room, graded week by week at the close. The kill was published as a fixed number against a line that moves; from 1 August it is restated with the line named and the level dated, and the original wording stands as posted. Resolved 21 August: five weekly closes later price has gone the other way, 4,602.660 tonight, and the line was never approached. NEVER REACHED. Untested lines are not wins, the standard applied to 3,942.430 on 7 August. The floor itself lives on as the epoch line inside the year-to-year standing call. See note 10.",
      "grade": "notest",
      "grade_word": "never reached",
      "pre_standard": false,
      "graded_on_close": "2026-07-24",
      "graded_at": "2026-07-24T23:00:00+02:00",
      "written": "2026-07-25",
      "commit": "https://github.com/TheSimplifier7/record/commit/3d124fa"
    },
    {
      "id": "lrow-28",
      "anchor": "r-2026-07-21-silver-61-02",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-21-silver-61-02",
      "named": "2026-07-21",
      "type": "Map",
      "metal": "Silver",
      "level": "61.02",
      "book": null,
      "call": "61.022 remains the ceiling; the lockout stands until the close says otherwise.",
      "kill": "a weekly close back above 61.022 grades the break a trap and puts the recovery on two legs.",
      "result": "Silver walked back to its own door and was turned away. The weekly close printed 58.181, below the ceiling. The kill never fired; the lockout held its third week.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-24",
      "graded_at": "2026-07-24T23:00:00+02:00",
      "written": "2026-07-25",
      "commit": "https://github.com/TheSimplifier7/record/commit/3d124fa"
    },
    {
      "id": "lrow-29",
      "anchor": "r-2026-07-12-gold-4098",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-12-gold-4098",
      "named": "2026-07-12",
      "type": "Map",
      "metal": "Gold",
      "level": "4,098",
      "book": null,
      "call": "4,098 remains the floor after Friday's reclaim.",
      "kill": "a weekly close beneath the floor, and the reclaim is itself reclassified.",
      "result": "The kill fired. Gold closed the week at 4,018.44, eighty dollars beneath the floor. The reclaim is reclassified: a bounce inside a falling market, not a repair. A miss, and it stays. First graded a pass on 17 July with the kill quoted backwards; corrected 18 July, see note 03.",
      "grade": "miss",
      "grade_word": "wrong",
      "pre_standard": false,
      "graded_on_close": "2026-07-17",
      "graded_at": "2026-07-17T23:00:00+02:00",
      "written": "2026-07-18",
      "commit": "https://github.com/TheSimplifier7/record/commit/a500967"
    },
    {
      "id": "lrow-30",
      "anchor": "r-2026-07-12-silver-61-02",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-12-silver-61-02",
      "named": "2026-07-12",
      "type": "Map",
      "metal": "Silver",
      "level": "61.02",
      "book": null,
      "call": "Locked out of its own house. The floor broken 7 July at 61.02 works the opposite job now, the ceiling, with a falling mean sixty cents above it.",
      "kill": "a weekly close back above 61.02 grades the break a trap.",
      "result": "No knock. Silver closed 55.911, five dollars below the ceiling, and never approached it all week. The break stands and the lockout held.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-17",
      "graded_at": "2026-07-17T23:00:00+02:00",
      "written": "2026-07-18",
      "commit": "https://github.com/TheSimplifier7/record/commit/a500967"
    },
    {
      "id": "lrow-31",
      "anchor": "r-2026-07-12-gold-4305-32",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-12-gold-4305-32",
      "named": "2026-07-12",
      "type": "Map",
      "metal": "Gold",
      "level": "4,305.32",
      "book": null,
      "call": "No repair claimed below the first stair.",
      "kill": "a weekly close above 4,305.32 is the first stair won back; until then, standing is all this is.",
      "result": "The stair was never reached, stated as no test, untested lines are not wins. The standing itself failed at the floor instead. Ungraded in the 17 July Verdict, entered 18 July with the correction.",
      "grade": "notest",
      "grade_word": "never reached",
      "pre_standard": false,
      "graded_on_close": "2026-07-17",
      "graded_at": "2026-07-17T23:00:00+02:00",
      "written": "2026-07-18",
      "commit": "https://github.com/TheSimplifier7/record/commit/a500967"
    },
    {
      "id": "lrow-32",
      "anchor": "r-2026-07-08-gold-4098",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-08-gold-4098",
      "named": "2026-07-08",
      "type": "Level Watch",
      "metal": "Gold",
      "level": "4,098",
      "book": null,
      "call": "Gold followed silver through its floor, closing 4,074, one night behind the weaker twin.",
      "kill": "written into the post before the close, a weekly close back above 4,098 grades the break a trap.",
      "result": "The reclaim printed. Friday's weekly close came in at 4,120.38, back above the line, and the break graded a trap at the first close that could judge it. Both branches were written before the candle closed; the tape picked this one.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-10",
      "graded_at": "2026-07-10T23:00:00+02:00",
      "written": "2026-07-11",
      "commit": "https://github.com/TheSimplifier7/record/commit/093f6ae"
    },
    {
      "id": "lrow-33",
      "anchor": "r-2026-07-07-silver-61-02",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-07-silver-61-02",
      "named": "2026-07-07",
      "type": "Level Watch",
      "metal": "Silver",
      "level": "61.02",
      "book": null,
      "call": "Silver closed below its March floor at 59.95, the first floor of this correction to go.",
      "kill": "named with the break, a weekly close back above 61.02 grades it a trap.",
      "result": "No reclaim on the first close. The weekly close printed 59.83, below the line. The break stood and the old floor became the ceiling. The reclaim signal stayed live and was graded week by week at the close. This is silver's first named level with a published kill and the anchor its later standing claims are measured from. Resolved on the 7 August close; see the Verdict row of that date and note 08. Graded 21 August for its own week, the week the break stood, under the standard applied to the gold Level Watch of 8 July; the final resolution of the level is the 7 August Verdict row, WRONG. See note 10.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-10",
      "graded_at": "2026-07-10T23:00:00+02:00",
      "written": "2026-07-11",
      "commit": "https://github.com/TheSimplifier7/record/commit/093f6ae"
    },
    {
      "id": "lrow-34",
      "anchor": "r-2026-07-05-gold-4098-4307",
      "url": "https://thesimplifier7.github.io/record/#r-2026-07-05-gold-4098-4307",
      "named": "2026-07-05",
      "type": "Map",
      "metal": "Gold",
      "level": "4,098 / 4,307",
      "book": null,
      "call": "4,098 is a floor again after Friday's vote. The gate at 4,307 decides what the bounce becomes. Nothing has turned below the 4,500 turn.",
      "kill": "a close back beneath 4,098 and this read dies.",
      "result": "The week threw the hardest test available: a midweek close below the floor. The weekly close answered at 4,120.38, above the line. The floor call survived its own published kill. The gate was never reached, stated as no test. The turn stays open.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": false,
      "graded_on_close": "2026-07-10",
      "graded_at": "2026-07-10T23:00:00+02:00",
      "written": "2026-07-11",
      "commit": "https://github.com/TheSimplifier7/record/commit/093f6ae"
    },
    {
      "id": "lrow-35",
      "anchor": "r-2026-06-29-gold-4500-3809",
      "url": "https://thesimplifier7.github.io/record/#r-2026-06-29-gold-4500-3809",
      "named": "2026-06-29",
      "type": "Map",
      "metal": "Gold",
      "level": "4,500 / 3,809",
      "book": null,
      "call": "Falling swing. The turn needs a close above the 4,500 mean; the catch below is the weekly slow mean at 3,809.",
      "kill": null,
      "result": "The turn at 4,500 was never tested. The catch at 3,809 was never reached. Both stated as no test, untested lines are not wins. The one claim the tape answered held: no close above 4,500, every green day stayed a bounce. Gold reclaimed the 4,098 floor on the vote, closing 4,174.76.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": true,
      "graded_on_close": "2026-07-03",
      "graded_at": "2026-07-03T23:00:00+02:00",
      "written": "2026-07-04",
      "commit": "https://github.com/TheSimplifier7/record/commit/7e9b339"
    },
    {
      "id": "lrow-36",
      "anchor": "r-2026-06-22-gold-4098-4633",
      "url": "https://thesimplifier7.github.io/record/#r-2026-06-22-gold-4098-4633",
      "named": "2026-06-22",
      "type": "Map",
      "metal": "Gold",
      "level": "4,098 / 4,633",
      "book": null,
      "call": "Caught between a floor to hold at 4,098 and a turn at 4,633. Until a close clears the turn, every green day is a bounce in a falling market.",
      "kill": null,
      "result": "The turn at 4,633 was never reached, nor the gate at 4,307. Every bounce failed. Gold lost the floor and closed 4,088.97. The gradeable sentence on this row is the bounce claim, and it is what the grade is given on.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": true,
      "graded_on_close": null,
      "graded_at": null,
      "written": "2026-06-30",
      "commit": null
    },
    {
      "id": "lrow-37",
      "anchor": "r-2026-06-15-gold-4307",
      "url": "https://thesimplifier7.github.io/record/#r-2026-06-15-gold-4307",
      "named": "2026-06-15",
      "type": "Map",
      "metal": "Gold",
      "level": "4,307",
      "book": null,
      "call": "Reclaimed the floor at 4,098 and the gate at 4,307. The gate is the line between a bounce and a repair: hold it and it repairs, lose it and it was a bounce.",
      "kill": null,
      "result": "Gold gave 4,307 straight back. The 4,209 first published here was Thursday's close and not the week's; the week closed roughly fifty dollars lower. The repair stalled exactly there, the floor held. It was a bounce. Corrected 3 August 2026, see note 07. The grade stands.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": true,
      "graded_on_close": null,
      "graded_at": null,
      "written": "2026-06-30",
      "commit": null
    },
    {
      "id": "lrow-38",
      "anchor": "r-2026-06-11-gold-4098",
      "url": "https://thesimplifier7.github.io/record/#r-2026-06-11-gold-4098",
      "named": "2026-06-11",
      "type": "Verdict",
      "metal": "Gold",
      "level": "4,098",
      "book": null,
      "call": "After the staircase broke, one signal named in advance: a close back above 4,098 is the first sign of life. Not a bottom, a sign.",
      "kill": null,
      "result": "It printed that night. Gold touched the 4,020s and closed back above 4,098, then bounced near 4,210. The signal fired; the bounce later failed, as called.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": true,
      "graded_on_close": null,
      "graded_at": null,
      "written": "2026-06-30",
      "commit": null
    },
    {
      "id": "lrow-39",
      "anchor": "r-2026-06-07-gold-4308",
      "url": "https://thesimplifier7.github.io/record/#r-2026-06-07-gold-4308",
      "named": "2026-06-07",
      "type": "Read",
      "metal": "Gold",
      "level": "4,308",
      "book": null,
      "call": "Stated bearish with the exit named: gold sits on the last shelf at 4,308, open air to 4,100 beneath.",
      "kill": "only a weekly close back above the line turns it bullish.",
      "result": "Gold broke the shelf and fell through the open air into the 4,020s. The bearish call held. This is the anchor the week to week standing call is measured from, in gold. It named no silver level.",
      "grade": "pass",
      "grade_word": "held",
      "pre_standard": true,
      "graded_on_close": null,
      "graded_at": null,
      "written": "2026-06-30",
      "commit": null
    }
  ],
  "standing": [
    {
      "clock": "The direction, week to week",
      "position": "Bullish in gold, silver and platinum",
      "stated": "2026-08-21",
      "restated": "2026-08-28",
      "as_of": "2026-09-11",
      "read": "close",
      "status": "retired",
      "kill": "the epoch midline crossing the slow mean reversion line, per metal, ruled 29 August 2026. Position against the line is the confirmation and never the trigger. No price close ends this call, which is why three closes beneath the line in silver, 28 August, 4 September and 11 September, and one each in gold and platinum on 11 September, have not ended it. The line reads: gold 4,357.307, silver 66.6878, platinum 1,797.35, all at the 11 September close. The weekly rows in the ledger below carry their own kills, published before each week, and those ARE price closes",
      "board": [
        {
          "metal": "Gold",
          "line": "4,357.307",
          "close": "4,347.760",
          "distance": "-9.547",
          "pct": "-0.22%",
          "since": "2026-08-07",
          "state": "live"
        },
        {
          "metal": "Silver",
          "line": "66.6878",
          "close": "64.4810",
          "distance": "-2.2068",
          "pct": "-3.31%",
          "since": "2026-08-21",
          "state": "live"
        },
        {
          "metal": "Platinum",
          "line": "1,797.35",
          "close": "1,791.50",
          "distance": "-5.85",
          "pct": "-0.33%",
          "since": "2026-08-19",
          "state": "live"
        }
      ],
      "marks": [
        {
          "name": "Gold",
          "move": "+0.13%",
          "from": "since the 7 Aug close, down from +2.04% last week · 9.547 under the line at 4,357.307, the first weekly close under it since the clock turned"
        },
        {
          "name": "Silver",
          "move": "-3.31%",
          "from": "2.2068 under the line at 66.6878 · third weekly close under it · bullish since 21 Aug, the claim intact under the 29 Aug ruling"
        },
        {
          "name": "Platinum",
          "move": "-0.33%",
          "from": "5.85 under the line at 1,797.35 · first weekly close under it · bullish since the 19 Aug close, the claim intact under the 29 Aug ruling"
        }
      ]
    },
    {
      "clock": "The direction, year to year",
      "position": "Up, on the secular clock, until the weekly cross",
      "stated": "2026-06-07",
      "restated": null,
      "as_of": "2026-08-01",
      "read": "read",
      "status": "live",
      "kill": "the weekly fast mean crossing down through the weekly slow mean, confirmed by price below the epoch line. Gold 4,721 over 3,834, silver 77.47 over 46.36 on 1 Aug",
      "board": null,
      "marks": [
        {
          "name": "Gold",
          "move": "4,721 over 3,834",
          "from": "weekly fast mean over weekly slow mean"
        },
        {
          "name": "Silver",
          "move": "77.47 over 46.36",
          "from": "weekly fast mean over weekly slow mean"
        }
      ]
    }
  ],
  "notes": [
    {
      "n": "01",
      "kind": "standard",
      "title": "Standard tightened 3 July 2026.",
      "body": "Every claim now ships with the condition that would grade it a miss; untested levels are stated as no test rather than counted as wins; partial grades count against the record in any ratio quoted. Earlier entries were graded under the prior convention, are marked pre-standard in the ledger, and are preserved unaltered. No grade is ever restamped to fit a rulebook written after it. Rows carrying no published kill are marked as such."
    },
    {
      "n": "02",
      "kind": "standard",
      "title": "Standing calls, added 31 July 2026.",
      "body": "The directional position runs on a longer clock than a level and resolves on a different question, so it is stated separately and graded separately. Standing calls never enter the tally: averaging a call that runs for months with one that resolves on a Friday would flatter both. Each carries the close that ends it, and every figure beside it carries the date it is measured from. When one ends it stays, with the date and what ended it."
    },
    {
      "n": "03",
      "kind": "correction",
      "title": "Correction, 18 July 2026.",
      "body": "The Verdict posted 17 July graded the 12 July gold floor claim a pass. The published kill, a weekly close beneath 4,098, had in fact fired at 4,018.44. The correct grade is a miss and the ledger carries it. A third claim from the same map, the stair at 4,305.32, went ungraded in that Verdict and is stated here as no test. The erroneous grades, this correction, and the grading process failure behind them are all logged and all remain."
    },
    {
      "n": "04",
      "kind": "correction",
      "title": "Correction, 30 July 2026.",
      "body": "The Updated field on this page was rendered from the visitor's own device clock rather than from the file. While that line was live, the page told every reader it had been updated on the day they opened it, whatever the file's actual state. No call, no grade and no figure in the ledger was affected. The field is now set by hand against the commit. The fault is logged here rather than fixed quietly, because a page that claims nothing is hidden does not get to hide this."
    },
    {
      "n": "05",
      "kind": "standard",
      "title": "Claim shapes, stated 31 July 2026.",
      "body": "A claim that price stays on one side of a line is graded on whether the close crossed it, whether or not the line was approached. A claim that requires price to reach a level is stated as no test when it does not. Every standing pass now carries a note saying how close price came, so a grade earned at arm's length cannot be read as one earned under pressure. Where a shape is arguable, the harsher grade is taken."
    },
    {
      "n": "06",
      "kind": "correction",
      "title": "Correction, 1 August 2026. Two kills renamed on the standing calls.",
      "body": "Both standing calls named the wrong line. Neither error changed a grade and both are corrected here rather than quietly. First, the week to week call published its kill as a weekly close above “the falling weekly fast mean, 4,283.” The level was right and has not changed: 4,283 was the 464 swing slow mean on 31 July and the same falling line reads 4,277 today. The name was wrong. The weekly fast mean is a different line and stands at 4,721 in gold, so a reader who opened a weekly chart to check this call found a number 438 dollars away from the one published here and could not verify a claim this page exists to make verifiable. Second, the year to year call published its kill as a weekly close beneath the slow mean. The grading standard has said since 3 July that a secular turn is dated by the weekly fast mean crossing the slow mean and never by a price close, because price can sit below the slow mean for weeks with the trend intact. The card contradicted the standard and the card was wrong. This second correction makes the call harder to end, not easier, which is the direction that flatters this desk and the direction it is required to distrust most. It is made anyway, because the rule it now matches was published on 3 July and has not moved. Silver's own swing kill, which had never been published on either card, is stated for the first time at 67.54."
    },
    {
      "n": "07",
      "kind": "correction",
      "title": "Correction, 3 August 2026. Four rows described the tape wrongly, and every error flattered this desk.",
      "body": "Four rows on this page misstated how close price came to a line. The gold row for 27 July said the line itself was not touched. Gold traded above 4,098.275 on four of the five days that week and closed above it on the Thursday. A nearest approach published as 4,079.46 cannot be reproduced and has been removed rather than restated. The floor row said price held ninety-nine dollars above 3,942.430 all week; ninety-nine was the distance from the closing price, and the week's low came roughly half that distance from the line. The silver row said price never came within five percent of 61.022 at any point in the week; the high came within about one and a half percent and the Thursday close was about three percent under. The 15 June row published 4,209 as the week's close; that was Thursday's close and the week closed roughly fifty dollars lower. No grade changes. Not one. Every error above made this record look better than the tape did, which is the direction this desk is required to distrust most, and note 05 was written on 31 July for the express purpose of preventing exactly this. The gate existed and it was not run, which is the worse failure of the two. Separately, and for the first time on any surface this desk publishes: every close quoted on this page is read from TradingView TVC, gold and silver spot against the dollar, at 23:00 CET. Until today no page or post named a source, and a reader checking a close against a different provider would have found a small difference with no way to know why. Two figures briefly published here on 3 August, 4,042.67 and 57.6569, came from a second provider during this correction and were themselves inconsistent with the source named above; they are withdrawn and the original readings stand."
    },
    {
      "n": "08",
      "kind": "correction",
      "title": "Resolution, 7 August 2026. Two wrongs, one untested, and the swing clock in gold turns.",
      "body": "Friday 7 August close (TradingView TVC, 23:00 CET): gold 4,342, silver 63.45. The standing ceiling claims at gold 4,098.275 and silver 61.02 both had their kills fire; both are graded wrong and stay on the page. Gold 3,942.430 was never reached and is stated as no test. Gold has been bearish since 13 March. That structure ended on the 7 August close above the 464 swing slow mean at 4,268. From that close the swing clock in gold turns bullish; a weekly close back below 4,268 now kills the new call. Silver remains below its own swing line and continues under the prior claim. Graded per metal under the ruling made before the close. Two rulings disclosed: the swing kill that named both metals in one sentence is graded per metal, and silver 61.02 receives its grade under the same standard applied to every gold restatement. Both rulings taken in the direction that costs this desk more. Tally after this close: twelve correct, three wrong, three never tested."
    },
    {
      "n": "09",
      "kind": "correction",
      "title": "Correction, 16 August 2026. A caption misdated its own claim.",
      "body": "The 9 August post restated the standing swing call as though named that night. The call it restated was stated 7 August, and the page carried the correct date throughout. The post is not deleted and its reach, roughly nineteen thousand impressions before this note was published, is stated rather than hidden. The error was the caption's date framing, not the level, the kill, or any grade. Logged the week it was caught, under the same rule as every other correction on this page. This note itself first went live on the morning of 17 August with a drafting blank still inside it; the blank was removed and the note republished the same day, the interval logged rather than tidied."
    },
    {
      "n": "10",
      "kind": "correction",
      "title": "Resolution, 21 August 2026. Two passes, one correction, two aged rows closed, and the swing clock in silver turns.",
      "body": "Friday 21 August close (TradingView TVC, 23:00 CEST): gold 4,602.660, silver 68.9460. The 16 August gold call, holds above the 464 swing slow mean, passes; the line read 4,308.037 tonight and the close sat 294.62 above it. The 16 August silver call passes, with a correction disclosed. The stance was published before the week, on X on 10 August: silver needs to clear the line at 66.86, bullish unconfirmed. The row staged on 16 August carried that conditional in its text and then named a close above the line as its kill, which inverted the stance; the page's own 14 August grade of no test, when the line was touched and not closed through, shows the claim was always the conditional and never that the door holds. Tonight the line read 66.6600 and the close printed 68.9460, through it. The conditional resolved as named: PASS. The kill wording stands as posted and is disclosed here as the error. Disclosed too: this row went live graded WRONG on the first 21 August push and was corrected the same night, on the evidence above, before the next close. From this close the swing clock in silver turns bullish, graded per metal under the 7 August standard, and a weekly close back below 66.6600 now kills the new call. Gold continues, 6.00% above the 7 August close. Two aged rows resolve tonight. Gold 3,810 of 21 July is stated as no test under the standard applied to 3,942.430 on 7 August: five weekly closes, price went the other way, the line never approached; the floor lives on as the epoch line in the year-to-year call. Silver 61.02 of 7 July receives its own week's grade, a pass, under the standard applied to the gold Level Watch of 8 July; its final resolution was already graded WRONG in the 7 August Verdict row. Both rulings taken the way the page already rules, not the way that flatters it. Tally after this close: sixteen correct, three wrong, five never tested, nothing open."
    },
    {
      "n": "11",
      "kind": "correction",
      "title": "Resolution, 28 August 2026. Two held, one broke, platinum's first grade, and a ruling on what ends a swing call.",
      "body": "Friday 28 August close (TradingView TVC, market shut, read 02:02 to 02:04 CEST on 29 August): gold 4,453.670, silver 66.3700, platinum 1,821.80. Three floors were named on Sunday 23 August with the market shut, each published with the weekly close that would kill it. The gold call passes; the 464 swing slow mean read 4,308.037, the same read as 21 August, and the close sat 145.63 above it. The platinum call passes on the first platinum row this page has ever carried; the line read 1,790.84, the same read as the level named on Sunday, and the close sat 30.96 above it, same method, same rule, third metal. The silver call fails; the line read 66.6600 and the close printed 66.3700, 0.29 under it, so the kill fired exactly as it was published on Sunday night. The row grades wrong and it stays, where every wrong row on this page stays. The standing week-to-week call is restated, not ended, and on 29 August it was restated a second time by a ruling that reached it and not the rows. THE SWING CLOCK TURNS ON A CROSS, NOT A PRICE CLOSE: what dates a turn is the epoch midline crossing the slow mean reversion line, and position against the line is the confirmation, never the trigger. That is not a new idea on this page, it is the standard already applied to the secular clock since 1 August under note 06, extended by the founder to the swing clock and read on his own chart on 29 August. Under it, silver's swing claim did not end on the 28 August close. Silver closed 0.29 under its line and the structure never turned, so the standing direction stands in all three metals. THE RULING DOES NOT REACH THE ROW, and the distinction is the reason this note is long. The weekly rows are a different object on a different clock: each names a level before the week and publishes a price close as the kill, and each is graded on exactly what it published. The silver row of 23 August named a weekly close back below 66.66 as its kill, the close printed 66.3700, and that kill fired. The row grades WRONG and it stays wrong. Note 01 has said since 3 July that no grade is ever restamped to fit a rulebook written after it, and note 03 is the precedent where this desk applied that in the direction that cost it a win. A desk that re-reads its own rulebook on a Saturday and finds its Friday loss was never a loss has stopped keeping a record. The tally is unchanged by the ruling: eighteen correct, four wrong. What the ruling does change, and this is the process defect worth naming, is that the Sunday rows have been publishing a kill that does not match the method the desk actually trades. The row's kill was honest, published and fired; it was also the wrong question to ask of a swing claim. From the next Sunday call, a row naming swing direction states the event that truly ends it, and a row naming a price level keeps a price kill and says which of the two it is. Gold's mark is 2.57% above the 7 August close, down from 6.00% last week. The entry keeps the status live and keeps its stated date of 21 August because nothing was newly stated tonight. Two disclosures with it. First, the paragraph under the standing calls told every reader that silver continues under its own line. That sentence was written on 7 August and was true then. It stopped being true on the 21 August close and was not moved with the entry that night, so it stood stale for one week. It is corrected tonight and logged rather than tidied. Second, platinum's close and its line read were entered separately from gold's and silver's, off the founder's own chart, because the capture taken at the close had the chart's own label covered. Platinum's swing clock is dated from the 19 August close, read off the same chart and supplied on 29 August; for one day this page carried platinum with a graded row and no dated clock, and said so rather than inventing a start. The grade was never in doubt; the two prints were, and they were read before this note published. Tally after this close: eighteen correct, four wrong, five never tested, twenty-seven rows, nothing open."
    },
    {
      "n": "12",
      "kind": "correction",
      "title": "Correction, 29 August 2026. The year-to-year card called a chart read a weekly close.",
      "body": "The standing year-to-year call has carried the line \"At the close of 1 Aug 2026\" under its figures since it was last refreshed. 1 August 2026 was a Saturday, and there is no weekly close on a Saturday. The figures beneath that line are a chart read taken that day, which the entry's own working note has always said: it records that the chart of record was read on 1 August for the date the secular bull began. The date was right and the verb was wrong, and the card now reads \"At the 1 Aug 2026 read\". No figure moved. The two returns on that card, gold since 28 May 2019 and silver since July 2020, are the same figures they were, unrecomputed, and the rule against touching them without reframing the chart to the marker still stands. Why it is logged rather than quietly fixed: it survived four weeks because it was the only date on that card and nothing sat beside it to disagree with. On 29 August a second graphic arrived on the same screen, stamped at a real weekly close, and two graphics of different vintages three hundred pixels apart cannot each say \"close\" when only one of them means it. That is note 09's defect in a new place, a surface dating a claim it did not make that day, and this page logs those rather than tidying them. Every date on both standing cards is now derived from one field per card and printed with the verb that field names, so the head, the graphic and the marks can no longer drift apart."
    },
    {
      "n": "13",
      "kind": "correction",
      "title": "Disclosure, 29 August 2026. Ten of the eighteen correct rows sit on two levels, and both of those levels broke.",
      "body": "This page grades one row per week, and the founder's ruling of 16 August gives the Sunday call a row every week even when the level has not moved. The cost of that rule is arithmetic: a level that holds for eight weeks produces eight passes. Until tonight the reader was never told, and the reader is the only person the count is for. Gold 4,098 was named in eight rows between 11 June and 7 August and six of them are graded correct. Silver 61.02 was named in five rows between 7 July and 7 August and four of them are graded correct. That is thirteen of the twenty-seven rows on this page, ten of the eighteen correct rows, and three of the four wrongs. Both levels broke. Gold's kill fired first on 12 July at 4,018.44, which note 03 corrected from a pass to a miss six days later, and the level was named again, held two more weeks, and went for good on the 7 August close. Silver's kill fired on that same close. Both wrong rows are still here, where every wrong row on this page stays. The clearest illustration is note 10. On 21 August this desk graded the silver 61.02 row of 7 July a pass for its own week, two weeks after the level itself had already failed, and that note says the rulings were taken the way the page already rules and not the way that flatters it. That was true of the ruling. It was not true of the arithmetic underneath it. The correct count went up on a level that had already broken, and nothing on the page said so. The desk knew before the reader did, which is the part worth logging. The working comment written into this file on 16 August, the day the rule was made, states that a level holding for ten weeks produces ten passes, calls it the flattery the standing block warns about, and says it is accepted here on purpose. Accepted on purpose is defensible, and the reason given still holds. Accepted on purpose and visible only to whoever opens the source file is a different thing and a worse one. Note 02 has refused since 31 July to average a standing call into the tally, on the ground that a claim running for months and a claim resolving on a Friday cannot share a denominator. That is the same fault. It was sitting inside the weekly rows the whole time, unreached by the rule written to catch it. From tonight the tally is published twice, and the count that makes this desk look worse is printed with it rather than after it. Read by week it is eighteen correct, four wrong, five never tested, twenty-seven rows, unchanged. Read by level it is fourteen levels named, seven ended held, three ended broken, four never tested, with four of those levels carried across more than one week, producing twelve of the eighteen correct rows, and two of those four finally broken. Both figures are derived from the ledger on every load rather than typed, so neither can drift from the rows it counts. The grouping rule is stated so a stranger can reproduce the number without asking. A level family is the metal plus the first number in the level field, whole units for gold and platinum, two decimals for silver. It is deliberately the generous rule. It counts gold 4,305, 4,307 and 4,308 as three separate levels when they are plainly one line moving, so the concentration reported is the smallest defensible figure rather than the largest. A stricter grouping makes this page look worse, not better, and anyone who applies one will find the desk said so first. Why both counts are published and not only the harsher one. Each of those thirteen rows was staged before its week with a kill that could have fired on any Friday, and on three of those Fridays it did. A week that was genuinely at risk and survived is a week that survived, and striking it would be an edit of exactly the kind this page refuses everywhere else. But eight surviving weeks are not eight ideas. They are one idea, tested eight times, that was finally wrong. A reader shown only the weekly number is being told this desk was right eighteen times, when what it did was be right about fewer things for longer. Neither figure alone is the truth, so both stand, each beside its base. No grade changes. Not one. No row is moved, renamed or removed, and the 16 August rule stands, because a weekly claim that resolves at a named close is the thing this page exists to record and refusing to record it would leave most weeks blank. What changes is what a reader can see without opening the file. This is published tonight rather than whenever someone else assembles it, which on this ledger takes about ten minutes with a sort and a pencil, because a page that keeps its wrong calls up forever does not get to keep the flattering arithmetic to itself."
    },
    {
      "n": "14",
      "kind": "standard",
      "title": "Copy for the reader, 3 September 2026. The fold, the legend, the standing sentence and the seat door.",
      "body": "Four pieces of copy on this page were rewritten for a stranger arriving from a post rather than for the desk that keeps it. The opening line now names the reader's own experience before it states the rule. A legend under the tally defines held, wrong, never reached and kill in plain words. One plain sentence sits above the standing cards. The seat door says what a person can do at the end of twelve weeks rather than how the programme is structured, and moves the structure to the terms page. A next verdict line is derived from the Updated stamp. The tally note is split into its two counts, by week and by level, side by side. A strip of squares draws every row on the record oldest to newest in its grade colour, each a link to its row. The ledger carries a divider per naming Sunday with that week's grades, and an opened row that states its graded close draws the close against the level as named; a row that does not state a close draws nothing. The standing cards carry a plain face: one row per metal with the close drawn against its own line, a state chip, and what ends the call in plain words; the desk graphic, the clocks and the marks are unchanged behind the fold. The notes fold like the rows. A link underline that had been drawing a line through the next row of text is corrected. No call, no grade, no figure, no kill and no date in the ledger was touched. The replaced copy is kept in the source beside the new, which is where this file keeps everything it replaces."
    },
    {
      "n": "15",
      "kind": "standard",
      "title": "Provenance per row, the ladder, and the record as data, 4 September 2026.",
      "body": "Every graded row now names the close it was graded on, the date it was written to this file, and the commit that wrote it, linked to the repository so the timestamp a reader checks is GitHub's and not this desk's. Rows named before the standard of 3 July were carried in the record's first upload of 30 June and say so instead. A ladder above the standing calls draws every level ever named on the price axis, per metal, with how many weeks it was named and how the last week that carried it ended, derived from the rows. The misses are gathered above the ledger with the first words of their own results. The ledger is also published as record.json beside this page, so anyone can check the arithmetic without trusting the page that states it."
    },
    {
      "n": "16",
      "kind": "standard",
      "title": "Open rows, 4 September 2026. Named before, written before, graded in place, and one metal deliberately without a number.",
      "body": "For the first time this page carries rows that are named and not yet graded. Two floors went out on the map of 30 August, gold 4,350 and platinum 1,797.40, and both are written into this file tonight, before the close that will decide them. That is the whole reason for writing them now rather than after. Every row above this one was written to the record in the hours following the close that graded it, which is fast and still, strictly, afterwards; a reader had to take this desk's word that the level had been named before the week rather than fitted to it. These two rows do not ask for that word. They sit in a commit whose timestamp belongs to GitHub, hours before the tape exists, and the provenance line under them links to the commits of the day so the reader checks that timestamp somewhere this desk cannot edit. From tonight the standing intention is that every Sunday call is written to this file open, in the week it is named, and graded in place. SILVER CARRIES NO ROW THIS WEEK, and that absence is deliberate rather than an oversight, so it is stated here rather than left as a gap. Silver's floor at 66.66 broke on the 28 August close and is graded wrong above, where it stays. The map published on 30 August did not give silver a replacement number. It gave it a condition to earn one back, a weekly close above 66.80, and a condition is not a call: there is nothing in it that can be wrong on Friday, so there is nothing in it to grade, and a row that cannot lose is not a row this page will carry. Two metals were named and two metals are on the clock. A reader who counts three metals on the standing cards and two rows in the week should read that as the desk declining to name a number it had not earned, and should hold it against the record if silver quietly reappears with a number that was never published on a Sunday. THE OPEN STATE, on every surface. The tally gains a fourth figure, counted and never estimated, and open rows stay out of the resolved base entirely: they are not correct, not wrong, and not never reached, and folding them into any of those three would move a number that has not been earned. The strip draws an open square in its own colour, the legend defines the word in plain language, and the ladder draws an open line in that same colour rather than in the never-reached grey. TWO DEFECTS THE OPEN STATE EXPOSED, both caught before this file was published and both logged anyway. First, the next verdict line, which derives the next grading date from the Updated stamp rather than from the visitor's clock under note 04, was written when the stamp was only ever set after a grade. On that assumption a Friday stamp meant the week had already been ruled and the next verdict was the following Friday. Setting the stamp before a grade, which is what writing an open row requires, turned that assumption into a false statement: the page would have told a reader on Friday evening that the next verdict was a week away while two rows sat open and due at that evening's close. It now names that Friday whenever anything is still open. Second, the ladder colours a band by the last grade its lines received, and the map from grade to colour sent anything that was not a pass or a miss to the never-reached grey. Platinum's 1,790.84 and its 1,797.40 sit within half a percent of each other and therefore cluster into one band, so an open line joining a graded pass would have turned that pass grey on sight. Neither error reached a reader. Both are written down because a page that logs only the faults it is caught at is keeping a different kind of record than this one claims to keep. As always, no call, no grade, no figure, no kill and no date above this note was touched, and every line of copy replaced tonight is kept in the source beside the line that replaced it."
    },
    {
      "n": "17",
      "kind": "correction",
      "title": "Resolution, 4 September 2026. Two floors held, and the metal with no number stayed without one.",
      "body": "Friday 4 September close, read off the TVC chart with the market shut at 03:03 to 03:04 CEST on 5 September: gold 4,430.520, platinum 1,819.60, silver 66.1940. Two floors were named on 30 August with the market shut, each published with the weekly close that would kill it, and both are graded here. Gold's 464 swing slow mean reads 4,350.628 tonight, the same read as the chart label the level was taken from on Sunday, and the close sat 80.52 above the number as published. PASS. Platinum's line reads 1,797.40, the same figure as the level named, and the close sat 22.20 above it. PASS. Second row for platinum on this page, second pass, same method, same rule. SILVER, WHICH IS THE ROW THAT IS NOT HERE. Silver's floor broke on the 28 August close, graded wrong above and staying there. The map of 30 August did not replace that number, it published a condition to earn one back: a weekly close above 66.80, which is the 464 swing slow mean, read 66.8082 that night. Tonight the condition was tested and it was not met. Silver closed 66.1940 with the line reading 66.8082, so the close sat 0.61 under the line and 0.61 under the number as published. Silver therefore carries no row for a second week and gets no level, no direction and no stated position until a weekly close clears that line. This is written out at length for one reason. A metal that goes wrong and then quietly stops appearing in a weekly ledger is indistinguishable, to a reader, from a metal that was dropped because it was embarrassing. Note 16 promised that the absence would be stated every week it holds. This is the first of those statements and there will be one every Friday silver stays under its line. The condition is unchanged, it was published before the week, and it is not moving down to meet the price. THE PROVENANCE ON THESE TWO ROWS CHANGED SHAPE TONIGHT, and the change is the point of note 16. Both rows were written into this file before the close that graded them, and while they were open their provenance line said so and linked to the commits of the day they were written, so a reader could check GitHub's timestamp rather than this desk's word. That line is now replaced with the graded shape, and the shape it replaced is kept in the source beside it, which is where this file keeps everything it replaces. The commit is named by its day rather than by a hash because the hash does not exist until the push that writes this note. Tally after this close: twenty correct, four wrong, five never tested, twenty-nine rows, nothing open. Both grades tonight were passes, which is the direction this desk is required to distrust most, so it is worth stating plainly what they are and are not. Two floors named on a Sunday held through a Friday close. That is two weeks of two claims, not two ideas proven, and the count by level published beside the count by week exists so that a reader can see the difference without being told."
    },
    {
      "n": "18",
      "kind": "standard",
      "title": "Four additions, 5 September 2026. The countdown, the counter, the crowd and the guest ledger.",
      "body": "Four things were added to this page today and none of them touches a grade, a row, a figure or a date in the ledger above. Each is stated here with the rule that governs it. THE COUNTDOWN. An open row now shows how long until the close that grades it, and the next verdict line carries the same clock while anything is open. Note 04 forbids deriving the Updated stamp from the visitor's clock, because that stamp is a claim about this file and a visitor's clock cannot know it. The countdown is a different object. The close that grades a row is a fixed fact derived from the row's own naming date, Friday at 23:00 in Madrid, and the visitor's clock is used only to say how far away that fixed point is. If a visitor's clock is wrong the countdown is wrong and nothing on the record moves. When the close has passed and the row is still open, the line says so instead of counting into negative numbers. THE COUNTER. Until today this page had no idea how many people read it. A record that grades itself to the cent and cannot count its own readers is missing the one number every other decision about it depends on. A privacy-respecting counter is wired in, no cookies, no personal data, open source, and it loads only when its address is filled in, on the same fail-closed rule as the doors. It cannot see, change or influence a single row. GRADED FROM THE CROWD. The sentence at the top of this page has always been about other people's calls: the ones that were wrong and disappeared. From today, now and then, one of them does not disappear. A claim made in public about gold, silver or platinum, quoted exactly with a link to where it was made, is given the kill it was published without, and graded on the same close in the same words as every row above it. The rules are written into the source beside the array and the ones that matter most are these. The claim, never the person: a handle is recorded because a quote without a source is a fabrication, and the person is not mentioned again, in the row or anywhere built from it. The kill is assigned by this desk and the rule for assigning it is stated, because the finding is not that the claim was wrong, it is that the claim was published with no way to be wrong. These rows never enter the tally, for the same reason the standing calls never do under note 02: they are a different object and averaging them in would flatter or damn this desk with somebody else's arithmetic. And when this desk has a wrong row of its own in the same week, that row is named first in any post that carries a crowd grade, because a standard applied from a clean sheet is a stone thrown. THE GUEST LEDGER. Other people's levels, graded here. A stranger submits a metal, a level and the weekly close that would prove it wrong before Sunday's close, and on Friday it is graded on the TVC close by this desk's standard, held, wrong or never reached, and kept whichever way it went. Eight rows a week at most, in the order received, the rest held over and counted rather than dropped. Never in the tally. Never edited after grading. Never given a video. The desk does not comment on a guest's level, it grades it. Why this exists is the reason this page exists: a record kept by one man is a diary, and a record where strangers are graded as hard as the owner is a standard. The only proof that a standard is neutral is a stranger graded wrong on it, in public, beside the owner's own wrong rows, and this section is where that proof accumulates or fails to. Both new sections render nothing while their arrays are empty, and empty is an acceptable state. This page has never padded a section to look busy and it will not start with these."
    },
    {
      "n": "19",
      "kind": "correction",
      "title": "Correction, 6 September 2026. The standing card and the ladder's close marker stood one week stale.",
      "body": "The upload of 5 September graded the two rows of 30 August on the 4 September close and left the standing swing-clock entry, its board and its marks at the 28 August close. The card said so in its own date, so it did not misstate a figure, but for a day and a half the page carried rows graded on 4 September above a standing card and a ladder marker that read 28 August, and a reader comparing the two would have been right to ask which week this page was on. Refreshed tonight, before the two open rows of 6 September went up: gold closed 4,430.520 against its line at 4,350.628, 79.892 above, and its mark since the 7 August close is now +2.04%, down from +2.57%; silver closed 66.1940 against 66.8082, 0.6142 under, its second weekly close under the line, and under the 29 August ruling the claim stands because the event that ends it is the cross and not the close; platinum closed 1,819.60 against 1,797.40, 22.20 above. Every replaced figure is kept in the source beside the new one. The Friday routine now carries the standing refresh as a named step so the card and the rows cannot be graded on different weeks again. Nothing in the ledger, the tally or any grade changes."
    },
    {
      "n": "20",
      "kind": "standard",
      "title": "The dark ground, 6 September 2026. The page is dark, and paper stays behind a switch.",
      "body": "This page launched on paper, the founder's chart canvas at the time, and paper was the wrong ground for two thousand words read on a phone at midnight: hard to read, and out of step with the chart the videos are cut from, which went to a near-black canvas today for the same reason. So the ground is dark for everyone from tonight, the same #141210 as the video chart, the tally card and the embed, and it is a ruling, not a preference detected from a device. The launch palette is not deleted: a Paper switch in the masthead restores it on the device that presses it, and that choice never leaves the browser. The dark ground is the same ink scale inverted and the same four brand colours, slate, copper, gold and flat, and every colour on it was measured against the ground the way the paper palette was measured: nothing sits under the 5.29 to 1 floor the paper comment states, and the measurements are in the source beside the values. No figure, grade, date, kill or rule changes with the ground."
    },
    {
      "n": "21",
      "kind": "standard",
      "title": "The first crowd row, 9 September 2026. Named before, graded after, applied to somebody else's number.",
      "body": "The section below the ledger carries its first claim tonight, and the shape of it is worth stating because it is not the shape note 18 assumed. Note 18 imagined a claim quoted and graded in the same breath, on the same Friday close as the rows above. The first claim the desk could source that passed every test, an institution's zero to three month target with a level, a window and no condition, does not resolve for thirteen weeks, and none of the loud claims in the market resolve at all, because they carry no clock. So the crowd row goes on the page the way this desk's own rows have gone on since 30 August: open, with its kill assigned and its grade date stated, and graded in place when that close prints. That is the more honest version of the format and it is now the standing one. A crowd row is named on the day it is written here and graded on the last weekly close inside the claim's own window; the countdown under it counts to that close and not to this page's naming week, and the code was changed tonight to make it do so. TWO RULES TOUCHED. Rule 2 of the crowd section said the link to the post. It now says the link to where the claim was made, in any dated public medium, because the loud claims are made everywhere at once and a dated article is the harder exhibit to delete; the founder ruled on 7 September that he will not search X for a claim to grade, and the desk does the finding from the open record instead, on a fixed day, by six tests written into the collision file. Rule 1 stands exactly: the source is recorded on the row because a quote without a source is a fabrication, and it is not spoken or written anywhere built from the row. WHAT THE FIRST ROW IS AND IS NOT. It is a target of 4,800 with a window that closes on 4 December, published without the close that would grade it. The finding is that sentence and only that sentence. The row does not say the number is wrong, does not say who will be right, and carries no adjective about the desk that published it. What the page adds is the kill the note did not carry, stated with the rule for assigning it, and the closes the note did not mention: the same desk's target has read 4,000, 4,500 and 4,800 since June, each move after the price had moved, and the same desk saw the second half of 2026 at 2,500 to 2,700 when it wrote in June 2025, against a 4 September close of 4,430.520. Those are dated figures with sources and they are on the row's frames, not in this note, because this note is about the rule and the frames are about the claim. Nothing above this note was touched. The tally does not know this row exists."
    },
    {
      "n": "22",
      "kind": "standard",
      "title": "The board on the page, 9 September 2026. The banks' targets, a third door, and five words under the masthead.",
      "body": "Three things changed on the page tonight and none of them touches a row, a grade, a figure or a date in the ledger. THE BOARD. The crowd section now carries the banks' gold targets as a standing ledger under the collision rows: each one an exact quote with its date and its source, the kill this desk assigned by one rule, the window, and the grade when the window shuts. The rule is the rule the rows above are graded by, stated once on the board: held if a weekly close on the target's side of it prints on or before the last weekly close inside the window, wrong if the window closes without one, a range on its nearer bound. What is not a target is not on the board and is listed by name with the reason on the full page, targets.html: averages for a year, toward, approach, could, and anything with no window. A revised target is a new row and the first one stays; that is the whole point of keeping it. A row whose window has already closed stays open until the founder has read the closes off his own chart, and says so; the first such row is on the board tonight. The board is never in the tally. THE THIRD DOOR. The indicator every level on this page is read from is available by invitation, on the reader's own chart, and it has a door beside the free one and the seat. The door leads to a page that describes the tool and sells nothing until its own terms and links exist, so it can never lead anywhere dead, which is the rule every door on this page has kept since the first one. THE FIVE WORDS. A row under the masthead names the five places this page now is: the record, the banks graded, the wrong ones, the Method, the seat. The record was one page with two doors at the bottom, and a stranger should not have to scroll to the footer to find that the banks are graded here. The lede of the crowd section was rewritten to say what the section has become; the replaced sentence is kept in the source. Nothing above this note was touched. The tally does not know any of this exists."
    },
    {
      "n": "23",
      "kind": "standard",
      "title": "The ladder stood up and the notes folded, 9 September 2026.",
      "body": "Two surfaces on this page were rebuilt tonight and neither touches a row, a grade, a figure, a kill or a date in the ledger. THE LADDER. Since 4 September it drew each metal on a horizontal axis with its labels stacked on shelves above the line. Three panels of height went to the empty space between levels, gold's labels still ran into each other whenever two lines sat close, and phones were given a list instead of the picture. It now stands the way a ladder stands on a desk: one column per metal, price running up a spine, every level ever named drawn as a bar at its price, and the last close on the standing cards drawn as a line across the whole column, so what sits above the close and what sits below it is read on sight. Every level states its distance from that close in the metal's own units, derived on every load from the standing card's figure and never typed. Levels within half a percent still join one band, and a band whose lines sit within a tenth of a percent of each other, one line named at two precisions or moved by a dollar, is labelled with the line named the most weeks rather than as a range of forty cents; its count line says how many lines it holds and every one of them is listed when the label is hovered. Each level is a link to the last row that named it. Labels are measured once they are on the page and pushed apart so none overprints and none crosses the close line, which is the only way the record has found to keep a plot honest as it fills. THE NOTES. Twenty-two of them, each already behind its own fold, still stood as twenty-two summary lines and a screen of the footer. They now sit behind one fold, shut on load, whose face carries the count and the newest note's title; a note reached by a link from a row opens the fold and then itself, and the open all control is inside. Every note is on the page in full, one click further away than it was. The plot this replaces and the head line it replaces are in the edit history and the source."
    },
    {
      "n": "24",
      "kind": "standard",
      "title": "One word per grade, the kill and the close on every line, 9 September 2026.",
      "body": "Six surface changes, made after the founder asked whether the page was perfect and the desk said no; none touches a row, a grade, a figure, a kill or a date. ONE VOCABULARY. The row chips read Pass, Miss and No test while the dividers, the strip, the legend, the tally cards and the standing cards read held, wrong and never reached, and the tally and the masthead read correct. A stranger was translating between four words for one grade. Every surface now reads held, wrong, never reached and open. The desk's own results keep PASS and MISS in their prose, because that is the word the grade was given in and the record does not rewrite its own past. THE LINE. Every ledger row now carries its kill and, once graded, its close on the summary line, read off the row's own kill sentence and result by the same rule the gauge uses, so the ledger can be audited without opening thirty-one rows; a row that states neither shows nothing there. THE BOARD. The banks' quotes were behind a fold on the record page, so the section whose first rule is the claim, never the person, showed a bank and a number and hid the claim. The quote is on the face; the kill folds. THE FACE. The standing card's plain face now states the date of its figures, because the head named the restatement of 28 August and the rows carried the 4 September close with nothing between them saying so. THE NEXT VERDICT line printed its Friday twice. THE SEAT's price is said on the call and not printed on the form, ruled by the founder, and method.html, which carried two prices and no way to ask, now says how to ask. Every replaced line is kept in the source."
    },
    {
      "n": "25",
      "kind": "standard",
      "title": "The tick book opens, the board takes silver, six places on every page, 9 September 2026.",
      "body": "THE TICK BOOK. From tonight this desk keeps a second ledger, tick.html, for the tactical clock: a metal, a side, a level and a kill on the five-minute chart, written and posted before the trade, graded on the 23:00 Madrid close of the day the row names. Held if the grading close sits on the row's side of the level and no five-minute close printed through the kill first; wrong if one did, at any time before the grading close; never reached if price never traded to the level. Two timestamps per row that this desk cannot edit, the post that named it and the commit that wrote it. No size and no profit or loss on any row, ever; the book is a claim about price and a close and nothing else, and it never enters the weekly tally. The founder has said in private that the day trading is where the edge is. Until tonight nothing public could support or refute that sentence. The book opens empty, and it is published empty on purpose, so the first row cannot be the one that looked good afterwards. THE BOARD. The first silver row: Citi's 13 January note named silver at 100 in the same breath as gold at 5,000, and silver crossed 100 on Friday 23 January and ended that week above 103.50, as reported. Held, ten days after it was published. The board's copy now says gold, silver and platinum, each row carries its metal, and the page never guesses a metal for a row that does not state one. THE PAGES. Every page carries the same six places under its masthead, the record, the banks, the wrong ones, the tick book, the Method and the seat, so a reader who lands anywhere can reach everywhere. The Method page carries the chart itself, gold on the 8 hour with the tool on it and the July break that the record grades wrong in frame, because a page selling a chart with no chart on it was selling a sentence. THE TRIMS. The desk's paragraph under the standing cards folds behind one line and the crowd lede is four sentences shorter; every replaced word is in the source. Nothing in the ledger, the tally or any grade changes."
    },
    {
      "n": "26",
      "kind": "correction",
      "title": "Resolution, 11 September 2026. Two floors broke, both by less than ten dollars, and both stay.",
      "body": "Friday 11 September close, read off the founder's 8h TVC charts stamped 12 September with the market shut, 03:09 CEST for gold and 03:10 for silver and platinum: gold 4,347.760, silver 64.4810, platinum 1,791.50. Two floors were named on 6 September before the week opened, each published with the weekly close that would kill it, and both kills fired. GOLD. The floor was 4,357, taken from a line labelled 4,357.307 and rounded down to the dollar. The close printed 4,347.760, which is 9.24 under the number as published and 9.547 under the line as it reads tonight, unchanged from Sunday. MISS. The rounding is disclosed because the Sunday session raised it as a standard question: rounding a floor down makes the test easier by the cents dropped, 30.7 of them this week. It did not decide anything tonight, because the close sat under the label as well as under the number, and it has not decided any grade on this page. Whether floors round up or print to the cent from here is the founder's ruling and it will be logged as its own note when he makes it; it changes no grade already written. PLATINUM. The floor was 1,797.40, named on 30 August, held on 4 September, and kept for a second week although the line had drifted five cents under it, because a floor does not follow a line down. The close printed 1,791.50, which is 5.90 under the number as published and 5.85 under the line. MISS. Third row for platinum, first wrong. SILVER, THE ROW THAT IS STILL NOT HERE. The door published on 30 August was a weekly close above 66.80. Tonight silver closed 64.4810 with the line reading 66.6878, so the close sat 2.319 under the door and 2.207 under the line. The condition is not met for a third week and silver carries no row again, stated here rather than left as a gap, for the reason note 17 gave. THE STANDING CARD. All three metals closed under their lines on the same night, gold and platinum for the first time since their clocks turned and silver for the third week running. Under the ruling of 29 August that is confirmation moving against the claim and not the event that ends it; the event is the epoch midline crossing the slow mean reversion line, read off the founder's chart and dated. No such read has been ruled tonight, so the card stands as last read and says so in its own words. If the founder rules a cross on any metal, the state, the date and the reason go on the card and this note is not edited. THE TALLY. Twenty-six rows resolved, twenty held, six wrong, five never reached, none open. The six wrong rows sit in the ledger below in the same type as the twenty, and misses.html lists them alone. THE TICK BOOK. The first rows in the second ledger, tick.html, are written from Monday 14 September, each posted before the trade and graded on that day's 23:00 close, under note 25 and the rule printed on that page."
    },
    {
      "n": "27",
      "kind": "correction",
      "title": "Correction, 13 September 2026. The wrong-calls page printed a total its own table contradicted.",
      "body": "The page that lists the wrong calls in full, misses.html, is generated from the ledger on this page so that it can never name a wrong row the record does not, or omit one it does. One sentence on it was not generated. The closing paragraph read that two of these four were the same idea, restated for weeks while it held, and then wrong, and the word four was typed by hand on 5 September, when four rows on this record had been graded wrong. The 11 September close graded two more. The upload of 12 September therefore put six wrong rows in the table with a sentence under them that still said four, and it stood that way until tonight. No grade is touched by this and no row is affected. The tally on this page, the tally line on that page and the count in its own head were all six throughout, so the page contradicted itself in one paragraph rather than misstating the record. That is still the defect worth naming, because the reason that page exists is a count that cannot drift, and a number typed once will always drift eventually. Both numbers in that sentence are now counted out of the ledger. The pair the sentence points at is found by the rule the ladder already uses under note 23, which is that two levels are the same line when they sit within a tenth of a percent of each other on the same metal, and today that pair is gold 4,098 named on 12 July and gold 4,098.275 named on 7 August, which is what the sentence has always been about. If a week comes when no pair repeats, the sentence about repeats is not printed at all rather than reworded into something the ledger does not say. The verifier that refuses an upload gained a check for it, so a prose count on that page which disagrees with the ledger now stops the upload instead of shipping. The replaced sentence is kept in the generator's source with its date, in the ordinary way. The three older mentions of four wrong rows that remain on that page sit inside notes 11, 13 and 17, where they are correct: each records the tally on the day it was written, and a note is never restamped to match a later count."
    },
    {
      "n": "28",
      "kind": "standard",
      "title": "Map of 13 September 2026. Three doors and no floor, the first week with nothing open.",
      "body": "The 11 September close graded gold and platinum wrong by 9.24 and 5.90 (note 26). Silver's floor broke on 28 August by 3.3 percent (note 11), and the map of 30 August did not replace it with a lower number; it published a door, a weekly close above 66.80, and the door has not been met on three Fridays or lowered on any of them (notes 16, 17). Tonight the same rule is applied to gold and platinum. A broken floor is not replaced by a lower one. Gold's door is 4,357 and platinum's is 1,797.40, the numbers that broke, and neither comes down to meet the price. The rule was written at a 3.3 percent break and is applied at 0.21 and 0.33 without adjustment, because a rule that is obeyed only when it costs nothing is not a rule. No row is named tonight. This is the first week the page carries nothing open, and the absence is stated here rather than left for a reader to explain to himself. The three conditions are read on the 18 September close: a metal whose weekly close prints above its door earns a floor back on the map of 20 September, read off the chart that night. The fixed frame of the Sunday map, one number on the page, is not spoken this week; the reason there is none takes its place. Closes read off the founder's own 8h TVC charts stamped 13 September with the market shut, 21:54 to 21:57 CEST: gold 4,347.760, silver 64.4810, platinum 1,791.50."
    },
    {
      "n": "29",
      "kind": "standard",
      "title": "Map of 20 September 2026. The rows move to the opening box, by the founder's ruling, and the three doors are closed out.",
      "body": "Friday 18 September close, read off the founder's own 8h TVC charts stamped 19 September with the market shut, 00:06 CEST for gold and silver and 00:07 for platinum: gold 4,377.290, silver 66.2210, platinum 1,800.40. No row was open. Note 28 named three conditions in place of rows and said they would be read tonight. Read: gold closed 20.29 above its door of 4,357, met; platinum 3.00 above 1,797.40, met; silver 0.579 under 66.80, not met, the fourth Friday running. THE RULING. On 20 September the founder retired the 464 swing slow mean floors as the object the weekly rows are named on. From tonight the rows name the opening box: the first sixteen hours of each week, Sunday's open through Monday's Asia and London morning, boxed and frozen, with its edges run right for the rest of the week, drawn by THE METHOD v4.0 shipped the same day. A metal that closed the week above the top of its box holds the top. A metal still inside holds the bottom and has its top stated beside it. The kill is what it has always been, a weekly close through the published number, and the grade is on the number as published. THE DOORS ARE CLOSED OUT, NOT CARRIED. Gold and platinum met theirs on this close and would have earned floors back under note 28; those floors are not named, because the object they belonged to is retired the same night, and a floor carried from a retired object would be one number living under two rules. Silver's door was not met on four Fridays and is retired unmet. All three last reads stand in the ledger and here. WHERE THE NEW LINES SIT. Every new line is under the door it replaces and the figures are given so nobody finds them later: gold 4,337 against 4,357, twenty under; silver 64.08 against 66.80, 2.72 under; platinum 1,746 against 1,797.40, 51.40 under. Notes 16, 17 and 28 refused to lower a broken floor within one object, and that rule stands within the new one: a box edge that breaks is not replaced by a lower edge the following Sunday. Tonight is a change of object, ruled and logged, not a lowering inside one. THE BOX OF 14 SEPTEMBER, read off the founder's 8h TVC charts stamped 20 September, market shut, 23:19 CEST for gold and silver and 23:20 for platinum: gold 4,251 to 4,337, the Fed wick of 16 September under the bottom without a close, the week closed out the top by 40; silver 62.31 to 64.08, closed out the top by 2.14; platinum 1,746 to 1,803, closed inside, 2.60 under the top. Published to the dollar and to the cent as written on each row and graded on those numbers. THE STANDING CARD stands at the 11 September read. It is graded on the swing clock, a different object on a different rule, and the 18 September charts carried no read of its line, so it is not refreshed tonight rather than refreshed against a line nobody read. Stated here the way note 19 stated a stale card on 6 September. Three rows are named for the week of 20 September, gold 4,337, silver 64.08 and platinum 1,746, written to this file before the week opened."
    },
    {
      "n": "30",
      "kind": "correction",
      "title": "Disclosure, 22 September 2026. The rows of 20 September were read before the open and written eighteen minutes after it.",
      "body": "The three rows named on 20 September, gold 4,337, silver 64.08 and platinum 1,746, each say they were named before the week opened, and note 29 says they were written to this file before the week opened. The first is true of the reads. The second is not. The levels were read off the founder's 8h TVC charts at 23:19 and 23:20 CEST on 20 September with the market shut. The market reopened at 00:00 CEST on 21 September, the end of the same daily break from 23:00 to 00:00 that the founder's 30 second charts of 21 September show. The rows were written to this file in commit d048757 at 00:17 CEST, eighteen minutes after the open, and the map that published them went out after it too. Nothing in the rows came from the new week: each level is an edge of the box of the week of 14 September, which froze on 14 September, and no bar of the new week can change it. The rows stand as written and are graded on the numbers as published. The sentence that describes them is corrected here and not in the rows, because a published row is not edited to fit a fact found afterwards."
    },
    {
      "n": "31",
      "kind": "standard",
      "title": "The second crowd row, 23 September 2026. One house, six numbers, graded on one rule.",
      "body": "The crowd row named today is a gold target of 6,200 published on 29 January 2026 for March, June and September 2026, in one sentence. The rule on this page grades a level against a window, so one sentence with three windows is three rows on the bank board, each graded on its own close, and the crowd row is the third, the one still open when it was named. It grades on the close of Friday 25 September. THE CHAIN. The same house said 5,000 by September 2026 on 29 December 2025, with gold at 4,322.40 as reported, and 5,000 by March on 6 January 2026. Both are on the board as held: weekly closes above 5,000 printed in February, inside both windows. The first number was right, seven months early, and the board says so in the same word it gives this desk. On 29 January 2026, the day after gold's high of the year, 5,589.38 on 28 January as reported, the target went to 6,200 for all three windows. March shut without a weekly close at or above it, and so did June, on a close of 4,088.97 read off the founder's chart. On 27 May the year-end target was cut from 5,900 to 5,500, which was already a row. HOW THE CLAIMS WERE SORTED. A claim that names a path, up to one level and then back to another, is graded on its first leg. The return leg is a second direction on the same clock and goes below the board with that reason: the 4,800 by the end of 2026 in the note of 29 December and the 5,900 by the end of 2026 in the note of 29 January. The 5,400 on midterm turmoil is conditional, towards 4,800 is not a level, and the 7,200 and 4,600 scenarios name two directions. All five are listed by name. THE FIGURES. Every quote was fetched at its source on 22 September 2026. Closes before June 2026 are as reported, under the rule this page has used for them since 8 September, and are replaced by the founder's chart read when he supplies it; the year's high as reported still stood on 22 September. No grade here turns on the decimals: the nearest a 6,200 window came is the year's high, more than 600 under the line, and the 5,000 rows rest on February closes, the one printed figure being 5,232.50 on 27 February as reported, 232.50 over the line. THE 5,500 ROW'S NOTE is corrected, the old wording kept in the source: it said the 5,900 was listed below the board, and until today it was not. None of these rows enters the tally. The desk's own gold row for the week, 4,337, grades on the same close as the September window, and both stay whatever prints."
    },
    {
      "n": "32",
      "kind": "standard",
      "title": "The first screen, the ladder and the standing calls, rebuilt. Written 22 September 2026, uploaded 23 September.",
      "body": "Seven surfaces on this page changed, and none of them touches a row, a grade, a kill, a figure or a date in the ledger. Each is stated here with the reason. THIS WEEK. The open rows, the last Friday close and the room between them now sit on the first screen, under the masthead. Until today a reader arriving from a post met the thesis, the tally and the strip, and the three lines the post was about sat at the top of the ledger several screens down. When nothing is open the same space shows the week last graded, with its grades and its close. THE CLOSES. The page now keeps the Friday closes it grades on in one table, CLOSES, each figure copied from the row or note that first printed it, and filled with every Friday grade from the same reads. Nothing in it is new. This Week and the ladder anchor on its last entry. THE LADDER. It read its close off the standing card, frozen at 11 September by note 29, so every distance it printed was a week old, and it joined lines within half a percent, so 4,337 on the opening box, 4,350 and 4,357 on the retired floors printed as one blue band with one held, one wrong and one open inside it. It now anchors on the last Friday close, joins only lines within a tenth of a percent, the same line named at two precisions, draws every line inside three percent of the close at its exact price with one square for each week it was named, and lists the rest at the edge of the column with their distance. On a phone the metals sit behind tabs, one at a time. The count of levels in the tally keeps its own family rule, stated beside it on 29 August, which is deliberately the generous one; the ladder groups for drawing and the tally groups for counting, and neither is changed to match the other. THE ONES THIS DESK GOT WRONG moved from under the standing cards to under the count, where the claim it proves is made. THE STANDING CALLS. The week-to-week entry was drawn on the 464 swing slow mean, the object the founder retired on 20 September, and it was last read on 11 September. It had not ended: its kill is the cross, and the cross had not printed at that read. It is therefore not ended, not restated and not graded. It carries a third status, retired, beside live and ended, with the date and the reason, and it stays on the page as it stood, every figure included. The year-to-year entry is the same call with the same kill in plainer words: its position read The bull is intact and now reads Up, on the secular clock, until the weekly cross. Its two marks, a run of 217.79 percent in gold since 28 May 2019 and 230.38 percent in silver since July 2020, had been flagged in the source since 1 August as not reproducible, because the chart measures the visible range and the same frame read 221.87 percent. A run figure is a mark of the market and not of this page, so they are withdrawn from the face, kept in the source, and replaced by the two pairs the kill itself reads, fast mean over slow mean at the 1 August read. That read is seven weeks old, and the entry says so on its face until the founder reads it again. THE SILVER ROW OF 16 AUGUST. The row reads that silver stays below the line, 66.60 at naming, and that its kill is a weekly close above it. Silver closed 68.9460. The row is graded held on the conditional posted on X on 10 August, and note 10 has recorded since 21 August why the kill as staged on the page inverted it. It is the only row on this page whose kill as written fired and whose grade is held. The grade is not restamped, because grades are never restamped. What changed is where it is said: the row's own summary line now says it at every width, and the count beside the tally says it, so no reader meets the held without the sentence. THE LINK CARD. Every link to this page and to the bank board posted on X showed one held row from 30 August, three weeks stale, with a sign-off since retired. The card is now regenerated from record.json every Sunday and Friday: the week's lines or the week's grades, and the count, wrong rows included. Every replaced line of copy is kept in the source beside the line that replaced it."
    },
    {
      "n": "33",
      "kind": "standard",
      "title": "From 28 September the rows are named on Monday evening, on the week's own box. Written 24 September 2026.",
      "body": "The week is judged against its own opening box: the first sixteen hours from Sunday's open, frozen on Monday afternoon. Until now the rows were named on Sunday night, before that box existed, so they sat on the box of the week before; the three rows of 20 September are edges of the box of 14 September. From 28 September the rows are named on Monday evening, on the week's own box, after it freezes and before the first 8h close after that, which is midnight in Madrid in most weeks, and they are written to this file before the post that names them. The kill does not change, a weekly close through the published number, and neither does the grade, the Friday close. The cost is stated here so nobody has to work it out: a row is now named with about 86 percent of the week still to trade, where until now it was named before the open. The rows of 20 September grade on 25 September as written. The cadence line under the masthead moves with this note, as do the lines on this page, in its description and on the method, advisers and year pages that promised the levels on a Sunday or before the open; the replaced copy is kept in the source, where this file keeps everything it replaces."
    },
    {
      "n": "34",
      "kind": "standard",
      "title": "The call grammar, from 28 September 2026. Side, line, room and odds on every row, and a third grade word. Written 24 September 2026.",
      "body": "From the rows of 28 September every row on this page says four things, and the file refuses a row that leaves one out. THE SIDE. Slate, the metal holds above the line, or copper, it stays below it. The side is the founder's read and nothing on this page picks it: a rule that read the side off the chart was written down before it was tested, run on the founder's own 8h exports on 23 September, and failed on the copper side, so it sets no row. THE LINE. A weekly close through it grades the row wrong, as every kill on this page always has. It is an edge of the week's own opening box, or a level of the founder's own inside the range of the week before, so no line sits so far from price that it cannot lose. THE ROOM. The next level on the called side. THE ODDS. The chance the line holds, printed on the row the moment it is named, from the tint THE METHOD draws on the box, which was fitted on weeks before June 2025 and kept its calibration on the weeks after; an own level takes the same formula and says so. THE GRADE has a third word. HIT, the week closed at or through the room. HELD, it closed on the called side of the line, short of the room. WRONG, it closed through the line. A hit is a held row, counted once, as held, and marked, so no week is counted twice. Never reached does not exist under the grammar, because the line and the room are both on the page and the close decides. THE ODDS BOARD under the tally adds up the odds each graded row was named at, which is what the rows would have held if the read added nothing, and sets the rows that held beside that figure. Held past it is the read beating the odds; held short of it is not, and the board says so in the same type. It counts the rows by side as well, so a lean either way is on the page rather than left to a reader's arithmetic. Rows named before 28 September keep the words they were graded in. The grammar is frozen to 13 November 2026."
    },
    {
      "n": "35",
      "kind": "standard",
      "title": "The page answers a pointer, 24 September 2026. The loupe, the crosshair, an address for every row, and a page for every bank row.",
      "body": "Five surfaces changed, and none of them touches a row, a grade, a kill, a figure or a date. THE LOUPE. A square on the strip, a line on the ladder and the close each open a small card on a pointer, on keyboard focus or on a first tap: the row's metal and level, its grade, the close it was graded on and how far that close sat from the level, the kill in the row's own words, and any flag the row carries. The close is the one the row's result prints, or where the result prints none, the Friday close in CLOSES for the date PROVENANCE says the row was graded on; the two agree on every row that carries both. Rows from the first upload carry no graded date and show no close. THE CROSSHAIR. Across the ladder a hairline follows the pointer and prints the price at that height on the column's own scale, and its distance from the last Friday close. Within a few pixels of a named line it holds to that line, the other lines dim, and the card lists the rows that named it, newest first, each with its grade and, where the row named one level, its close. THE STRIP. On a wide screen the squares are grouped by the month the row was named, the key marks the rows named before the standard of 3 July, and the arrow keys walk the squares. AN ADDRESS FOR EVERY ROW. A row was reached by its position in the array, lrow and a number, and new rows go on top every week, so every such link pointed at a different row the week after it was shared. Every row now also carries an address built from its own date, metal and level, r-2026-09-06-gold-4357 for the gold row of 6 September, which never moves. A link inside the page writes that address into the bar, the wrong-calls page links by it, record.json carries it on every row as anchor, and a button in every opened row copies it, or on a phone hands it to the share sheet. An old positional link still opens whatever row sits at that position today. A PAGE FOR EVERY BANK ROW. Each target on the bank board links to a page of its own under banks, with the quote, the kill this desk assigned, the grade, the house's other rows, what the house said that the board leaves off, and a link card of its own. The pages are written from record.json by make_bank_pages.js, and verify_record.js checks that every row has its page and its card. Three lines that promised the rows before the open, which note 33 did not reach, move with it in this upload: the brief for producers, the first line of the README and the ledger line on the year page. An opened row draws its gauge from the level to the close, and a row reached by its address is marked in gold for two seconds. Motion is kept to what the reader does, with one exception, the dot beside the countdown, which breathes while the clock runs; a reader who has asked the system for reduced motion gets none of it."
    },
    {
      "n": "36",
      "kind": "correction",
      "title": "Correction, 24 September 2026. The bank pages went up without their folder, the crosshair left a mark on the ladder, and the chart now reads over time.",
      "body": "Note 35 was uploaded at 17:34 CEST on 24 September. Two things in it did not work, and both are corrected here rather than tidied. THE BANK PAGES. The upload put every file in the top folder of the site, so the fourteen pages note 35 placed under banks never reached it. Each was replaced by the source of its own link card, which carried the same name, and every link from the board to a row's page opened nothing. The pages now sit beside the record itself, named for the house, the metal, the level and the window, ubs-gold-6200-2026-09-30.html for the UBS row that grades on 25 September, which is where a plain upload puts them; the link cards sit beside them as images, and their sources are no longer written to the site at all. verify_record.js now refuses a page that is not the page it names, or that was written before its row's grade. THE CROSSHAIR. After a pointer left a column of the ladder, the tag that printed its distance from the close stayed on the column as an empty bar. It no longer does. THE CHART OVER TIME. The section that held the ladder now opens on the same rows over time: the Friday close as a line, every level named drawn from the day it was named to the close that graded it, in the colour it was graded, and a thin tick from the level to that close, so the margin of every held row and every wrong one reads at a glance. A card names the week or the row under the pointer, and a click opens the row. The ladder is one tab away, by price, as notes 32 and 35 left it, and a reader who chooses it keeps it on that device. The chart draws from CLOSES, CALLS and PROVENANCE and nothing else; the four gold rows of June, named before the closes on file begin, are counted under it and stay in the ledger. Nothing in this note touches a row, a grade, a kill, a figure or a date."
    },
    {
      "n": "37",
      "kind": "standard",
      "title": "The phone, the counter, a card for every page, and the Method's charts, 24 September 2026.",
      "body": "Four things changed, and none of them touches a row, a grade, a kill, a figure or a date. THE PHONE. On a phone this page opened on its title, two paragraphs and a table of terms, with This Week below the first screen, and it ran to about twenty-three screens. From tonight, on a screen 760 pixels wide or less, This Week sits under the title, a card per metal. The standing calls show their state, their dates and the live kill, and the rest opens under Open the call. The banks' targets show their count and the next to grade, the table opens under Show the board here, and the whole board has its own page. The ledger opens on the four latest weeks named, with a line under it that shows every row in one tap, and a link to any row, a square on the strip or a line on the chart opens the whole ledger first, so no row sits behind a fold it did not ask for. Running text on a phone is 12 pixels at least, the buttons and the nav are a thumb's height, the squares of the strip are larger, and the nav is one row that scrolls. A wider screen reads as it did. On every screen the nav's first link now lands on the ledger; it has pointed at nothing since 8 September. THE COUNTER. Note 18 wired a counter in and left it off until it had an address. It has one: GoatCounter, open source, no cookies and no personal data, on every page but the embed and the seal, which say they track nothing and still do not. It counts visits, and on this page three events by path alone: a row reached by its address, a row's link copied or shared, and a door used. It cannot see, change or influence a row. THE CARDS. The pages beside the record showed no picture when their link was posted, or another page's: the wrong ones, the Year, the producers' page, the advisers' page and the embed page had none, the board borrowed this page's week card, the tick book showed a held row from 30 August and the Method its chart of 9 September. Each now has a card of its own, drawn from record.json by make_link_cards.js, and verify_record.js refuses an upload whose cards were drawn from figures the record no longer holds. A card's address now carries a fingerprint of the card, the bank rows' cards included, so a messenger fetches a changed card again instead of showing the old one. The tally cards, the first frame of every video, were drawn in a stand-in face on 6, 12 and 21 September for want of the record's own; they carry it now, with the same four numbers. THE METHOD. Its chart of 9 September gives way to two the founder took at 18:12 on 24 September, gold on the 8 hour and on the 1 hour, with the three gold weeks this page grades wrong, named 12 July, 7 August and 6 September, in frame on both. The chart of 9 September stays in the folder and in the page's source."
    },
    {
      "n": "38",
      "kind": "correction",
      "title": "Disclosure, 24 September 2026. Every version of this file, read before the standing offer went up.",
      "body": "Note 40 puts $1,000 on this page's history, so the desk read that history first: every version of this file from the first upload of 30 June to tonight, row by row, for any level, kill, side, date or grade that changed after it was published and any row that went missing, and the bank board's rows and grades the same way. Three things turned up that meet the offer's terms. One is already on the page: the silver row of 16 August was pushed graded WRONG at 23:11 CEST on 21 August and changed to PASS at 23:19, which note 10 discloses with its reasons. Two were not, and are here. THE FIRST EVENING. This file was first uploaded at 21:04 CEST on 30 June. At 22:30 the same evening the gold row of 22 June, uploaded at 21:04 naming 4,098 alone with the text: while price is capped below the 4,098 floor, every green day is a bounce in a falling market, not a turn, was changed to name 4,098 and 4,633, with the text it carries now. Its grade is given on the bounce claim, which both texts make, as its result says. THE BLANK PAGE. At 15:19 CEST on 2 August this file was overwritten by mistake with a single line of text, and for sixteen minutes the page showed no calls. At 15:35 it was put back in full. Every row, level, kill, date and grade came back as it was. Besides these the history holds wording changes that move no figure, side or date, among them the kill sentences of six early rows labelled on 27 July in a commit titled layout only. The commits: d828c30 and 77bf1b5 on 30 June, 9da7551 and e5a11b5 on 2 August, a7337ea and a0e26cb on 21 August. If this read missed something, the offer covers it."
    },
    {
      "n": "39",
      "kind": "correction",
      "title": "Disclosure, 24 September 2026. The box the rows of 20 September were read from is not the box the chart draws for that week today.",
      "body": "The three rows of 20 September, gold 4,337, silver 64.08 and platinum 1,746, are each an edge of the box the metal opened the week of 14 September in, read on the night of 20 September off THE METHOD as it then stood, version 4.0. Version 4.0 started the week at Monday 00:00 UTC, which left the week's first bar, from the metals' open at 22:00 UTC on Sunday, out of the box. On 22 September version 4.4 moved the start of the week to that open, and every version since draws the week of 14 September with that bar inside it. On the founder's 8h TVC exports of 23 September that box runs gold 4,252.70 to 4,355.34, silver 62.303 to 64.471 and platinum 1,749.184 to 1,805.50. A reader who opens the chart today finds the gold top at 4,355.34, not 4,337, the silver top at 64.471, not 64.08, and the platinum low at 1,749.184, not 1,746. The rows are not changed. They are graded on the numbers as published, which is what this page promised when they were named, and each keeps the words it was named in. From 28 September the rows are named on the Monday, on the week's own box once it has frozen (note 33), so the box a reader finds on the chart is the box the row was read from. If a later version of the tool draws a named week's box differently, this page says so."
    },
    {
      "n": "40",
      "kind": "standard",
      "title": "Two rulings, 24 September 2026. A standing offer of $1,000 for proof of tampering, and no forecast without its kill.",
      "body": "THE STANDING OFFER. From today the foot of this page carries an offer, ruled by the founder on 24 September. Anyone who finds a call in the ledger whose level, kill, side or date changed after it was first published, a call dated earlier than the day it was first published, a call removed, or a grade on this page changed after it was given, the bank board's included, is paid $1,000, and the finding is logged on this page as a note. The standing calls are outside it, because their lines are read off the chart each week by design. The wrong calls do not qualify: they are the point of the page. What the page has already disclosed does not qualify, in its notes or on the row, and before the offer went up the desk read every version of this file and disclosed what it found (note 38). Every change to this file has been a public commit since the first upload, so the proof, if it exists, is already public and needs no one's permission to find. A claim is made by sending the row's address and the commit or post that shows it to @TheSimplifier7 on X. The desk checks each claim against the public history and answers here, whichever way it goes, and the first to claim a finding is the one paid. NO FORECAST WITHOUT ITS KILL. The email door on this page said: no signals, no forecasts. From today it says: no signals, no forecast without its kill. The founder ruled on 24 September that a forecast goes public from this desk only as a row on this page, with the close that would prove it wrong and the date by which it is graded, and never as a number on its own. The desk's own gold number of 4,850, shown on X on 17 September on one axis beside the banks' targets, is the first case. It goes on the bank board as a row of its own, quoted from the post that published it and graded by the board's one rule, and until it is there it is not repeated anywhere."
    },
    {
      "n": "41",
      "kind": "standard",
      "title": "What a price is made of, 25 September 2026. An animation from the record's closes down to the trades, the people, the machines and the metal, on this page and the Method's.",
      "body": "Under The Chart this page now carries an animation, drawn in the browser by descent.js, that plays only when pressed. It walks one price down through what makes it. It opens on the record itself: the weekly closes on this page and every level named against them, drawn from this page's own data, so it moves with every grade. It goes into the last week on the record, folds that week into one candle, opens the candle into its trades and the trades into the orders behind them. Then it goes behind the orders: to the people who traded these metals face to face in the New York pits until July 2015, to the programs that now take part in most of the trading in gold and silver futures, to the matching engines outside Chicago where orders meet, and to the vaults where the metal lies and, when it is traded, usually does not move. Everything below the record is an illustration, and the screen says so while it runs. The week, the trades and the orders are a seeded model between the last two closes on the record, with no price printed on them; the facts in the words below them carry their sources on screen: CME Group, the CFTC, the LBMA, the LPPM and the Bank of England. It ends back on the record, on this week's lines. Gold, silver or platinum can be followed down, and the words are under it as text. On the Method page it plays by itself when it comes into view, and never by itself on a device set to reduce motion. The counter notes two events by path alone, the animation started and the animation finished. It touches no row, grade, kill, figure or date."
    },
    {
      "n": "42",
      "kind": "standard",
      "title": "Resolution, 25 September 2026. Gold's floor broke by 52.87, silver's held by 0.176 and platinum's by 31.80. The last rows graded in the old words.",
      "body": "Friday 25 September close, read off the founder's 4h TVC charts stamped 25 September with the market shut, 23:07 CEST for gold and 23:08 for silver and platinum: gold 4,284.130, silver 64.2560, platinum 1,777.80. Three floors were named on 20 September, each an edge of the box its metal opened the week of 14 September in, each published with the weekly close that would kill it. GOLD. The floor was 4,337. The close printed 4,284.130, 52.87 under it. MISS. Gold printed under the line on Monday, closed back over it on the 8h chart on Tuesday night at 4,357.68, and went back under on Wednesday, bars carried by the founder's 8h export of 23 September. SILVER. The floor was 64.08. Silver traded under it during the week, 63.8480 on the founder's chart at 08:21 CEST on Friday, and the weekly close printed 64.2560, 0.176 above it. PASS, at arm's length, and the row says so. PLATINUM. The floor was 1,746. Platinum went through it on Wednesday, a low of 1,740.486 on the 8h export, 5.51 under it, and the weekly close printed 1,777.80, 31.80 above it. PASS. THE LAST ROWS IN THE OLD WORDS. These three are the last rows this page grades as held, wrong or never reached. From 28 September every row is named under the call grammar of note 34, with its side, its room and its odds, and graded HIT, HELD or WRONG on the close. THE BANK BOARD AND THE CROWD. The September window of the UBS target of 6,200, the crowd row of 23 September, shut on the same close: 4,284.130, 1,915.87 under the line. No weekly close at or above 6,200 printed inside the window. Wrong, on the board and in the crowd, in the same word this page gives itself, and the desk's own gold row graded wrong on the same close. Tonight's two rows give gold's high of 2026 from the founder's weekly chart, 5,598.75, in the week of 26 to 30 January; the March and June rows still carry it as reported, 5,589.38 on 28 January, and no grade turns on the difference. THE STANDING CARD. The charts read tonight carry no read of the standing line, so the card stands as last read and says so. THE TALLY. Twenty-nine rows resolved, twenty-two held, seven wrong, five never reached, none open. Neither the bank board nor the crowd is in it."
    },
    {
      "n": "43",
      "kind": "correction",
      "title": "Correction, 28 September 2026. The first 8h close after the box freezes prints at 23:00 in Madrid, not midnight.",
      "body": "Note 33 says the rows are named on Monday evening, after the week's box freezes and before the first 8h close after that, which is midnight in Madrid in most weeks. The hour is wrong. The 8h bar that opens at 16:00 CEST on Monday does not run its eight hours: it ends at 23:00 CEST, when the metals stop for the daily break from 23:00 to 00:00 that note 30 cites, and the next bar opens at midnight. The founder's 8h TVC chart of gold, stamped 18:47 CEST tonight, shows that bar with 4 hours 12 minutes left to run. The deadline for a Monday row is therefore 23:00 CEST. In the weeks when Europe and New York change their clocks on different dates, the last week of October and the second half of March, the freeze, the break and the deadline each come an hour earlier in Madrid, 22:00 for the deadline. The commit time of each Monday upload shows whether its rows beat it. No row was named under note 33 before tonight, so no row is affected, and note 33 stands as written, corrected here and not in its own words."
    },
    {
      "n": "44",
      "kind": "standard",
      "title": "Resolution, 2 October 2026. Gold HELD, silver HIT, platinum HIT: the first rows graded under the call grammar.",
      "body": "Friday 2 October close, read off the founder's 8h TVC charts stamped 2 October with the market shut, 23:00 CEST: gold 4,142.710, silver 60.4470, platinum 1,699.70. Three lines were named on 28 September under the call grammar of note 34, each the high of the box its metal opened the week in, each with its side, its room, its odds at naming and the weekly close that would make it wrong. Each is graded by arithmetic on the close: through the line, WRONG; at or through the room, HIT, counted once as held; otherwise HELD. GOLD. Stays below 4,280.50, room 4,135.74, odds at naming 78. The close printed 4,142.710, 137.79 under the line and 6.97 above the room, 4,135.74. HELD. SILVER. Stays below 64.212, room 60.932, odds at naming 71. The close printed 60.4470, 3.765 under the line and 0.485 under the room, 60.932. HIT. PLATINUM. Stays below 1,775.80, room 1,720.72, odds at naming 71. The close printed 1,699.70, 76.10 under the line and 21.02 under the room, 1,720.72. HIT. THE ODDS BOARD. Three rows graded under the grammar, three held, two of them hits. The odds they were named at add up to 2.20. Held against that sum is read over many weeks; this is one week, three rows. THE BANK BOARD. No window shut on this close. THE STANDING CARD. The charts read tonight carry no read of the standing line, so the card stands as last read and says so. THE TALLY. Thirty-two rows resolved, twenty-five held, seven wrong, five never reached, none open. Neither the bank board nor the crowd is in it."
    },
    {
      "n": "45",
      "kind": "standard",
      "title": "The third crowd row, 7 October 2026. A second-quarter target, named after its window shut.",
      "body": "The crowd row named today is a gold target of 5,800 for the second quarter of 2026, in one sentence of a report published by Kitco on 16 February 2026: commodity analysts at ANZ expected prices to hit 5,800 an ounce in the second quarter, raised from a previous target of 5,400. It is the first crowd row named after its window shut, so it is named graded, on the rule the board has used for shut windows since 8 September. THE RULE. Held if a weekly close at or above 5,800 printed on or before the last weekly close inside the window, Friday 26 June 2026; wrong if the window closed without one. The quote says hit, a level reached, and the row grades those words as reported, as the board grades every reported target. THE CLOSE. The window's last weekly close was Friday 26 June, 4,088.97, read off the founder's chart, 1,711.03 under the line. Wrong. No weekly close at or above 5,800 printed inside the window: gold's high of 2026 on the founder's weekly chart is 5,598.75, in the week of 26 to 30 January, before the report and 201.25 under the line. HOW THE CLAIMS WERE SORTED. The 5,800 is on the board as one row. The 5,400 it was raised from is reported without its window and is listed below the board with that reason. THE FIGURES. The quote was fetched at its source on 6 October 2026. The weekly closes inside the window before 26 June are not the desk's reads, and the grade does not need them: no price in the window reached the line, on the founder's own weekly high. No grade here turns on the decimals. THE DESK'S OWN ROWS. The last were graded on the close of 2 October and none was wrong, so this edition has no wrong row of the desk's to name first. The three rows named on 5 October grade on the close of 9 October and stay whatever prints. None of these rows enters the tally."
    },
    {
      "n": "46",
      "kind": "standard",
      "title": "Resolution, 9 October 2026. Gold WRONG, silver HELD, platinum WRONG: the first rows graded wrong under the call grammar.",
      "body": "Friday 9 October close, read off the founder's TVC charts stamped 9 October with the market shut: gold 4,194.085 and silver 60.7790 on the 10 minute charts at 23:00 CEST, platinum 1,686.90 on the 8h chart at 23:14 CEST. Three lines were named on 5 October under the call grammar of note 34: the high of the box gold and silver opened the week in and the low of platinum's, each with its side, its room, its odds at naming and the weekly close that would make it wrong. Each is graded by arithmetic on the close: through the line, WRONG; at or through the room, HIT, counted once as held; otherwise HELD. GOLD. Stays below 4,170.41, room 4,124.48, odds at naming 61. The close printed 4,194.085, 23.675 above the line and 69.605 above the room, 4,124.48. WRONG. SILVER. Stays below 62.036, room 60.326, odds at naming 60. The close printed 60.7790, 1.257 under the line and 0.453 above the room, 60.326. HELD. PLATINUM. Holds above 1,699.00, room 1,740.40, odds at naming 59. The close printed 1,686.90, 12.10 under the line and 53.50 under the room, 1,740.40. WRONG. The two wrong rows are the first under the grammar, and they stay on this page as graded, as every wrong row has since 30 June. THE ODDS BOARD. Six rows graded under the grammar, four held, two of them hits, and two wrong. The odds they were named at add up to 4.00. Held against that sum is read over many weeks; this is two weeks, six rows. THE BANK BOARD. No window shut on this close. THE STANDING CARD. The charts read tonight carry no read of the standing line, so the card stands as last read and says so. THE TALLY. Thirty-five rows resolved, twenty-six held, nine wrong, five never reached, none open. Neither the bank board nor the crowd is in it."
    },
    {
      "n": "47",
      "kind": "standard",
      "title": "The record as a table, 10 October 2026. This week first, a table and a page for each metal, the row opened in the ledger's ink, and the week's odds read after each close.",
      "body": "Eight rulings the founder made on 7 and 8 October 2026 are on this page from today, and none of them touches a row, a grade, a figure or a date, or what would prove a call wrong. Every count is still recomputed from the rows on every load, the typed count is still checked against it before every upload, and no percentage is printed in headline type. Every function that carried a rule on the page before is carried across word for word, and the standing calls are drawn by the same functions, gates and checks as before. THIS WEEK FIRST. The page opens on the week's lines: each metal's line and its side, the last close on file and how far it sits from the line, the odds it was named at, and the clock to the Friday close that grades it. A bar of five places sits at the foot of a phone, this week, the results, The Fan, the banks and the doors, and along the top of a wide screen with the rules beside them; on a phone the rules are one tap from this week and from the results. The Monday Level is one tap away on every screen. THE TABLE. Under this week sits a table with a row for each metal: the rows graded, held and wrong, and its form, the last five grades, newest first. Each metal has a page of its own with every line it has carried. Never reached and open rows count for neither side, as before, and the table says so under it. The banks' targets are counted the same way, a row for each bank, and are never in this desk's count. THE ROW, OPENED. A row opens in one sheet in the ledger's ink: the call, what would prove it wrong and the grade, word for word, in serif type, the line and the close drawn to scale, and the public commit that wrote it. THE RUNNING SCORE. From the week named Monday 19 October 2026, after each close from Tuesday to Thursday, the founder reads the same chart again, on the 8h bar that closes at 23:00 Madrid time, and adds one dated line to days.txt beside this page with a snapshot of the chart: each metal's close, and the odds the chart prints of a weekly close above and below the week's range, in 100, as printed. The range runs from a line's next level to the line, so one of the two is the odds of a close through the line, and the page shows the chance the line holds as 100 less that figure, beside the odds it was named at, which never change. A line in days.txt is never edited or removed: a line the page cannot read as typed stays, and the day is typed again on the next line. A day with no reading says so, and a figure later found wrong against its snapshot is corrected by a note here, never in place. Friday's close grades the line with the week's readings beside it, so the odds are graded too. The readings are on this page only. verify_record.js checks every line before an upload: its date, its figures and its snapshot, that no line was written before the close it reads or after the close that graded its week, and that the file has only grown. THE WORDS, THE GROUND AND THE RETURNING READER. The frame speaks in plain words and the rows keep their own, verbatim. The ground follows the reader's device, white by day and navy when the device is set dark, with a switch that holds the reader's own choice on that device only. A returning reader is told in one line what was graded and named since his last visit, kept in his own browser and nowhere else. Every grade is drawn as a mark beside its word, never instead of it: a gold dot for held, a gold dot in a ring for a hit, a red square crossed for wrong, a grey ring for never reached and a blue ring round a point for open. THE FAN. The page of the indicator the weekly lines are read from takes the same look and says in one page what the tool shows: the higher timeframes' structure inside the chart a trader acts on, and the week's range with its odds, under the founder's gold 10 minute chart of 8 October 2026. Its prices are unchanged. Access is asked for by a direct message to the desk on X, and the line that said to reply to the Monday Level email is kept in a comment beside it. The page as it stood is kept in a comment at the foot of that file. RETIRED WITH IT. The action bar, the ladder of every level by price and the loupe across it, which one chart for each metal replaces, and the folds a phone opened on under note 37, which the views and the one sheet replace, except the ledger's four latest weeks, which stay. The address of every row (note 35) and the counter's three events (note 37) are unchanged. The page as it stood before today is whole in the edit history at the commit before this one, and its markup is kept in a comment at the foot of this file."
    },
    {
      "n": "48",
      "kind": "standard",
      "title": "The cards in navy, 10 October 2026. The image a link to this record shows is drawn in the record's own navy and type.",
      "body": "By the founder's ruling of 10 October 2026, every card the record draws for a link is drawn in the navy and the Barlow type this page has worn since note 47: the week card under a link to the record, the card of each page, the card of each bank target, and the tally cards a grade post carries. Held is gold and wrong is red, as on the page. Nothing a card counts has changed. Every figure on a card is still drawn from record.json and never typed, and verify_record.js still checks each page's card against it before every upload. Four page cards and the tally cards now carry gold's Friday closes since 26 June as a line, read off the founder's chart and held in record.json beside every other close. The week card's address now carries its fingerprint, as the other cards' addresses do, so a messenger fetches a card that was drawn again instead of showing the old one. The cards no longer close on the old sign-off; the address closes them. No row, grade, figure, date or line was touched."
    }
  ],
  "targets": [
    {
      "published": "2026-09-04",
      "source": "Citi Research",
      "level": 4800,
      "side": "above",
      "window": "2026-12-04",
      "windowText": "0 to 3 months",
      "quote": "The bank sets its 0-3 month point price target for gold at $4,800 per ounce and its 6-12 month forecast at $5,000 per ounce.",
      "link": "https://www.investing.com/news/commodities-news/citi-says-the-dip-in-gold-prices-is-a-buying-opportunity-93CH-4889056",
      "kill": "Held if a weekly close at or above 4,800 prints on or before the 4 December 2026 close; wrong if the window closes without one.",
      "note": "First raised to 4,800 from 4,500 on 24 August 2026 (Reuters). Also the first crowd row on the record, named 9 September.",
      "grade": "pending",
      "result": "",
      "page": "citi-gold-4800-2026-12-04.html"
    },
    {
      "published": "2026-09-02",
      "source": "Goldman Sachs Research",
      "level": 4900,
      "side": "above",
      "window": "2026-12-31",
      "windowText": "year-end 2026",
      "quote": "Goldman Sachs sees gold price at $4,900/oz by year-end.",
      "link": "http://www.kitco.com/news/article/2026-09-02/goldman-sachs-sees-gold-price-4900oz-year-end-investors-hedging-through",
      "kill": "Held if a weekly close at or above 4,900 prints on or before the last weekly close of 2026; wrong if the year closes without one.",
      "note": "Set at 4,900 on 19 June 2026, cut from 5,400; restated 2 September. Quote is the report's headline; the analysts' text projects gains into the second half without restating the figure.",
      "grade": "pending",
      "result": "",
      "page": "goldman-sachs-gold-4900-2026-12-31.html"
    },
    {
      "published": "2026-05-27",
      "source": "UBS",
      "level": 5500,
      "side": "above",
      "window": "2026-12-31",
      "windowText": "year-end 2026",
      "quote": "Swiss banking giant UBS has cut its year-end 2026 gold price forecast from $5,900 to $5,500 per ounce.",
      "link": "https://www.kitco.com/news/article/2026-05-27/ubs-lowers-2026-gold-price-forecast-5500oz-markets-are-rediscovering",
      "kill": "Held if a weekly close at or above 5,500 prints on or before the last weekly close of 2026; wrong if the year closes without one.",
      "note": "The 5,900 it was cut from was published on 29 January 2026 as the return leg of a path from 6,200 and is listed below the board with that reason; the 6,200 is on the board as three rows, one per window. A March 2026 note reading toward 5,900 to 6,200 this year is listed below the board: toward is not a level.",
      "grade": "pending",
      "result": "",
      "page": "ubs-gold-5500-2026-12-31.html"
    },
    {
      "published": "2025-12-29",
      "source": "UBS, as reported by Kitco",
      "metal": "Gold",
      "level": 5000,
      "side": "above",
      "window": "2026-09-30",
      "windowText": "by September 2026",
      "quote": "Gold prices will reach $5,000 per ounce by September 2026, and any increase in political or economic turmoil surrounding the U.S. midterm elections could push the yellow metal to $5,400",
      "link": "https://www.kitco.com/news/article/2025-12-29/ubs-sees-5000-gold-q3-2026-potential-5400-if-us-political-and-economic",
      "kill": "Held if a weekly close at or above 5,000 prints on or before the last weekly close of September 2026; wrong if the window closes without one.",
      "note": "The first leg of a path. The same note saw a pull back to 4,800 by the end of 2026, and 5,400 if the midterm elections brought turmoil; both are listed below the board with the reason. Gold was 4,322.40 as reported on the day of publication.",
      "grade": "pass",
      "result": "Held. Weekly closes above the line printed in February, seven months inside the window: Friday 20 February closed above 5,000 and Friday 27 February printed 5,232.50 as reported. Where price went afterwards is not the question the rule asks.",
      "page": "ubs-gold-5000-2026-09-30.html"
    },
    {
      "published": "2026-01-06",
      "source": "UBS, as reported by Kitco",
      "metal": "Gold",
      "level": 5000,
      "side": "above",
      "window": "2026-03-31",
      "windowText": "by March 2026",
      "quote": "We now expect gold prices to rally to USD 5,000/oz by March, stay there to September, and moderate towards USD 4,800/oz by the end of 2026.",
      "link": "https://www.kitco.com/news/article/2026-01-06/gold-will-reach-5000oz-q1-2026-amid-broader-commodities-rally-ubs",
      "kill": "Held if a weekly close at or above 5,000 printed on or before the last weekly close of March 2026; wrong if the window closed without one.",
      "note": "The same house restating its 29 December target with a March window. Only the first clause carries a level, a window and one direction. Stay there to September names no close to grade, and towards 4,800 is not a level; it is listed below the board.",
      "grade": "pass",
      "result": "Held. Friday 20 February closed above 5,000 and Friday 27 February printed 5,232.50 as reported, inside the window. Where price went after the window is not the question the rule asks.",
      "page": "ubs-gold-5000-2026-03-31.html"
    },
    {
      "published": "2026-01-29",
      "source": "UBS, as reported by Investing.com",
      "metal": "Gold",
      "level": 6200,
      "side": "above",
      "window": "2026-03-31",
      "windowText": "for March 2026",
      "quote": "UBS raised its XAU/USD price target to $6200 per ounce for March, June and September 2026, up from $5000.",
      "link": "https://www.investing.com/news/commodities-news/ubs-raises-gold-price-target-for-2026-here-are-the-new-forecasts-4474171",
      "kill": "Held if a weekly close at or above 6,200 printed on or before the last weekly close of March 2026; wrong if the window closed without one.",
      "note": "One level for three windows in one sentence, graded as three rows, each on its own window. Published the day after gold's high of 2026. The 5,900 by the end of 2026 in the same note is listed below the board.",
      "grade": "miss",
      "result": "Wrong. The window's last weekly close was Friday 27 March 2026. Gold's high of 2026, 5,589.38 on 28 January as reported, printed the day before the target was published and still stands as reported on 22 September 2026, so no weekly close at or above 6,200 printed inside the window. Graded on reported figures under the rule for closes before June 2026; the founder's chart read replaces them when supplied. The line sits more than 600 above the year's high, so the grade does not turn on the decimals.",
      "page": "ubs-gold-6200-2026-03-31.html"
    },
    {
      "published": "2026-01-29",
      "source": "UBS, as reported by Investing.com",
      "metal": "Gold",
      "level": 6200,
      "side": "above",
      "window": "2026-06-30",
      "windowText": "for June 2026",
      "quote": "UBS raised its XAU/USD price target to $6200 per ounce for March, June and September 2026, up from $5000.",
      "link": "https://www.investing.com/news/commodities-news/ubs-raises-gold-price-target-for-2026-here-are-the-new-forecasts-4474171",
      "kill": "Held if a weekly close at or above 6,200 printed on or before the last weekly close of June 2026; wrong if the window closed without one.",
      "note": "The second of the three windows. By the time it shut the same house had cut its year-end target from 5,900 to 5,500, on 27 May, which is its own row.",
      "grade": "miss",
      "result": "Wrong. The window's last weekly close was Friday 26 June 2026, 4,088.97 read off the founder's chart, 2,111.03 under the line. No weekly close at or above 6,200 printed inside the window: gold's high of 2026 is 5,589.38 on 28 January, as reported. The grade does not turn on the decimals.",
      "page": "ubs-gold-6200-2026-06-30.html"
    },
    {
      "published": "2026-01-29",
      "source": "UBS, as reported by Investing.com",
      "metal": "Gold",
      "level": 6200,
      "side": "above",
      "window": "2026-09-30",
      "windowText": "for September 2026",
      "quote": "UBS raised its XAU/USD price target to $6200 per ounce for March, June and September 2026, up from $5000.",
      "link": "https://www.investing.com/news/commodities-news/ubs-raises-gold-price-target-for-2026-here-are-the-new-forecasts-4474171",
      "kill": "Held if a weekly close at or above 6,200 prints on or before the last weekly close of September 2026, Friday 25 September; wrong if the window closes without one.",
      "note": "The third window, and the crowd row of 23 September. The last weekly close before it was named, Friday 18 September, was 4,377.290, 1,822.71 under the line. Grades on the founder's read of the 25 September close.",
      "grade": "miss",
      "result": "Wrong. The window's last weekly close was Friday 25 September 2026, 4,284.130 read off the founder's chart, 1,915.87 under the line. No weekly close at or above 6,200 printed inside the window: gold's high of 2026 on the founder's weekly chart is 5,598.75, in the week of 26 to 30 January. The grade does not turn on the decimals. See note 42.",
      "page": "ubs-gold-6200-2026-09-30.html"
    },
    {
      "published": "2026-02-25",
      "source": "J.P. Morgan Global Research",
      "level": 6300,
      "side": "above",
      "window": "2026-12-31",
      "windowText": "year-end 2026",
      "quote": "Gold price will rise 22% above current level to reach $6,300 by year-end 2026.",
      "link": "https://www.kitco.com/news/article/2026-02-25/gold-price-will-rise-22-above-current-level-reach-6300-year-end-2026-jp",
      "kill": "Held if a weekly close at or above 6,300 prints on or before the last weekly close of 2026; wrong if the year closes without one.",
      "note": "Quote is the report's headline. On 23 June 2026 the same house moved 6,300 to end-2027 and held about 6,000 for Q4 2026; that revision is its own row. This row stays as published.",
      "grade": "pending",
      "result": "",
      "page": "j-p-morgan-gold-6300-2026-12-31.html"
    },
    {
      "published": "2026-06-23",
      "source": "J.P. Morgan Global Research, as reported",
      "level": 6000,
      "side": "above",
      "window": "2026-12-31",
      "windowText": "Q4 2026",
      "quote": "Q4 2026: ~$6,000/oz. End of 2027: ~$6,300/oz potential.",
      "link": "https://www.canadianminingreport.com/blog/jpmorgan-s-latest-gold-forecast-what-the-world-s-largest-bank-is-telling-investors",
      "kill": "Held if a weekly close at or above 6,000 prints on or before the last weekly close of 2026; wrong if the year closes without one.",
      "note": "Reported second-hand; the desk will replace the link with the primary note when it has one. The 2026 average of 5,243 in the same note is not on the board: an average cannot be graded on a close.",
      "grade": "pending",
      "result": "",
      "page": "j-p-morgan-gold-6000-2026-12-31.html"
    },
    {
      "published": "2026-05-19",
      "source": "MKS PAMP",
      "level": 5800,
      "side": "above",
      "window": "2026-12-31",
      "windowText": "second half of 2026",
      "quote": "Gold is still expected to average $4,500/oz in 2026, with a new higher all-time high of $5,800/oz a fair target for the second half of the year.",
      "link": "https://www.kitco.com/news/article/2026-05-19/gold-will-hit-5800-ath-december-silver-has-highest-upside-platinum-has",
      "kill": "Held if a weekly close at or above 5,800 prints on or before the last weekly close of 2026; wrong if the year closes without one.",
      "note": "The 4,500 average in the same sentence is not on the board.",
      "grade": "pending",
      "result": "",
      "page": "mks-pamp-gold-5800-2026-12-31.html"
    },
    {
      "published": "2025-06-17",
      "source": "Citi Research",
      "level": 2700,
      "side": "below",
      "window": "2026-12-31",
      "windowText": "the second half of 2026",
      "quote": "Gold could return to around $2,500-$2,700/oz by the second half of 2026.",
      "link": "https://finance.yahoo.com/news/citi-sees-gold-below-3-091750805.html",
      "kill": "A range, graded on its nearer bound: held if a weekly close at or below 2,700 prints on or before the last weekly close of 2026; wrong if the second half closes without one.",
      "note": "Fifteen months old and still inside its window. The 4 September 2026 weekly close was 4,430.520. It is graded when the window closes, not before, because that is the rule.",
      "grade": "pending",
      "result": "",
      "page": "citi-gold-2700-2026-12-31.html"
    },
    {
      "published": "2026-01-13",
      "source": "Citi Research",
      "metal": "Silver",
      "level": 100,
      "side": "above",
      "window": "2026-03-31",
      "windowText": "0 to 3 months, by March",
      "quote": "Raised their 0-3 month target for gold to $5,000 per ounce and silver to $100 per ounce.",
      "link": "https://www.kitco.com/news/article/2026-01-13/gold-will-hit-5000oz-silver-100oz-march-gold-will-be-vulnerable-correction",
      "kill": "Held if a weekly close at or above 100 printed on or before the last weekly close of March 2026; wrong if the window closed without one.",
      "note": "The same note as the gold 5,000 row, graded on its own metal. Graded 9 September 2026 on a reported close; the founder's chart read replaces the figure when supplied.",
      "grade": "pass",
      "result": "Held. Silver crossed 100 for the first time on Friday 23 January 2026 and ended that week above 103.50 as reported, ten days after the target was published and inside the window. Where price went after the window is not the question the rule asks.",
      "page": "citi-silver-100-2026-03-31.html"
    },
    {
      "published": "2026-01-13",
      "source": "Citi Research",
      "metal": "Gold",
      "level": 5000,
      "side": "above",
      "window": "2026-03-31",
      "windowText": "0 to 3 months, by March",
      "quote": "Raised their 0-3 month target for gold to $5,000 per ounce and silver to $100 per ounce.",
      "link": "https://www.kitco.com/news/article/2026-01-13/gold-will-hit-5000oz-silver-100oz-march-gold-will-be-vulnerable-correction",
      "kill": "Held if a weekly close at or above 5,000 printed on or before the last weekly close of March 2026; wrong if the window closed without one.",
      "note": "Graded 8 September 2026, five months after the window shut, because the board did not exist until then. The desk's own weekly closes begin in June 2026; the February closes are as reported and the figure is replaced by the founder's chart read when supplied. The grade does not depend on the decimals.",
      "grade": "pass",
      "result": "Held. Spot gold first traded above 5,000 on 26 January 2026 and weekly closes above the line printed in February, inside the window: Friday 20 February closed above 5,000 and Friday 27 February printed 5,232.50 as reported. Where price went after the window is not the question the rule asks.",
      "page": "citi-gold-5000-2026-03-31.html"
    },
    {
      "published": "2026-02-16",
      "source": "ANZ, as reported by Kitco",
      "metal": "Gold",
      "level": 5800,
      "side": "above",
      "window": "2026-06-30",
      "windowText": "in the second quarter of 2026",
      "quote": "In their latest note on gold, commodity analysts at ANZ said they expect prices to hit $5,800 an ounce in the second quarter, up sharply from their previous target of $5,400 an ounce.",
      "link": "https://www.kitco.com/news/article/2026-02-16/anz-sees-gold-hitting-5800-ounce-second-quarter",
      "kill": "Held if a weekly close at or above 5,800 printed on or before the last weekly close of June 2026; wrong if the window closed without one.",
      "note": "Raised from a previous target of 5,400 in the same report, as reported; the 5,400 carries no window there and is listed below the board with that reason. The desk's own weekly closes begin in June 2026, and the window's last close is one of them.",
      "grade": "miss",
      "result": "Wrong. The window's last weekly close was Friday 26 June 2026, 4,088.97 read off the founder's chart, 1,711.03 under the line. No weekly close at or above 5,800 printed inside the window: gold's high of 2026 on the founder's weekly chart is 5,598.75, in the week of 26 to 30 January, before the target was published. The grade does not turn on the decimals.",
      "page": "anz-gold-5800-2026-06-30.html"
    }
  ],
  "targets_left_off": [
    [
      "31 Aug 2026",
      "TD Securities",
      "gold falling toward the lower end of the recent $4,200 to $4,700 trading range by year-end",
      "toward, and a range with no side stated; not a level"
    ],
    [
      "2 Sep 2026",
      "RBC",
      "gold will approach $5,000 in 2026, $5,300 in 2027",
      "approach is not a level"
    ],
    [
      "31 Aug 2026",
      "Crescat Capital",
      "gold price is going to $20,000",
      "no window"
    ],
    [
      "1 Sep 2026",
      "World Gold Council",
      "at risk of pullback to $4,215 as short-term momentum fades",
      "at risk of, and no window; a technical note, not a target"
    ],
    [
      "23 Jun 2026",
      "J.P. Morgan",
      "2026 average lowered to $5,243 from $5,708",
      "an average for a year cannot be graded on a close"
    ],
    [
      "8 Jul 2026",
      "Bank of America",
      "2026 average $4,360, cut from $5,093",
      "an average; primary source not yet in the desk's hands"
    ],
    [
      "Mar 2026",
      "UBS",
      "gold prices should rise toward USD 5,900 to 6,200/oz this year",
      "toward, and a range"
    ],
    [
      "29 Dec 2025",
      "UBS",
      "any increase in political or economic turmoil surrounding the U.S. midterm elections could push the yellow metal to $5,400",
      "could, and a condition"
    ],
    [
      "29 Dec 2025",
      "UBS",
      "expect the gold price to pull back to $4,800 per ounce by the end of 2026",
      "the return leg of a path; the first leg, 5,000 by September, is on the board"
    ],
    [
      "6 Jan 2026",
      "UBS",
      "moderate towards USD 4,800/oz by the end of 2026",
      "towards is not a level"
    ],
    [
      "29 Jan 2026",
      "UBS",
      "expects prices to ease modestly to $5,900 by the end of 2026, following the U.S. midterm elections",
      "the return leg of a path; the first leg, 6,200, is on the board as three rows"
    ],
    [
      "29 Jan 2026",
      "UBS",
      "an upside scenario of $7200 per ounce and a downside case of $4600",
      "two scenarios, two directions"
    ],
    [
      "16 Feb 2026",
      "ANZ",
      "up sharply from their previous target of $5,400 an ounce",
      "the number the target was raised from, reported without its window"
    ]
  ],
  "closes": [
    {
      "date": "2026-06-26",
      "Gold": "4,088.97"
    },
    {
      "date": "2026-07-03",
      "Gold": "4,174.76"
    },
    {
      "date": "2026-07-10",
      "Gold": "4,120.38",
      "Silver": "59.83"
    },
    {
      "date": "2026-07-17",
      "Gold": "4,018.44",
      "Silver": "55.911"
    },
    {
      "date": "2026-07-24",
      "Gold": "4,052.60",
      "Silver": "58.181"
    },
    {
      "date": "2026-07-31",
      "Gold": "4,041.18",
      "Silver": "57.564"
    },
    {
      "date": "2026-08-07",
      "Gold": "4,342",
      "Silver": "63.45"
    },
    {
      "date": "2026-08-14",
      "Gold": "4,376.160",
      "Silver": "64.7080"
    },
    {
      "date": "2026-08-21",
      "Gold": "4,602.660",
      "Silver": "68.9460"
    },
    {
      "date": "2026-08-28",
      "Gold": "4,453.670",
      "Silver": "66.3700",
      "Platinum": "1,821.80"
    },
    {
      "date": "2026-09-04",
      "Gold": "4,430.520",
      "Silver": "66.1940",
      "Platinum": "1,819.60"
    },
    {
      "date": "2026-09-11",
      "Gold": "4,347.760",
      "Silver": "64.4810",
      "Platinum": "1,791.50"
    },
    {
      "date": "2026-09-18",
      "Gold": "4,377.290",
      "Silver": "66.2210",
      "Platinum": "1,800.40"
    },
    {
      "date": "2026-09-25",
      "Gold": "4,284.130",
      "Silver": "64.2560",
      "Platinum": "1,777.80"
    },
    {
      "date": "2026-10-02",
      "Gold": "4,142.710",
      "Silver": "60.4470",
      "Platinum": "1,699.70"
    },
    {
      "date": "2026-10-09",
      "Gold": "4,194.085",
      "Silver": "60.7790",
      "Platinum": "1,686.90"
    }
  ],
  "crowd": {
    "note": "Claims made in public by others, quoted exactly, given the kill they were published without, graded on the same close. Not this desk's calls. Not in the tally. See note 18.",
    "rows": [
      {
        "id": "crow-0",
        "published": "2026-09-04",
        "source": "Citi Research, via Investing.com and Reuters",
        "quote": "The bank sets its 0-3 month point price target for gold at $4,800 per ounce and its 6-12 month forecast at $5,000 per ounce.",
        "link": "https://www.investing.com/news/commodities-news/citi-says-the-dip-in-gold-prices-is-a-buying-opportunity-93CH-4889056",
        "metal": "Gold",
        "kill_assigned": "Held if a weekly close at or above 4,800 prints on or before the 4 December 2026 close; wrong if the window closes without one. Assigned because the claim names a level and a window and no condition, so the level and the window are the only honest kill available. Graded on the TVC weekly close, as every row on this page, which is stated here because a bank's target is usually read as a print on any day and this page grades closes and nothing else.",
        "graded_on_close": "2026-12-04",
        "result": null,
        "grade": "pending",
        "grade_word": "open"
      },
      {
        "id": "crow-1",
        "published": "2026-01-29",
        "source": "UBS, as reported by Investing.com",
        "quote": "UBS raised its XAU/USD price target to $6200 per ounce for March, June and September 2026, up from $5000.",
        "link": "https://www.investing.com/news/commodities-news/ubs-raises-gold-price-target-for-2026-here-are-the-new-forecasts-4474171",
        "metal": "Gold",
        "kill_assigned": "Held if a weekly close at or above 6,200 prints on or before the 25 September 2026 close; wrong if the window closes without one. Assigned because the claim names a level and three windows and no condition, so the level and the last window are the only honest kill available. The March and June windows are graded on the bank board, each on its own close. Graded on the TVC weekly close, as every row on this page.",
        "graded_on_close": "2026-09-25",
        "result": "Wrong. The window shut on the close of Friday 25 September 2026, 4,284.130 read off the founder's chart, 1,915.87 under the line. No weekly close at or above 6,200 printed inside it: gold's high of 2026 on the founder's weekly chart is 5,598.75, in the week of 26 to 30 January. The same house's first number, 5,000 by September, is on the board as held. See note 42.",
        "grade": "miss",
        "grade_word": "wrong"
      },
      {
        "id": "crow-2",
        "published": "2026-02-16",
        "source": "ANZ, as reported by Kitco",
        "quote": "In their latest note on gold, commodity analysts at ANZ said they expect prices to hit $5,800 an ounce in the second quarter, up sharply from their previous target of $5,400 an ounce.",
        "link": "https://www.kitco.com/news/article/2026-02-16/anz-sees-gold-hitting-5800-ounce-second-quarter",
        "metal": "Gold",
        "kill_assigned": "Held if a weekly close at or above 5,800 printed on or before the 26 June 2026 close; wrong if the window closed without one. Assigned because the claim names a level and a window and no condition, so the level and the window are the only honest kill available. Graded on the TVC weekly close, as every row on this page.",
        "graded_on_close": "2026-06-26",
        "result": "Wrong. The window shut on the close of Friday 26 June 2026, 4,088.97 read off the founder's chart, 1,711.03 under the line. No weekly close at or above 5,800 printed inside it: gold's high of 2026 on the founder's weekly chart is 5,598.75, in the week of 26 to 30 January, before the target was published. See note 45.",
        "grade": "miss",
        "grade_word": "wrong"
      }
    ]
  },
  "guest": {
    "note": "Levels submitted by others, graded by this desk's standard on the same close. Eight per week. Not in the tally. See note 18.",
    "rows": []
  },
  "tick_book": {
    "note": "Intraday calls on the five-minute clock, named and posted before the trade, graded on the 23:00 Madrid close of the day named. Never in the weekly tally. See tick.html.",
    "tally": {
      "held": 0,
      "wrong": 0,
      "never_reached": 0,
      "open": 0
    },
    "rows": []
  }
}
