Colour scheme: system

Which numbers on this site recompute

Every number on the pages of this record, in page order, with what it means and what kind of number it is. The list is measured from the pages every time the site is built, and the build stops if it no longer matches them.

A number recomputes only if a test recomputes it from a published file every time the site is built and the sentence that prints it reads it from that test’s source. Every other number says what it is instead.

The record

176 entries in all. 81 are the numbers this page tells you it has, each tied to its source: 59 recompute from the published files, 12 are read from the summary file, 2 are choices of the analysis and 8 come from my unpublished capture archive. The other 95: 68 are not measurements (dates, section numbers, ids, rule numbers, word counts), 24 are not bound to a source on the page yet and 3 are drawings, one entry each.

  1. 18 · Opening · The questions put to the engines, all about one company. · recomputes from the published files
  2. 4 July · Opening · “…whose first captures fall between 4 July and 25 August…” · not a measurement
  3. 25 August 2026 · Opening · “…first captures fall between 4 July and 25 August 2026. The questions are…” · not a measurement
  4. 81 · Opening · “…Write to denes@denescsaszar.com. 81 numbers in the…” · not a measurement
  5. 59 · Opening · “…81 numbers in the text of this page, 59 of them can be…” · not a measurement
  6. 1 · Opening · “…or click any number. Map of this page 1 Change measured…” · not a measurement
  7. 338 words · Opening · “…against each cell’s own noise floor 338 words 2 Wording, not…” · not a measurement
  8. 2 · Opening · “…cell’s own noise floor 338 words 2 Wording, not…” · not a measurement
  9. 176 words · Opening · “…exploratory, not pre-registered 176 words 3 The counts…” · not a measurement
  10. 3 · Opening · “…not pre-registered 176 words 3 The counts 571…” · not a measurement
  11. 571 words · Opening · “…176 words 3 The counts 571 words 4 The floor test…” · not a measurement
  12. 4 · Opening · “…176 words 3 The counts 571 words 4 The floor test…” · not a measurement
  13. 570 words · Opening · “…counts 571 words 4 The floor test 570 words 5 What cannot be…” · not a measurement
  14. 5 · Opening · “…words 4 The floor test 570 words 5 What cannot be…” · not a measurement
  15. 184 words · Opening · “…cannot be checked about the instrument 184 words 6 The verifier…” · not a measurement
  16. 6 · Opening · “…about the instrument 184 words 6 The verifier 173…” · not a measurement
  17. 173 words · Opening · “…184 words 6 The verifier 173 words 7 The method…” · not a measurement
  18. 7 · Opening · “…words 6 The verifier 173 words 7 The method 107…” · not a measurement
  19. 107 words · Opening · “…verifier 173 words 7 The method 107 words 8 The analysis…” · not a measurement
  20. 8 · Opening · “…173 words 7 The method 107 words 8 The analysis tools…” · not a measurement
  21. 79 words · Opening · “…107 words 8 The analysis tools 79 words 9 So does it…” · not a measurement
  22. 9 · Opening · “…8 The analysis tools 79 words 9 So does it tell us…” · not a measurement
  23. 612 words · Opening · “…9 So does it tell us anything? 612 words 10 Colophon 52…” · not a measurement
  24. 10 · Opening · “…does it tell us anything? 612 words 10 Colophon 52…” · not a measurement
  25. 52 words · Opening · “…us anything? 612 words 10 Colophon 52 words This page’s…” · not a measurement
  26. Opening · “…This page’s numbers, sourced.…” · a drawing; its numbers are not listed one by one
  27. 1 · 1 · Change measured against each cell’s own noise floor · “…1 · Change measured…” · not a measurement
  28. 103 · 1 · Change measured against each cell’s own noise floor · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  29. 85 · 1 · Change measured against each cell’s own noise floor · Cells whose answers changed between waves by more than they vary on their own inside one sitting. · recomputes from the published files
  30. 103 · 1 · Change measured against each cell’s own noise floor · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  31. 33 · 1 · Change measured against each cell’s own noise floor · Of the cells whose floor is measured in both waves, the ones that moved beyond their own floor. · recomputes from the published files
  32. 43 · 1 · Change measured against each cell’s own noise floor · Cells whose floor is measured in both the earlier and the later wave, with at least two answers in each. · recomputes from the published files
  33. 52 · 1 · Change measured against each cell’s own noise floor · Of the cells whose floor rests on two or more pairs, the ones that moved beyond their own floor. · recomputes from the published files
  34. 65 · 1 · Change measured against each cell’s own noise floor · Cells whose floor rests on at least two pairs of repeat answers, so it is less of a one-off. · recomputes from the published files
  35. 15 July 2026 · 1 · Change measured against each cell’s own noise floor · “…one documented exception, mistral on 15 July 2026, and 768 rows…” · not a measurement
  36. 768 · 1 · Change measured against each cell’s own noise floor · Rows in the whole capture archive that carry no record of how they were captured. · from my unpublished capture archive
  37. 2 · 2 · Wording, not agreement; exploratory, not pre-registered · “…2 · Wording, not…” · not a measurement
  38. 17 September 2026 · 2 · Wording, not agreement; exploratory, not pre-registered · “…captured by the same person. Corrected 17 September 2026. Sign-in state was…” · not a measurement
  39. 3 September 2026 · 2 · Wording, not agreement; exploratory, not pre-registered · “…variable is not a control. Corrected 3 September 2026. The analysis was…” · not a measurement
  40. 4 July 2026 · 2 · Wording, not agreement; exploratory, not pre-registered · “…were, in stages: the first ten on 4 July 2026 and the full…” · not a measurement
  41. 9 July · 2 · Wording, not agreement; exploratory, not pre-registered · “…4 July 2026 and the full eighteen on 9 July; the floor test…” · not a measurement
  42. 22 August · 2 · Wording, not agreement; exploratory, not pre-registered · “…the floor test itself was written on 22 August, after six of the…” · not a measurement
  43. 3 · 3 · The counts · “…3 · The counts The…” · not a measurement
  44. 0.047 · 3 · The counts · Inside one cell: how much less alike the answer is at the longer gap than at the shorter one. The typical (median) difference, on the 0 to 1 scale. · recomputes from the published files
  45. 87 · 3 · The counts · Cells where the answer at the longer gap is less alike than at the shorter gap. · recomputes from the published files
  46. 108 · 3 · The counts · Cells measured at two or more different time gaps, so a longer and a shorter gap can be compared. · recomputes from the published files
  47. 37 · 3 · The counts · Cells where two pairs sit at the same gap in days, so which one counts as the shortest or the longest is a choice. · recomputes from the published files
  48. 108 · 3 · The counts · Cells measured at two or more different time gaps, so a longer and a shorter gap can be compared. · recomputes from the published files
  49. −0.042 · 3 · The counts · The mildest result across five reasonable ways of settling those ties. · recomputes from the published files
  50. −0.066 · 3 · The counts · The strongest result across those five ways of settling the ties. · recomputes from the published files
  51. 83 · 3 · The counts · Under the five tie-break rules, at least this many cells are less alike at the longer gap. · recomputes from the published files
  52. 93 · 3 · The counts · Under the five tie-break rules, at most this many cells are less alike at the longer gap. · recomputes from the published files
  53. 108 · 3 · The counts · Cells measured at two or more different time gaps, so a longer and a shorter gap can be compared. · recomputes from the published files
  54. 3 · The counts · “…Every cell against its own noise floor. One row per cell, 103 rows, sorted by that cell’s own floor. Each row runs from the floor to the between-wave similarity. 85 run left, 18 run right. The floor is not one threshold: it runs from 0.036 to 0.600, which is wider than most of the differences it judges.…” · a drawing; its numbers are not listed one by one
  55. section 1 · 3 · The counts · “…counting a column. The two subsets in section 1 are not rows…” · not a measurement
  56. at 2 or more · 3 · The counts · “…is the cells CSV’s within_pairs column at 2 or more. Waves 7…” · not a measurement
  57. 7 · 3 · The counts · Capture waves: separate sittings in which the questions were asked again. Not every wave holds every question. · recomputes from the published files
  58. 2026-07-04 · 3 · The counts · “…column at 2 or more. Waves 7 2026-07-04 to 2026-08-25,…” · not a measurement
  59. 2026-08-25 · 3 · The counts · “…at 2 or more. Waves 7 2026-07-04 to 2026-08-25, first captures…” · not a measurement
  60. 18 · 3 · The counts · The questions put to the engines, all about one company. · recomputes from the published files
  61. 111 · 3 · The counts · Question-and-engine pairs (cells) that could be compared between waves, with or without a noise floor. · recomputes from the published files
  62. 103 · 3 · The counts · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  63. 12 · 3 · The counts · Cells left out because the wording of the question changed between waves. · read from the summary file
  64. 0.266 · 3 · The counts · The typical noise floor: how alike two answers to the same question are inside one sitting, from 0 (no shared words) to 1 (the same words). · recomputes from the published files
  65. 0.200 · 3 · The counts · A quarter of the cells have a noise floor below this. · recomputes from the published files
  66. 0.321 · 3 · The counts · A quarter of the cells have a noise floor above this. · recomputes from the published files
  67. 0.183 · 3 · The counts · The typical similarity between answers from different waves, on the same 0 to 1 scale. Lower than the floor: the answers moved. · recomputes from the published files
  68. −0.065 · 3 · The counts · The typical change: similarity between waves minus the cell's own floor. Negative means less alike than noise alone would make them. · recomputes from the published files
  69. 95 · 3 · The counts · The level the uncertainty ranges are stated at: a choice of the analysis, not a value any published file holds. · a choice of the analysis
  70. −0.092 · 3 · The counts · Low end of the uncertainty range around the typical change. · read from the summary file
  71. −0.030 · 3 · The counts · High end of the uncertainty range around the typical change. · read from the summary file
  72. 18 · 3 · The counts · The resampling treats each question as one unit, so answers to the same question are not counted as independent evidence. · read from the summary file
  73. 2,000 · 3 · The counts · A setting of the analysis, as the summary file records it: how many times the data were resampled to get the range. · read from the summary file
  74. 0 · 3 · The counts · A setting of the analysis, as the summary file records it: the random seed, so the same resampling can be run again exactly. · read from the summary file
  75. 85 · 3 · The counts · Cells whose answers changed between waves by more than they vary on their own inside one sitting. · recomputes from the published files
  76. 103 · 3 · The counts · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  77. 40.9 · 3 · The counts · If cells moved only by chance, about this share of them would show as moved. · read from the summary file
  78. 5,000 · 3 · The counts · A setting of the analysis, as the summary file records it: how many random reshuffles the chance test ran. · read from the summary file
  79. 0 · 3 · The counts · A setting of the analysis, as the summary file records it: the random seed, so the same resampling can be run again exactly. · read from the summary file
  80. 5,000 · 3 · The counts · A setting of the analysis, as the summary file records it: how many random reshuffles the chance test ran. · read from the summary file
  81. −0.047 · 3 · The counts · Inside one cell: how much less alike the answer is at the longer gap than at the shorter one. The typical (median) difference, on the 0 to 1 scale. · recomputes from the published files
  82. 108 · 3 · The counts · Cells measured at two or more different time gaps, so a longer and a shorter gap can be compared. · recomputes from the published files
  83. 103 · 3 · The counts · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  84. 111 · 3 · The counts · Question-and-engine pairs (cells) that could be compared between waves, with or without a noise floor. · recomputes from the published files
  85. 95 · 3 · The counts · The level the uncertainty ranges are stated at: a choice of the analysis, not a value any published file holds. · a choice of the analysis
  86. −0.058 · 3 · The counts · Low end of the uncertainty range around the within-cell difference. · read from the summary file
  87. −0.031 · 3 · The counts · High end of the uncertainty range around the within-cell difference. · read from the summary file
  88. 87 · 3 · The counts · Cells where the answer at the longer gap is less alike than at the shorter gap. · recomputes from the published files
  89. 108 · 3 · The counts · Cells measured at two or more different time gaps, so a longer and a shorter gap can be compared. · recomputes from the published files
  90. 21 · 3 · The counts · Cells where the answer at the longer gap is more alike: the opposite direction. · recomputes from the published files
  91. −0.042 · 3 · The counts · The mildest result across five reasonable ways of settling those ties. · recomputes from the published files
  92. −0.066 · 3 · The counts · The strongest result across those five ways of settling the ties. · recomputes from the published files
  93. 83 · 3 · The counts · Under the five tie-break rules, at least this many cells are less alike at the longer gap. · recomputes from the published files
  94. 93 · 3 · The counts · Under the five tie-break rules, at most this many cells are less alike at the longer gap. · recomputes from the published files
  95. 108 · 3 · The counts · Cells measured at two or more different time gaps, so a longer and a shorter gap can be compared. · recomputes from the published files
  96. 18 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  97. 5 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  98. 111 · 3 · The counts · Question-and-engine pairs (cells) that could be compared between waves, with or without a noise floor. · recomputes from the published files
  99. 8 · 3 · The counts · Cells with no noise floor, because they had too few repeat answers. They stay in the file with the reason. · recomputes from the published files
  100. 103 · 3 · The counts · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  101. 103 · 3 · The counts · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  102. 18 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  103. 16 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  104. 16 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  105. 16 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  106. 16 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  107. 12 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  108. 12 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  109. 5 · 3 · The counts · Comparable cells for this engine, over all the rows of the cells file. A count of what happened to be captured, not a comparison between engines. · recomputes from the published files
  110. 4 · 4 · The floor test · “…4 · The floor test…” · not a measurement
  111. rather than 1 · 4 · The floor test · “…empty the comparison returns nothing rather than 1, because two draws…” · not a measurement
  112. 0 · 4 · The floor test · The fewest pairs of repeat answers in any row of the cells file. A cell with none has no floor and is left out. · recomputes from the published files
  113. 18 · 4 · The floor test · The most pairs of repeat answers behind any cell's floor. · recomputes from the published files
  114. 9,692 · 4 · The floor test · Separate answers in the capture archive, counted under one de-duplication key. · from my unpublished capture archive
  115. 28 · 4 · The floor test · Capture runs those answers fall in. · from my unpublished capture archive
  116. 29 August 2026 · 4 · The floor test · “…across 28 runs when I read it on 29 August 2026; a count I ran the…” · not a measurement
  117. 5,214 · 4 · The floor test · The same count a day earlier, under a different key, on a smaller archive. · from my unpublished capture archive
  118. 27 · 4 · The floor test · Capture runs in that smaller archive. · from my unpublished capture archive
  119. 6 · 4 · The floor test · “…instead of being implied by a number. 6 files CC BY 4.0 ·…” · not bound to a source on the page yet
  120. CC BY 4.0 · 4 · The floor test · “…of being implied by a number. 6 files CC BY 4.0 · 273 KB ·…” · not a measurement
  121. 273 · 4 · The floor test · “…by a number. 6 files CC BY 4.0 · 273 KB · checksummed…” · not bound to a source on the page yet
  122. 08284347 · 4 · The floor test · “…273 KB · checksummed floor-test-cells-08284347.csv one row per…” · not a measurement
  123. 13 · 4 · The floor test · “…row per question-by-engine cell size 13 KB contents 111…” · not bound to a source on the page yet
  124. 111 · 4 · The floor test · “…cell size 13 KB contents 111 rows sha256…” · not bound to a source on the page yet
  125. 08284347 · 4 · The floor test · “…floor-test-gap-pairs-08284347.csv one row per…” · not a measurement
  126. 81 · 4 · The floor test · “…what the within-cell test needs size 81 KB contents 957…” · not bound to a source on the page yet
  127. 957 · 4 · The floor test · “…test needs size 81 KB contents 957 rows sha256…” · not bound to a source on the page yet
  128. 08284347 · 4 · The floor test · “…floor-test-summary-08284347.json the aggregate…” · not a measurement
  129. 11 · 4 · The floor test · “…every seed and replicate count size 11 KB contents 24…” · not bound to a source on the page yet
  130. 24 · 4 · The floor test · “…replicate count size 11 KB contents 24 top-level keys…” · not bound to a source on the page yet
  131. 08284347 · 4 · The floor test · “…codebook-08284347.json the codebook:…” · not a measurement
  132. 17 · 4 · The floor test · “…README's words and allowed values size 17 KB contents both…” · not bound to a source on the page yet
  133. 08284347 · 4 · The floor test · “…floor-test-cells-08284347.sav the same table…” · not a measurement
  134. 24 · 4 · The floor test · “…for SPSS, generated from the CSV size 24 KB contents 111…” · not bound to a source on the page yet
  135. 111 · 4 · The floor test · “…from the CSV size 24 KB contents 111 rows sha256…” · not bound to a source on the page yet
  136. 08284347 · 4 · The floor test · “…floor-test-gap-pairs-08284347.sav the same table…” · not a measurement
  137. 125 · 4 · The floor test · “…for SPSS, generated from the CSV size 125 KB contents 957…” · not bound to a source on the page yet
  138. 957 · 4 · The floor test · “…from the CSV size 125 KB contents 957 rows sha256…” · not bound to a source on the page yet
  139. 1 September 2026 · 4 · The floor test · “…were rederived from the cells file on 1 September 2026 with all fourteen…” · not a measurement
  140. 3 September · 4 · The floor test · “…rederived from the gap-pairs file on 3 September. The summary…” · not a measurement
  141. 4 · The floor test · “…This page’s numbers, sourced.…” · a drawing; its numbers are not listed one by one
  142. 5 · 5 · What cannot be checked about the instrument · “…5 · What cannot be…” · not a measurement
  143. 26 August 2026 · 5 · What cannot be checked about the instrument · “…holds exactly one commit, dated 26 August 2026; everything before…” · not a measurement
  144. 4 July · 5 · What cannot be checked about the instrument · “…show from here that the instrument on 4 July was the instrument…” · not a measurement
  145. 25 August · 5 · What cannot be checked about the instrument · “…on 4 July was the instrument on 25 August. That is a real…” · not a measurement
  146. 6 · 6 · The verifier · “…6 · The verifier I run…” · not a measurement
  147. 29 August · 6 · The verifier · “…against my own archive. On the 29 August read, build…” · not a measurement
  148. 2136073e · 6 · The verifier · “…archive. On the 29 August read, build 2136073e, it ran eleven…” · not a measurement
  149. 2136073e · 6 · The verifier · “…since. Verifier output, build 2136073e. 4 of 11 checks…” · not a measurement
  150. 4 · 6 · The verifier · “…Verifier output, build 2136073e. 4 of 11 checks fail.…” · not bound to a source on the page yet
  151. 11 · 6 · The verifier · “…Verifier output, build 2136073e. 4 of 11 checks fail.…” · not bound to a source on the page yet
  152. 1,257 · 6 · The verifier · “…FAIL ambiguous picture references 1,257 FAIL duplicate…” · not bound to a source on the page yet
  153. 8 · 6 · The verifier · “…references 1,257 FAIL duplicate frames 8 FAIL cross-run…” · not bound to a source on the page yet
  154. 26 · 6 · The verifier · “…frames 8 FAIL cross-run references 26 FAIL coverage more…” · not bound to a source on the page yet
  155. 26 · 6 · The verifier · “…allows pass unkeyed observations 26, no cell address,…” · not bound to a source on the page yet
  156. 3,686 · 6 · The verifier · “…pictures re-hash to recorded values 3,686 checked, 0…” · not bound to a source on the page yet
  157. 0 · 6 · The verifier · “…to recorded values 3,686 checked, 0 mismatches pass…” · not bound to a source on the page yet
  158. 9,692 · 6 · The verifier · “…answers re-hash to recorded values 9,692 checked, 0…” · not bound to a source on the page yet
  159. 0 · 6 · The verifier · “…to recorded values 9,692 checked, 0 mismatches pass…” · not bound to a source on the page yet
  160. 0 · 6 · The verifier · “…pass duplicate observation ids 0 pass proof floor…” · not bound to a source on the page yet
  161. 26 · 6 · The verifier · Pictures recorded under one capture run and referenced from another. · from my unpublished capture archive
  162. f0dd057a · 6 · The verifier · “…number on this site. An older build, f0dd057a, failed three of…” · not a measurement
  163. 7.2 · 6 · The verifier · Coverage an older build of the verifier reported, counted a different way, so not comparable. · from my unpublished capture archive
  164. 24.1 · 6 · The verifier · Coverage this build of the verifier reports, counted over draw slots. · from my unpublished capture archive
  165. 7 · 7 · The method · “…7 · The method The…” · not a measurement
  166. 30 August 2026 · 7 · The method · “…carries no date anywhere, and until 30 August 2026 not one of them…” · not a measurement
  167. 8 · 8 · The analysis tools · “…8 · The analysis tools…” · not a measurement
  168. 9 · 9 · So does it tell us anything? · “…9 · So does it tell us…” · not a measurement
  169. 85 · 9 · So does it tell us anything? · Cells whose answers changed between waves by more than they vary on their own inside one sitting. · recomputes from the published files
  170. 103 · 9 · So does it tell us anything? · Cells with enough repeat answers inside one sitting to measure how much their answers vary on their own: their noise floor. · recomputes from the published files
  171. 30 August 2026 · 9 · So does it tell us anything? · “…It is the paraphrase study written on 30 August 2026, still a draft…” · not a measurement
  172. 31 October 2026 · 9 · So does it tell us anything? · “…files and which do not. Done on 31 October 2026: the study is…” · not a measurement
  173. 10.5281/zenodo.23780961 · 9 · So does it tell us anything? · “…the study is archived on Zenodo as 10.5281/zenodo.23780961, and the list is a…” · not a measurement
  174. 25 August 2026 · 9 · So does it tell us anything? · “…of the last wave, which began on 25 August 2026, were lost with a…” · not a measurement
  175. 10 · 10 · Colophon · “…10 · Colophon My own…” · not a measurement
  176. 31 October 2026 · 10 · Colophon · “…monospace, no analytics, last revised 31 October 2026. What changed here…” · not a measurement

↑ top

The method

51 entries in all. None of them is tied to a source yet. All 51: 36 are not measurements (dates, section numbers, ids, rule numbers, word counts), 15 are not bound to a source on the page yet and 0 are drawings, one entry each.

  1. 1 · Opening · “…Back to the record. Map of this page 1 What is measured…” · not a measurement
  2. 181 words · Opening · “…Map of this page 1 What is measured 181 words 2 The seven…” · not a measurement
  3. 2 · Opening · “…page 1 What is measured 181 words 2 The seven…” · not a measurement
  4. 256 words · Opening · “…181 words 2 The seven dimensions 256 words 3 The run…” · not a measurement
  5. 3 · Opening · “…2 The seven dimensions 256 words 3 The run protocol…” · not a measurement
  6. 412 words · Opening · “…256 words 3 The run protocol 412 words 4 Registration,…” · not a measurement
  7. 4 · Opening · “…3 The run protocol 412 words 4 Registration, and…” · not a measurement
  8. 388 words · Opening · “…and the half of it that was missing 388 words 5 What a failed…” · not a measurement
  9. 5 · Opening · “…of it that was missing 388 words 5 What a failed…” · not a measurement
  10. 546 words · Opening · “…words 5 What a failed check costs 546 words 6 What is…” · not a measurement
  11. 6 · Opening · “…What a failed check costs 546 words 6 What is…” · not a measurement
  12. 219 words · Opening · “…6 What is deliberately not here 219 words Pages of this site…” · not a measurement
  13. 1 · 1 · What is measured · “…1 · What is measured…” · not a measurement
  14. 2 · 2 · The seven dimensions · “…2 · The seven…” · not a measurement
  15. 3 July 2026 · 2 · The seven dimensions · “…by an independent methodology gate on 3 July 2026: it averaged the…” · not a measurement
  16. 3 · 3 · The run protocol · “…3 · The run protocol…” · not a measurement
  17. 31 August 2026 · 3 · The run protocol · “…particular mode. It was replaced on 31 August 2026 by a field that…” · not a measurement
  18. 15 July 2026 · 3 · The run protocol · “…moving from one mode to another on 15 July 2026, and 768 rows…” · not a measurement
  19. 768 · 3 · The run protocol · “…mode to another on 15 July 2026, and 768 rows carry no mode…” · not bound to a source on the page yet
  20. 15,171 · 3 · The run protocol · “…rows carry no mode at all. Across all 15,171 rows of the…” · not bound to a source on the page yet
  21. 4 · 4 · Registration, and the half of it that was missing · “…4 · Registration, and…” · not a measurement
  22. 31 August 2026 · 4 · Registration, and the half of it that was missing · “…Counted in the working repository on 31 August 2026, by the rule that…” · not a measurement
  23. 30 June 2026 · 4 · Registration, and the half of it that was missing · “…such files exist, the earliest dated 30 June 2026 and the most…” · not a measurement
  24. 30 August 2026 · 4 · Registration, and the half of it that was missing · “…2026 and the most recent two dated 30 August 2026. Exactly one…” · not a measurement
  25. 4 July 2026 · 4 · Registration, and the half of it that was missing · “…that run’s first capture: ten on 4 July 2026, fourteen on 5…” · not a measurement
  26. 5 July · 4 · Registration, and the half of it that was missing · “…ten on 4 July 2026, fourteen on 5 July, all eighteen on 9…” · not a measurement
  27. 9 July · 4 · Registration, and the half of it that was missing · “…fourteen on 5 July, all eighteen on 9 July. Two questions…” · not a measurement
  28. 4 July · 4 · Registration, and the half of it that was missing · “…on 9 July. Two questions asked on 4 July, the two that tell…” · not a measurement
  29. 22 August · 4 · Registration, and the half of it that was missing · “…the floor test, was written on 22 August, after six of the…” · not a measurement
  30. 30 August 2026 · 4 · Registration, and the half of it that was missing · “…at all. The first pair was written on 30 August 2026. Two of the eleven…” · not a measurement
  31. 30 August · 4 · Registration, and the half of it that was missing · “…Two of the eleven are themselves dated 30 August, so before that…” · not a measurement
  32. 5 · 5 · What a failed check costs · “…5 · What a failed check…” · not a measurement
  33. 29 August 2026 · 5 · What a failed check costs · “…A verifier ran against the archive on 29 August 2026 and failed it on…” · not a measurement
  34. 1,257 · 5 · What a failed check costs · “…Ambiguous picture references, at 1,257, and duplicate…” · not bound to a source on the page yet
  35. 8 · 5 · What a failed check costs · “…at 1,257, and duplicate frames, at 8, mean a screenshot…” · not bound to a source on the page yet
  36. 26 · 5 · What a failed check costs · “…were passing. Cross-run references, at 26, is the same layer…” · not bound to a source on the page yet
  37. 31 August 2026 · 5 · What a failed check costs · “…a published number was determined on 31 August 2026, and the answer is…” · not a measurement
  38. 103 · 5 · What a failed check costs · “…and the answer is yes. Twelve of the 103 published cells…” · not bound to a source on the page yet
  39. 82.5 · 5 · What a failed check costs · “…moved beyond their own floor goes from 82.5 percent to 85.7…” · not bound to a source on the page yet
  40. 85.7 · 5 · What a failed check costs · “…own floor goes from 82.5 percent to 85.7 percent, because…” · not bound to a source on the page yet
  41. 58.3 · 5 · What a failed check costs · “…percent, because those twelve move at 58.3 percent, well…” · not bound to a source on the page yet
  42. −0.065 · 5 · What a failed check costs · “…The median difference does not move: −0.065 with the twelve in…” · not bound to a source on the page yet
  43. −0.065 · 5 · What a failed check costs · “…move: −0.065 with the twelve in and −0.065 without them. The…” · not bound to a source on the page yet
  44. 0.0007 · 5 · What a failed check costs · “…−0.065 without them. The two differ by 0.0007. The interval on…” · not bound to a source on the page yet
  45. 0.062 · 5 · What a failed check costs · “…0.0007. The interval on that median is 0.062 wide, about ninety…” · not bound to a source on the page yet
  46. 2 September 2026 · 5 · What a failed check costs · “…shift this data can resolve. Corrected 2 September 2026. So the deviation…” · not a measurement
  47. 5 · 5 · What a failed check costs · “…it belongs, in the study, section 5. How a withdrawal…” · not bound to a source on the page yet
  48. 6 · 6 · What is deliberately not here · “…6 · What is…” · not a measurement
  49. 4 · 6 · What is deliberately not here · “…are public and they are in section 4: the earliest and…” · not bound to a source on the page yet
  50. 30 August 2026 · 6 · What is deliberately not here · “…The paraphrase study written on 30 August 2026, still a draft…” · not a measurement
  51. 31 October 2026 · 6 · What is deliberately not here · “…this page said before. Last revised 31 October 2026.…” · not a measurement

↑ top

How the result changed

71 entries in all. None of them is tied to a source yet. All 71: 37 are not measurements (dates, section numbers, ids, rule numbers, word counts), 34 are not bound to a source on the page yet and 0 are drawings, one entry each.

  1. 3 September 2026 · Opening · “…source on the study page. Corrected 3 September 2026. Map of this page…” · not a measurement
  2. 1 · Opening · “…3 September 2026. Map of this page 1 What I wanted to…” · not a measurement
  3. 68 words · Opening · “…this page 1 What I wanted to measure 68 words 2 What I first…” · not a measurement
  4. 2 · Opening · “…What I wanted to measure 68 words 2 What I first built…” · not a measurement
  5. 92 words · Opening · “…68 words 2 What I first built 92 words 3 What running…” · not a measurement
  6. 3 · Opening · “…2 What I first built 92 words 3 What running it…” · not a measurement
  7. 129 words · Opening · “…words 3 What running it produced 129 words 4 Where the…” · not a measurement
  8. 4 · Opening · “…What running it produced 129 words 4 Where the problems…” · not a measurement
  9. 351 words · Opening · “…words 4 Where the problems were 351 words 5 What changed,…” · not a measurement
  10. 5 · Opening · “…Where the problems were 351 words 5 What changed, and…” · not a measurement
  11. 261 words · Opening · “…What changed, and what justified it 261 words 6 What survived…” · not a measurement
  12. 6 · Opening · “…and what justified it 261 words 6 What survived 383…” · not a measurement
  13. 383 words · Opening · “…it 261 words 6 What survived 383 words 7 What the next…” · not a measurement
  14. 7 · Opening · “…words 6 What survived 383 words 7 What the next…” · not a measurement
  15. 282 words · Opening · “…7 What the next study inherits 282 words Pages of this site…” · not a measurement
  16. 1 · 1 · What I wanted to measure · “…1 · What I wanted to…” · not a measurement
  17. 4 July 2026 · 1 · What I wanted to measure · “…first ten questions were registered on 4 July 2026, before the first…” · not a measurement
  18. 9 July · 1 · What I wanted to measure · “…the full eighteen were registered on 9 July.…” · not a measurement
  19. 2 · 2 · What I first built · “…2 · What I first built…” · not a measurement
  20. 3 · 3 · What running it produced · “…3 · What running it…” · not a measurement
  21. 4 July · 3 · What running it produced · “…whose first captures fall between 4 July and 25 August…” · not a measurement
  22. 25 August 2026 · 3 · What running it produced · “…first captures fall between 4 July and 25 August 2026, giving 111…” · not a measurement
  23. 111 · 3 · What running it produced · “…4 July and 25 August 2026, giving 111 question-by-engine…” · not bound to a source on the page yet
  24. 0.183 · 3 · What running it produced · “…Median similarity between waves: 0.183. Read against the…” · not bound to a source on the page yet
  25. 29 August 2026 · 3 · What running it produced · “…A verifier run against the archive on 29 August 2026, build 2136073e,…” · not a measurement
  26. 2136073e · 3 · What running it produced · “…the archive on 29 August 2026, build 2136073e, failed it on four…” · not a measurement
  27. 4 · 4 · Where the problems were · “…4 · Where the problems…” · not a measurement
  28. 4 July · 4 · Where the problems were · “…The questions were registered between 4 July and 9 July; the…” · not a measurement
  29. 9 July · 4 · Where the problems were · “…were registered between 4 July and 9 July; the analysis that…” · not a measurement
  30. 22 August · 4 · Where the problems were · “…that ran over them was written on 22 August. Nothing had been…” · not a measurement
  31. 103 · 4 · Where the problems were · “…rest on a single repeat pair. Of the 103 cells with a…” · not bound to a source on the page yet
  32. 38 · 4 · Where the problems were · “…the 103 cells with a measurable floor, 38 have a floor that…” · not bound to a source on the page yet
  33. 60 · 4 · Where the problems were · “…median of exactly one repeat pair, and 60 draw their floor…” · not bound to a source on the page yet
  34. 57.7 · 4 · Where the problems were · “…rank blocks, the share that moved runs 57.7, 84.6, 92.0 and…” · not bound to a source on the page yet
  35. 84.6 · 4 · Where the problems were · “…the share that moved runs 57.7, 84.6, 92.0 and 96.2…” · not bound to a source on the page yet
  36. 92.0 · 4 · Where the problems were · “…the share that moved runs 57.7, 84.6, 92.0 and 96.2 percent.…” · not bound to a source on the page yet
  37. 96.2 · 4 · Where the problems were · “…that moved runs 57.7, 84.6, 92.0 and 96.2 percent. The…” · not bound to a source on the page yet
  38. 31 August · 4 · Where the problems were · “…Checked against the live surface on 31 August:…” · not a measurement
  39. 5 · 5 · What changed, and what justified it · “…5 · What changed, and…” · not a measurement
  40. 103 · 5 · What changed, and what justified it · “…checked rather than removed. Twelve of 103 cells carried more…” · not bound to a source on the page yet
  41. 82.5 · 5 · What changed, and what justified it · “…those twelve raises the figure from 82.5 to 85.7 percent.…” · not bound to a source on the page yet
  42. 85.7 · 5 · What changed, and what justified it · “…twelve raises the figure from 82.5 to 85.7 percent.…” · not bound to a source on the page yet
  43. 17 September 2026 · 5 · What changed, and what justified it · “…makes the criterion work. Corrected 17 September 2026. correction The…” · not a measurement
  44. 4 · 5 · What changed, and what justified it · “…because of the gap named in station 4.…” · not bound to a source on the page yet
  45. 6 · 6 · What survived · “…6 · What survived…” · not a measurement
  46. 85 · 6 · What survived · “…observed Of the cells with a floor, 85 of 103 moved…” · not bound to a source on the page yet
  47. 103 · 6 · What survived · “…Of the cells with a floor, 85 of 103 moved beyond their…” · not bound to a source on the page yet
  48. 40.9 · 6 · What survived · “…beyond their own noise floor, against 40.9 percent expected…” · not bound to a source on the page yet
  49. 5,000 · 6 · What survived · “…permutation null over shuffled labels, 5,000 replicates, seed…” · not bound to a source on the page yet
  50. 0 · 6 · What survived · “…labels, 5,000 replicates, seed 0. The median…” · not bound to a source on the page yet
  51. −0.065 · 6 · What survived · “…seed 0. The median difference is −0.065, with a 95 percent…” · not bound to a source on the page yet
  52. 95 · 6 · What survived · “…median difference is −0.065, with a 95 percent interval…” · not bound to a source on the page yet
  53. −0.092 · 6 · What survived · “…−0.065, with a 95 percent interval of −0.092 to −0.030,…” · not bound to a source on the page yet
  54. −0.030 · 6 · What survived · “…a 95 percent interval of −0.092 to −0.030, bootstrapped over…” · not bound to a source on the page yet
  55. 18 · 6 · What survived · “…to −0.030, bootstrapped over the 18 question clusters.…” · not bound to a source on the page yet
  56. 43 · 6 · What survived · “…as a headline. observed On the 43 cells whose floor…” · not bound to a source on the page yet
  57. 33 · 6 · What survived · “…waves rather than one, the share is 33 of 43. That is…” · not bound to a source on the page yet
  58. 43 · 6 · What survived · “…rather than one, the share is 33 of 43. That is 76.7…” · not bound to a source on the page yet
  59. 76.7 · 6 · What survived · “…one, the share is 33 of 43. That is 76.7 percent against…” · not bound to a source on the page yet
  60. 82.5 · 6 · What survived · “…33 of 43. That is 76.7 percent against 82.5 across all 103.…” · not bound to a source on the page yet
  61. 103 · 6 · What survived · “…76.7 percent against 82.5 across all 103. remains unknown…” · not bound to a source on the page yet
  62. 85 · 6 · What survived · “…alone is narrower and still enough: 85 of 103 is the…” · not bound to a source on the page yet
  63. 103 · 6 · What survived · “…is narrower and still enough: 85 of 103 is the headline…” · not bound to a source on the page yet
  64. 33 · 6 · What survived · “…enough: 85 of 103 is the headline and 33 of 43 is the same…” · not bound to a source on the page yet
  65. 43 · 6 · What survived · “…85 of 103 is the headline and 33 of 43 is the same count…” · not bound to a source on the page yet
  66. 3 September · 6 · What survived · “…“held constant and not known” until 3 September, which says two…” · not a measurement
  67. 7 · 7 · What the next study inherits · “…7 · What the next study…” · not a measurement
  68. 4 · 7 · What the next study inherits · “…paid for by errors named in station 4: the second by the…” · not bound to a source on the page yet
  69. 3 September 2026 · 7 · What the next study inherits · “…on this page evidences them. Corrected 3 September 2026. The baseline and…” · not a measurement
  70. 17 September 2026 · 7 · What the next study inherits · “…is what the rules are for. Corrected 17 September 2026. The method and…” · not a measurement
  71. 31 October 2026 · 7 · What the next study inherits · “…since this record began. Last revised 31 October 2026.…” · not a measurement

↑ top

The floor test

304 entries in all. None of them is tied to a source yet. All 304: 64 are not measurements (dates, section numbers, ids, rule numbers, word counts), 238 are not bound to a source on the page yet and 2 are drawings, one entry each.

  1. 08284347 · Opening · “…how much of that I can defend. Scope 08284347, read 29 August…” · not a measurement
  2. 29 August 2026 · Opening · “…I can defend. Scope 08284347, read 29 August 2026. Listed as…” · not a measurement
  3. 001 · Opening · “…Listed as Makefield study MF-2026-001. Back to the…” · not bound to a source on the page yet
  4. 1 · Opening · “…Back to the record. Map of this page 1 Scope, before the…” · not a measurement
  5. 472 words · Opening · “…page 1 Scope, before the finding 472 words 2 The floor…” · not a measurement
  6. 2 · Opening · “…Scope, before the finding 472 words 2 The floor comes…” · not a measurement
  7. 372 words · Opening · “…472 words 2 The floor comes first 372 words 3 The result…” · not a measurement
  8. 3 · Opening · “…2 The floor comes first 372 words 3 The result 310…” · not a measurement
  9. 310 words · Opening · “…comes first 372 words 3 The result 310 words 4 What others…” · not a measurement
  10. 4 · Opening · “…372 words 3 The result 310 words 4 What others have…” · not a measurement
  11. 218 words · Opening · “…have measured with the same measure 218 words 5 What this does…” · not a measurement
  12. 5 · Opening · “…with the same measure 218 words 5 What this does not…” · not a measurement
  13. 1,418 words · Opening · “…words 5 What this does not prove 1,418 words 6 How to check…” · not a measurement
  14. 6 · Opening · “…What this does not prove 1,418 words 6 How to check it…” · not a measurement
  15. 1,192 words · Opening · “…words 6 How to check it yourself 1,192 words Pages of this site…” · not a measurement
  16. 1 · 1 · Scope, before the finding · “…1 · Scope, before the…” · not a measurement
  17. 4 July · 1 · Scope, before the finding · “…whose first captures fall between 4 July and 25 August…” · not a measurement
  18. 25 August 2026 · 1 · Scope, before the finding · “…first captures fall between 4 July and 25 August 2026. Three of the…” · not a measurement
  19. 20260705 · 1 · Scope, before the finding · “…a cell: the second, makefield-cycle0-20260705, took one draw per…” · not a measurement
  20. 2026-07-04 · 1 · Scope, before the finding · “…any number below. makefield-cycle0 2026-07-04…” · not a measurement
  21. 20260705 · 1 · Scope, before the finding · “…2026-07-04 makefield-cycle0-20260705 2026-07-05…” · not a measurement
  22. 2026-07-05 · 1 · Scope, before the finding · “…2026-07-04 makefield-cycle0-20260705 2026-07-05…” · not a measurement
  23. 2026-07-05 · 1 · Scope, before the finding · “…2026-07-05 makefield-cycle0-v2 2026-07-05…” · not a measurement
  24. 2026-07-09 · 1 · Scope, before the finding · “…2026-07-05 makefield-cycle0-v3 2026-07-09 makefield-cycle1…” · not a measurement
  25. 2026-08-10 · 1 · Scope, before the finding · “…2026-07-09 makefield-cycle1 2026-08-10…” · not a measurement
  26. 20260811 · 1 · Scope, before the finding · “…2026-08-10 makefield-cycle1-20260811 2026-08-11…” · not a measurement
  27. 2026-08-11 · 1 · Scope, before the finding · “…2026-08-10 makefield-cycle1-20260811 2026-08-11 makefield-cycle2…” · not a measurement
  28. 2026-08-25 · 1 · Scope, before the finding · “…2026-08-11 makefield-cycle2 2026-08-25 A cell is one…” · not a measurement
  29. 2,952 · 1 · Scope, before the finding · “…crossed with one engine. Folding 2,952 records into cells…” · not bound to a source on the page yet
  30. 111 · 1 · Scope, before the finding · “…2,952 records into cells produced 111 comparable cells.…” · not bound to a source on the page yet
  31. 5 July · 1 · Scope, before the finding · “…re-run asked again on the morning of 5 July. Both questions…” · not a measurement
  32. 111 · 1 · Scope, before the finding · “…nothing from appearing in it. Of the 111, 103 have a…” · not bound to a source on the page yet
  33. 103 · 1 · Scope, before the finding · “…from appearing in it. Of the 111, 103 have a measurable…” · not bound to a source on the page yet
  34. 103 · 1 · Scope, before the finding · “…have a measurable noise floor. Those 103 are the…” · not bound to a source on the page yet
  35. 3 September 2026 · 1 · Scope, before the finding · “…toward more cells moving. Corrected 3 September 2026 and 17 September…” · not a measurement
  36. 17 September 2026 · 1 · Scope, before the finding · “…moving. Corrected 3 September 2026 and 17 September 2026. The capture…” · not a measurement
  37. 25 August 2026 · 1 · Scope, before the finding · “…to end with the wave that begins on 25 August 2026. That wave itself…” · not a measurement
  38. 27 August · 1 · Scope, before the finding · “…2026. That wave itself ran on into 27 August, and the record…” · not a measurement
  39. 26 · 1 · Scope, before the finding · “…the whole of it, so captures from 26 and 27 August are…” · not bound to a source on the page yet
  40. 27 August · 1 · Scope, before the finding · “…whole of it, so captures from 26 and 27 August are inside these…” · not a measurement
  41. 2 · 2 · The floor comes first · “…2 · The floor comes…” · not a measurement
  42. 5 · 2 · The floor comes first · “…and it is the subject of section 5. One rule in the…” · not bound to a source on the page yet
  43. 1 · 2 · The floor comes first · “…empty, the pair returns nothing, not 1. Two answers that…” · not bound to a source on the page yet
  44. 103 · 2 · The floor comes first · “…and the drop is counted. Across the 103 cells the noise…” · not bound to a source on the page yet
  45. 0.266 · 2 · The floor comes first · “…cells the noise floor has a median of 0.266, with quartiles at…” · not bound to a source on the page yet
  46. 0.200 · 2 · The floor comes first · “…a median of 0.266, with quartiles at 0.200 and 0.321, a…” · not bound to a source on the page yet
  47. 0.321 · 2 · The floor comes first · “…of 0.266, with quartiles at 0.200 and 0.321, a minimum of…” · not bound to a source on the page yet
  48. 0.036 · 2 · The floor comes first · “…at 0.200 and 0.321, a minimum of 0.036 and a maximum of…” · not bound to a source on the page yet
  49. 0.600 · 2 · The floor comes first · “…a minimum of 0.036 and a maximum of 0.600. The spread is the…” · not bound to a source on the page yet
  50. 0.036 · 2 · The floor comes first · “…is the point. A cell whose floor is 0.036 is one where the…” · not bound to a source on the page yet
  51. 3 · 3 · The result · “…3 · The result…” · not a measurement
  52. 0.183 · 3 · The result · “…similarity has a median of 0.183 against a floor…” · not bound to a source on the page yet
  53. 0.266 · 3 · The result · “…0.183 against a floor whose median is 0.266. Cell by cell, the…” · not bound to a source on the page yet
  54. −0.065 · 3 · The result · “…Cell by cell, the median difference is −0.065, with a 95 percent…” · not bound to a source on the page yet
  55. 95 · 3 · The result · “…median difference is −0.065, with a 95 percent confidence…” · not bound to a source on the page yet
  56. −0.092 · 3 · The result · “…a 95 percent confidence interval from −0.092 to −0.030,…” · not bound to a source on the page yet
  57. −0.030 · 3 · The result · “…confidence interval from −0.092 to −0.030, bootstrapped over…” · not bound to a source on the page yet
  58. 18 · 3 · The result · “…−0.092 to −0.030, bootstrapped over 18 question clusters,…” · not bound to a source on the page yet
  59. 2,000 · 3 · The result · “…over 18 question clusters, 2,000 replicates, seed…” · not bound to a source on the page yet
  60. 0 · 3 · The result · “…clusters, 2,000 replicates, seed 0. The clustering…” · not bound to a source on the page yet
  61. 85 · 3 · The result · “…cells rather than averaging them: 85 of 103 cells, 82.5…” · not bound to a source on the page yet
  62. 103 · 3 · The result · “…rather than averaging them: 85 of 103 cells, 82.5…” · not bound to a source on the page yet
  63. 82.5 · 3 · The result · “…than averaging them: 85 of 103 cells, 82.5 percent, are less…” · not bound to a source on the page yet
  64. 40.9 · 3 · The result · “…cell, holding the group sizes fixed, 40.9 percent would be…” · not bound to a source on the page yet
  65. 5,000 · 3 · The result · “…would be expected by chance, over 5,000 replicates, seed…” · not bound to a source on the page yet
  66. 0 · 3 · The result · “…by chance, over 5,000 replicates, seed 0. No replicate…” · not bound to a source on the page yet
  67. 85 · 3 · The result · “…seed 0. No replicate reached 85. The summary file…” · not bound to a source on the page yet
  68. 0.0002 · 3 · The result · “…too. The summary file reports p = 0.0002 and states in its…” · not bound to a source on the page yet
  69. 3 · The result · “…Every cell against its own noise floor. One row per cell, 103 rows, sorted by that cell’s own floor. Each row runs from the floor to the between-wave similarity. 85 run left, 18 run right. The floor is not one threshold: it runs from 0.036 to 0.600, which is wider than most of the differences it judges.…” · a drawing; its numbers are not listed one by one
  70. 08284347 · 3 · The result · “…floor, of cells with a floor. Scope 08284347. Engine Moved of…” · not a measurement
  71. 13 · 3 · The result · “…Moved of cells with a floor metaai 13 of 18 chatgpt 15…” · not bound to a source on the page yet
  72. 18 · 3 · The result · “…of cells with a floor metaai 13 of 18 chatgpt 15 of 16…” · not bound to a source on the page yet
  73. 15 · 3 · The result · “…with a floor metaai 13 of 18 chatgpt 15 of 16 claude 12 of…” · not bound to a source on the page yet
  74. 16 · 3 · The result · “…a floor metaai 13 of 18 chatgpt 15 of 16 claude 12 of 16…” · not bound to a source on the page yet
  75. 12 · 3 · The result · “…13 of 18 chatgpt 15 of 16 claude 12 of 16 copilot 14…” · not bound to a source on the page yet
  76. 16 · 3 · The result · “…13 of 18 chatgpt 15 of 16 claude 12 of 16 copilot 14 of 16…” · not bound to a source on the page yet
  77. 14 · 3 · The result · “…15 of 16 claude 12 of 16 copilot 14 of 16 gemini 13 of…” · not bound to a source on the page yet
  78. 16 · 3 · The result · “…15 of 16 claude 12 of 16 copilot 14 of 16 gemini 13 of 15…” · not bound to a source on the page yet
  79. 13 · 3 · The result · “…12 of 16 copilot 14 of 16 gemini 13 of 15 googleai 11…” · not bound to a source on the page yet
  80. 15 · 3 · The result · “…12 of 16 copilot 14 of 16 gemini 13 of 15 googleai 11 of 11…” · not bound to a source on the page yet
  81. 11 · 3 · The result · “…14 of 16 gemini 13 of 15 googleai 11 of 11 perplexity 7…” · not bound to a source on the page yet
  82. 11 · 3 · The result · “…of 16 gemini 13 of 15 googleai 11 of 11 perplexity 7 of 11…” · not bound to a source on the page yet
  83. 7 · 3 · The result · “…13 of 15 googleai 11 of 11 perplexity 7 of 11…” · not bound to a source on the page yet
  84. 11 · 3 · The result · “…15 googleai 11 of 11 perplexity 7 of 11…” · not bound to a source on the page yet
  85. 4 · 4 · What others have measured with the same measure · “…4 · What others have…” · not a measurement
  86. 2603.08924 · 4 · What others have measured with the same measure · “…Figure Sielinski, IQRush, arXiv 2603.08924 Repeat-query…” · not bound to a source on the page yet
  87. 0.29 · 4 · What others have measured with the same measure · “…similarity at domain level Jaccard 0.29 to 0.50 across…” · not bound to a source on the page yet
  88. 0.50 · 4 · What others have measured with the same measure · “…at domain level Jaccard 0.29 to 0.50 across three…” · not bound to a source on the page yet
  89. 2604.07585 · 4 · What others have measured with the same measure · “…University of St. Gallen, arXiv 2604.07585 Same-day,…” · not bound to a source on the page yet
  90. 0.32 · 4 · What others have measured with the same measure · “…same-query source overlap Jaccard 0.32 to 0.43 Jack and…” · not bound to a source on the page yet
  91. 0.43 · 4 · What others have measured with the same measure · “…source overlap Jaccard 0.32 to 0.43 Jack and others,…” · not bound to a source on the page yet
  92. 2605.27440 · 4 · What others have measured with the same measure · “…0.43 Jack and others, Unusual, arXiv 2605.27440 Cross-paraphrase…” · not bound to a source on the page yet
  93. 0.288 · 4 · What others have measured with the same measure · “…against a same-prompt rerun floor 0.288 for cosmetic…” · not bound to a source on the page yet
  94. 0.135 · 4 · What others have measured with the same measure · “…floor 0.288 for cosmetic rewording and 0.135 for rewording that…” · not bound to a source on the page yet
  95. 0.50 · 4 · What others have measured with the same measure · “…adds a constraint, against a floor of 0.50 to 0.61 The third…” · not bound to a source on the page yet
  96. 0.61 · 4 · What others have measured with the same measure · “…constraint, against a floor of 0.50 to 0.61 The third row…” · not bound to a source on the page yet
  97. 5 · 5 · What this does not prove · “…5 · What this does not…” · not a measurement
  98. 103 · 5 · What this does not prove · “…room to fall below it. Splitting the 103 cells by the…” · not bound to a source on the page yet
  99. 103 · 5 · What this does not prove · “…of the cell’s own noise floor. The 103 cells are sorted…” · not bound to a source on the page yet
  100. 26 · 5 · What this does not prove · “…and split into blocks of 26, 26, 25 and 26.…” · not bound to a source on the page yet
  101. 26 · 5 · What this does not prove · “…and split into blocks of 26, 26, 25 and 26. Block…” · not bound to a source on the page yet
  102. 25 · 5 · What this does not prove · “…and split into blocks of 26, 26, 25 and 26. Block…” · not bound to a source on the page yet
  103. 26 · 5 · What this does not prove · “…split into blocks of 26, 26, 25 and 26. Block Floors it…” · not bound to a source on the page yet
  104. 0.0357 · 5 · What this does not prove · “…Floors it contains Moved lowest block 0.0357 to 0.1985 15 of…” · not bound to a source on the page yet
  105. 0.1985 · 5 · What this does not prove · “…contains Moved lowest block 0.0357 to 0.1985 15 of 26, 57.7…” · not bound to a source on the page yet
  106. 15 · 5 · What this does not prove · “…Moved lowest block 0.0357 to 0.1985 15 of 26, 57.7…” · not bound to a source on the page yet
  107. 26 · 5 · What this does not prove · “…lowest block 0.0357 to 0.1985 15 of 26, 57.7 percent…” · not bound to a source on the page yet
  108. 57.7 · 5 · What this does not prove · “…block 0.0357 to 0.1985 15 of 26, 57.7 percent second…” · not bound to a source on the page yet
  109. 0.2021 · 5 · What this does not prove · “…15 of 26, 57.7 percent second block 0.2021 to 0.2655 22 of…” · not bound to a source on the page yet
  110. 0.2655 · 5 · What this does not prove · “…57.7 percent second block 0.2021 to 0.2655 22 of 26, 84.6…” · not bound to a source on the page yet
  111. 22 · 5 · What this does not prove · “…percent second block 0.2021 to 0.2655 22 of 26, 84.6…” · not bound to a source on the page yet
  112. 26 · 5 · What this does not prove · “…second block 0.2021 to 0.2655 22 of 26, 84.6 percent…” · not bound to a source on the page yet
  113. 84.6 · 5 · What this does not prove · “…block 0.2021 to 0.2655 22 of 26, 84.6 percent third…” · not bound to a source on the page yet
  114. 0.2656 · 5 · What this does not prove · “…22 of 26, 84.6 percent third block 0.2656 to 0.3174 23 of…” · not bound to a source on the page yet
  115. 0.3174 · 5 · What this does not prove · “…26, 84.6 percent third block 0.2656 to 0.3174 23 of 25, 92.0…” · not bound to a source on the page yet
  116. 23 · 5 · What this does not prove · “…percent third block 0.2656 to 0.3174 23 of 25, 92.0…” · not bound to a source on the page yet
  117. 25 · 5 · What this does not prove · “…third block 0.2656 to 0.3174 23 of 25, 92.0 percent…” · not bound to a source on the page yet
  118. 92.0 · 5 · What this does not prove · “…third block 0.2656 to 0.3174 23 of 25, 92.0 percent highest…” · not bound to a source on the page yet
  119. 0.3237 · 5 · What this does not prove · “…23 of 25, 92.0 percent highest block 0.3237 to 0.6000 25 of…” · not bound to a source on the page yet
  120. 0.6000 · 5 · What this does not prove · “…92.0 percent highest block 0.3237 to 0.6000 25 of 26, 96.2…” · not bound to a source on the page yet
  121. 25 · 5 · What this does not prove · “…percent highest block 0.3237 to 0.6000 25 of 26, 96.2…” · not bound to a source on the page yet
  122. 26 · 5 · What this does not prove · “…highest block 0.3237 to 0.6000 25 of 26, 96.2 percent The…” · not bound to a source on the page yet
  123. 96.2 · 5 · What this does not prove · “…block 0.3237 to 0.6000 25 of 26, 96.2 percent The middle…” · not bound to a source on the page yet
  124. 2 · 5 · What this does not prove · “…quartile values printed in section 2 gives 57.7, 81.5,…” · not bound to a source on the page yet
  125. 57.7 · 5 · What this does not prove · “…values printed in section 2 gives 57.7, 81.5, 95.8 and…” · not bound to a source on the page yet
  126. 81.5 · 5 · What this does not prove · “…printed in section 2 gives 57.7, 81.5, 95.8 and 96.2…” · not bound to a source on the page yet
  127. 95.8 · 5 · What this does not prove · “…printed in section 2 gives 57.7, 81.5, 95.8 and 96.2 instead.…” · not bound to a source on the page yet
  128. 96.2 · 5 · What this does not prove · “…section 2 gives 57.7, 81.5, 95.8 and 96.2 instead. Corrected…” · not bound to a source on the page yet
  129. 3 September 2026 · 5 · What this does not prove · “…81.5, 95.8 and 96.2 instead. Corrected 3 September 2026. The headline…” · not a measurement
  130. 82.5 · 5 · What this does not prove · “…September 2026. The headline share of 82.5 percent is…” · not bound to a source on the page yet
  131. 0.4 · 5 · What this does not prove · “…its floor” counts a cell that fell by 0.4 the same as a cell…” · not bound to a source on the page yet
  132. 0.003 · 5 · What this does not prove · “…by 0.4 the same as a cell that fell by 0.003. Of the 85 that…” · not bound to a source on the page yet
  133. 85 · 5 · What this does not prove · “…as a cell that fell by 0.003. Of the 85 that moved, 6 are…” · not bound to a source on the page yet
  134. 6 · 5 · What this does not prove · “…fell by 0.003. Of the 85 that moved, 6 are within 0.01 of…” · not bound to a source on the page yet
  135. 0.01 · 5 · What this does not prove · “…Of the 85 that moved, 6 are within 0.01 of their floor and…” · not bound to a source on the page yet
  136. 15 · 5 · What this does not prove · “…6 are within 0.01 of their floor and 15 are within 0.02.…” · not bound to a source on the page yet
  137. 0.02 · 5 · What this does not prove · “…0.01 of their floor and 15 are within 0.02. The median margin…” · not bound to a source on the page yet
  138. 0.084 · 5 · What this does not prove · “…are within 0.02. The median margin is 0.084 and the largest is…” · not bound to a source on the page yet
  139. 0.408 · 5 · What this does not prove · “…margin is 0.084 and the largest is 0.408. So the effect is…” · not bound to a source on the page yet
  140. 1 · 5 · What this does not prove · “…waves compared in a cell ranges from 1 day to 52. Calling…” · not bound to a source on the page yet
  141. 52 · 5 · What this does not prove · “…in a cell ranges from 1 day to 52. Calling the…” · not bound to a source on the page yet
  142. 1 · 5 · What this does not prove · “…not the same cell over time. Gap Moved 1 to 14 days 8 of…” · not bound to a source on the page yet
  143. 14 · 5 · What this does not prove · “…same cell over time. Gap Moved 1 to 14 days 8 of 12, 66.7…” · not bound to a source on the page yet
  144. 8 · 5 · What this does not prove · “…cell over time. Gap Moved 1 to 14 days 8 of 12, 66.7…” · not bound to a source on the page yet
  145. 12 · 5 · What this does not prove · “…over time. Gap Moved 1 to 14 days 8 of 12, 66.7 percent 15…” · not bound to a source on the page yet
  146. 66.7 · 5 · What this does not prove · “…time. Gap Moved 1 to 14 days 8 of 12, 66.7 percent 15 to 40…” · not bound to a source on the page yet
  147. 15 · 5 · What this does not prove · “…1 to 14 days 8 of 12, 66.7 percent 15 to 40 days 28 of…” · not bound to a source on the page yet
  148. 40 · 5 · What this does not prove · “…to 14 days 8 of 12, 66.7 percent 15 to 40 days 28 of 35,…” · not bound to a source on the page yet
  149. 28 · 5 · What this does not prove · “…8 of 12, 66.7 percent 15 to 40 days 28 of 35, 80.0…” · not bound to a source on the page yet
  150. 35 · 5 · What this does not prove · “…12, 66.7 percent 15 to 40 days 28 of 35, 80.0 percent 41…” · not bound to a source on the page yet
  151. 80.0 · 5 · What this does not prove · “…66.7 percent 15 to 40 days 28 of 35, 80.0 percent 41 days or…” · not bound to a source on the page yet
  152. 41 · 5 · What this does not prove · “…15 to 40 days 28 of 35, 80.0 percent 41 days or more 49 of…” · not bound to a source on the page yet
  153. 49 · 5 · What this does not prove · “…28 of 35, 80.0 percent 41 days or more 49 of 56, 87.5…” · not bound to a source on the page yet
  154. 56 · 5 · What this does not prove · “…35, 80.0 percent 41 days or more 49 of 56, 87.5 percent…” · not bound to a source on the page yet
  155. 87.5 · 5 · What this does not prove · “…80.0 percent 41 days or more 49 of 56, 87.5 percent Within a…” · not bound to a source on the page yet
  156. −0.047 · 5 · What this does not prove · “…its shorter one, the difference is −0.047 with a 95 percent…” · not bound to a source on the page yet
  157. 95 · 5 · What this does not prove · “…one, the difference is −0.047 with a 95 percent interval…” · not bound to a source on the page yet
  158. −0.058 · 5 · What this does not prove · “…−0.047 with a 95 percent interval from −0.058 to −0.031, and 87…” · not bound to a source on the page yet
  159. −0.031 · 5 · What this does not prove · “…a 95 percent interval from −0.058 to −0.031, and 87 of 108…” · not bound to a source on the page yet
  160. 87 · 5 · What this does not prove · “…interval from −0.058 to −0.031, and 87 of 108 cells are…” · not bound to a source on the page yet
  161. 108 · 5 · What this does not prove · “…from −0.058 to −0.031, and 87 of 108 cells are less…” · not bound to a source on the page yet
  162. 21 · 5 · What this does not prove · “…are less alike at the longer gap while 21 are more alike.…” · not bound to a source on the page yet
  163. 5 · What this does not prove · “…Each pair of capture dates, against the same cell’s shortest gap. Pooled, the pairs within a month sit at −0.011 and the pairs across the month boundary at −0.045. Row by row, the within-month rows run from −0.005 to −0.021, except the 4 July → 5 July row (10 cells, hollow), which is +0.075; the cross-month rows run from −0.026 to −0.078. Inside a cell the same step shows: of the 61 cells that have pairs of both kinds, 51 have a lower median difference across the month boundary than within a month. The figure cannot tell the month from the gap: every pair across the boundary is 32 days or more apart and every pair within a month is 15 days or less. That is the sense in which the result is not net of period.…” · a drawing; its numbers are not listed one by one
  164. 3 September 2026 · 5 · What this does not prove · “…strongest thing on this page and until 3 September 2026 it was the least…” · not a measurement
  165. 957 · 5 · What this does not prove · “…now: a third file carries all 957 wave pairs, and…” · not bound to a source on the page yet
  166. 37 · 5 · What this does not prove · “…pairs at the same gap in days, and 37 of the 108 cells…” · not bound to a source on the page yet
  167. 108 · 5 · What this does not prove · “…at the same gap in days, and 37 of the 108 cells do, so which…” · not bound to a source on the page yet
  168. −0.042 · 5 · What this does not prove · “…of making it, the median runs from −0.042 to −0.066 and the…” · not bound to a source on the page yet
  169. −0.066 · 5 · What this does not prove · “…it, the median runs from −0.042 to −0.066 and the count from…” · not bound to a source on the page yet
  170. 83 · 5 · What this does not prove · “…−0.042 to −0.066 and the count from 83 to 93 of 108. The…” · not bound to a source on the page yet
  171. 93 · 5 · What this does not prove · “…to −0.066 and the count from 83 to 93 of 108. The…” · not bound to a source on the page yet
  172. 108 · 5 · What this does not prove · “…−0.066 and the count from 83 to 93 of 108. The direction…” · not bound to a source on the page yet
  173. −0.047 · 5 · What this does not prove · “…would have been a finding. Read −0.047 as one defensible…” · not bound to a source on the page yet
  174. 108 · 5 · What this does not prove · “…point inside that range. Those 108 cells are not the…” · not bound to a source on the page yet
  175. 103 · 5 · What this does not prove · “…range. Those 108 cells are not the 103, and not a subset…” · not bound to a source on the page yet
  176. 111 · 5 · What this does not prove · “…not the 103, and not a subset of the 111 either. The…” · not bound to a source on the page yet
  177. 123 · 5 · What this does not prove · “…111 either. The gap-pairs file covers 123 cells, because a…” · not bound to a source on the page yet
  178. 123 · 5 · What this does not prove · “…pairs are in the file. Fifteen of the 123 carry a single…” · not bound to a source on the page yet
  179. 108 · 5 · What this does not prove · “…distinct gap and fall out, leaving 108: 100 of them are…” · not bound to a source on the page yet
  180. 100 · 5 · What this does not prove · “…gap and fall out, leaving 108: 100 of them are among…” · not bound to a source on the page yet
  181. 111 · 5 · What this does not prove · “…leaving 108: 100 of them are among the 111 and 8 are not.…” · not bound to a source on the page yet
  182. 8 · 5 · What this does not prove · “…108: 100 of them are among the 111 and 8 are not.…” · not bound to a source on the page yet
  183. 100 · 5 · What this does not prove · “…111 and 8 are not. Restricted to the 100, the median is…” · not bound to a source on the page yet
  184. −0.042 · 5 · What this does not prove · “…Restricted to the 100, the median is −0.042 with 79 of 100…” · not bound to a source on the page yet
  185. 79 · 5 · What this does not prove · “…to the 100, the median is −0.042 with 79 of 100 less alike…” · not bound to a source on the page yet
  186. 100 · 5 · What this does not prove · “…100, the median is −0.042 with 79 of 100 less alike at the…” · not bound to a source on the page yet
  187. 108 · 5 · What this does not prove · “…end of the range above. And two of the 108 should not be in…” · not bound to a source on the page yet
  188. −0.046 · 5 · What this does not prove · “…those two cells. Dropping both gives −0.046 over 106, with 85…” · not bound to a source on the page yet
  189. 106 · 5 · What this does not prove · “…cells. Dropping both gives −0.046 over 106, with 85 less…” · not bound to a source on the page yet
  190. 85 · 5 · What this does not prove · “…both gives −0.046 over 106, with 85 less alike at the…” · not bound to a source on the page yet
  191. 21 · 5 · What this does not prove · “…85 less alike at the longer gap and 21 more alike. The…” · not bound to a source on the page yet
  192. 3 September 2026 · 5 · What this does not prove · “…an engine composition effect. Found on 3 September 2026 by recomputing the…” · not a measurement
  193. 31 August 2026 · 5 · What this does not prove · “…a single pair. This one was found on 31 August 2026 while checking…” · not a measurement
  194. 103 · 5 · What this does not prove · “…pairs of a cell. Thirty-eight of the 103 cells have exactly…” · not bound to a source on the page yet
  195. 60 · 5 · What this does not prove · “…is that number. Counting more broadly, 60 of the 103 draw…” · not bound to a source on the page yet
  196. 103 · 5 · What this does not prove · “…Counting more broadly, 60 of the 103 draw their floor…” · not bound to a source on the page yet
  197. 60 · 5 · What this does not prove · “…difference Floor from one wave only 60 52, 86.7 percent…” · not bound to a source on the page yet
  198. 52 · 5 · What this does not prove · “…difference Floor from one wave only 60 52, 86.7 percent…” · not bound to a source on the page yet
  199. 86.7 · 5 · What this does not prove · “…Floor from one wave only 60 52, 86.7 percent −0.086…” · not bound to a source on the page yet
  200. −0.086 · 5 · What this does not prove · “…from one wave only 60 52, 86.7 percent −0.086 Floor from both…” · not bound to a source on the page yet
  201. 43 · 5 · What this does not prove · “…percent −0.086 Floor from both waves 43 33, 76.7 percent…” · not bound to a source on the page yet
  202. 33 · 5 · What this does not prove · “…−0.086 Floor from both waves 43 33, 76.7 percent…” · not bound to a source on the page yet
  203. 76.7 · 5 · What this does not prove · “…−0.086 Floor from both waves 43 33, 76.7 percent −0.032 All…” · not bound to a source on the page yet
  204. −0.032 · 5 · What this does not prove · “…from both waves 43 33, 76.7 percent −0.032 All cells, as…” · not bound to a source on the page yet
  205. 103 · 5 · What this does not prove · “…−0.032 All cells, as reported above 103 85, 82.5 percent…” · not bound to a source on the page yet
  206. 85 · 5 · What this does not prove · “…All cells, as reported above 103 85, 82.5 percent…” · not bound to a source on the page yet
  207. 82.5 · 5 · What this does not prove · “…All cells, as reported above 103 85, 82.5 percent −0.065 On…” · not bound to a source on the page yet
  208. −0.065 · 5 · What this does not prove · “…as reported above 103 85, 82.5 percent −0.065 On the…” · not bound to a source on the page yet
  209. 103 · 5 · What this does not prove · “…it. The share on that subset of the 103 is 76.7 percent…” · not bound to a source on the page yet
  210. 76.7 · 5 · What this does not prove · “…The share on that subset of the 103 is 76.7 percent and…” · not bound to a source on the page yet
  211. 2 · 5 · What this does not prove · “…and draws_late are each at least 2. The count of…” · not bound to a source on the page yet
  212. 1 · 5 · What this does not prove · “…a third column, within_pairs equal to 1. This is the…” · not bound to a source on the page yet
  213. 29 August 2026 · 5 · What this does not prove · “…A verifier ran against that archive on 29 August 2026 and failed it on…” · not a measurement
  214. 4 July 2026 · 5 · What this does not prove · “…were, in stages: the first ten on 4 July 2026 and the full…” · not a measurement
  215. 9 July · 5 · What this does not prove · “…4 July 2026 and the full eighteen on 9 July, each stage in a…” · not a measurement
  216. 22 August · 5 · What this does not prove · “…The floor test itself was written on 22 August, after six of the…” · not a measurement
  217. 30 August 2026 · 5 · What this does not prove · “…plan in this practice is dated 30 August 2026, itself still a…” · not a measurement
  218. 6 · 6 · How to check it yourself · “…6 · How to check it…” · not a measurement
  219. 3 September 2026 · 6 · How to check it yourself · “…counts. The gap-pairs CSV, added on 3 September 2026, has one row per…” · not a measurement
  220. 6 · 6 · How to check it yourself · “…counts and the exclusion reasons. 6 files CC BY 4.0 ·…” · not bound to a source on the page yet
  221. CC BY 4.0 · 6 · How to check it yourself · “…and the exclusion reasons. 6 files CC BY 4.0 · 273 KB ·…” · not a measurement
  222. 273 · 6 · How to check it yourself · “…exclusion reasons. 6 files CC BY 4.0 · 273 KB · checksummed…” · not bound to a source on the page yet
  223. 08284347 · 6 · How to check it yourself · “…273 KB · checksummed floor-test-cells-08284347.csv one row per…” · not a measurement
  224. 13 · 6 · How to check it yourself · “…row per question-by-engine cell size 13 KB contents 111…” · not bound to a source on the page yet
  225. 111 · 6 · How to check it yourself · “…cell size 13 KB contents 111 rows sha256…” · not bound to a source on the page yet
  226. 08284347 · 6 · How to check it yourself · “…floor-test-gap-pairs-08284347.csv one row per…” · not a measurement
  227. 81 · 6 · How to check it yourself · “…what the within-cell test needs size 81 KB contents 957…” · not bound to a source on the page yet
  228. 957 · 6 · How to check it yourself · “…test needs size 81 KB contents 957 rows sha256…” · not bound to a source on the page yet
  229. 08284347 · 6 · How to check it yourself · “…floor-test-summary-08284347.json the aggregate…” · not a measurement
  230. 11 · 6 · How to check it yourself · “…every seed and replicate count size 11 KB contents 24…” · not bound to a source on the page yet
  231. 24 · 6 · How to check it yourself · “…replicate count size 11 KB contents 24 top-level keys…” · not bound to a source on the page yet
  232. 08284347 · 6 · How to check it yourself · “…codebook-08284347.json the codebook:…” · not a measurement
  233. 17 · 6 · How to check it yourself · “…README's words and allowed values size 17 KB contents both…” · not bound to a source on the page yet
  234. 08284347 · 6 · How to check it yourself · “…floor-test-cells-08284347.sav the same table…” · not a measurement
  235. 24 · 6 · How to check it yourself · “…for SPSS, generated from the CSV size 24 KB contents 111…” · not bound to a source on the page yet
  236. 111 · 6 · How to check it yourself · “…from the CSV size 24 KB contents 111 rows sha256…” · not bound to a source on the page yet
  237. 08284347 · 6 · How to check it yourself · “…floor-test-gap-pairs-08284347.sav the same table…” · not a measurement
  238. 125 · 6 · How to check it yourself · “…for SPSS, generated from the CSV size 125 KB contents 957…” · not bound to a source on the page yet
  239. 957 · 6 · How to check it yourself · “…from the CSV size 125 KB contents 957 rows sha256…” · not bound to a source on the page yet
  240. 1 September 2026 · 6 · How to check it yourself · “…were rederived from the cells file on 1 September 2026 with all fourteen…” · not a measurement
  241. 3 September · 6 · How to check it yourself · “…rederived from the gap-pairs file on 3 September. The summary…” · not a measurement
  242. 1 September 2026 · 6 · How to check it yourself · “…than asserted: the first five on 1 September 2026, fourteen separate…” · not a measurement
  243. 3 September · 6 · How to check it yourself · “…all matching, and the sixth on 3 September when the file it…” · not a measurement
  244. 103 · 6 · How to check it yourself · “…in full. The first five are over the 103 cells that have a…” · not bound to a source on the page yet
  245. 111 · 6 · How to check it yourself · “…cells that have a floor, not over all 111 rows, and that…” · not bound to a source on the page yet
  246. 85 · 6 · How to check it yourself · “…headline. Sum text_moved_beyond_noise: 85 of 103. The gap…” · not bound to a source on the page yet
  247. 103 · 6 · How to check it yourself · “…Sum text_moved_beyond_noise: 85 of 103. The gap table.…” · not bound to a source on the page yet
  248. 14 · 6 · How to check it yourself · “…103. The gap table. Band gap_days at 14 and 40, over the…” · not bound to a source on the page yet
  249. 40 · 6 · How to check it yourself · “…The gap table. Band gap_days at 14 and 40, over the 103. ⚠…” · not bound to a source on the page yet
  250. 103 · 6 · How to check it yourself · “…Band gap_days at 14 and 40, over the 103. ⚠ Over all 111…” · not bound to a source on the page yet
  251. 111 · 6 · How to check it yourself · “…at 14 and 40, over the 103. ⚠ Over all 111 rows instead, two…” · not bound to a source on the page yet
  252. 28 · 6 · How to check it yourself · “…the three bands come out differently: 28 of 40 and 49 of 59…” · not bound to a source on the page yet
  253. 40 · 6 · How to check it yourself · “…bands come out differently: 28 of 40 and 49 of 59…” · not bound to a source on the page yet
  254. 49 · 6 · How to check it yourself · “…come out differently: 28 of 40 and 49 of 59 rather than…” · not bound to a source on the page yet
  255. 59 · 6 · How to check it yourself · “…out differently: 28 of 40 and 49 of 59 rather than 28 of…” · not bound to a source on the page yet
  256. 28 · 6 · How to check it yourself · “…28 of 40 and 49 of 59 rather than 28 of 35 and 49 of…” · not bound to a source on the page yet
  257. 35 · 6 · How to check it yourself · “…of 40 and 49 of 59 rather than 28 of 35 and 49 of 56.…” · not bound to a source on the page yet
  258. 49 · 6 · How to check it yourself · “…and 49 of 59 rather than 28 of 35 and 49 of 56. Corrected 1…” · not bound to a source on the page yet
  259. 56 · 6 · How to check it yourself · “…of 59 rather than 28 of 35 and 49 of 56. Corrected 1…” · not bound to a source on the page yet
  260. 1 September 2026 · 6 · How to check it yourself · “…than 28 of 35 and 49 of 56. Corrected 1 September 2026. The margin…” · not a measurement
  261. 85 · 6 · How to check it yourself · “…margin counts. Take delta_text for the 85 cells that moved:…” · not bound to a source on the page yet
  262. 6 · 6 · How to check it yourself · “…for the 85 cells that moved: 6 within 0.01, 15…” · not bound to a source on the page yet
  263. 0.01 · 6 · How to check it yourself · “…for the 85 cells that moved: 6 within 0.01, 15 within 0.02,…” · not bound to a source on the page yet
  264. 15 · 6 · How to check it yourself · “…85 cells that moved: 6 within 0.01, 15 within 0.02,…” · not bound to a source on the page yet
  265. 0.02 · 6 · How to check it yourself · “…that moved: 6 within 0.01, 15 within 0.02, median 0.084,…” · not bound to a source on the page yet
  266. 0.084 · 6 · How to check it yourself · “…6 within 0.01, 15 within 0.02, median 0.084, largest 0.408.…” · not bound to a source on the page yet
  267. 0.408 · 6 · How to check it yourself · “…15 within 0.02, median 0.084, largest 0.408. The…” · not bound to a source on the page yet
  268. 2 · 6 · How to check it yourself · “…and draws_late are each at least 2. Of the 43 such…” · not bound to a source on the page yet
  269. 43 · 6 · How to check it yourself · “…draws_late are each at least 2. Of the 43 such cells, 33…” · not bound to a source on the page yet
  270. 33 · 6 · How to check it yourself · “…each at least 2. Of the 43 such cells, 33 moved; of the…” · not bound to a source on the page yet
  271. 60 · 6 · How to check it yourself · “…43 such cells, 33 moved; of the other 60, 52 moved. The…” · not bound to a source on the page yet
  272. 52 · 6 · How to check it yourself · “…such cells, 33 moved; of the other 60, 52 moved. The…” · not bound to a source on the page yet
  273. 103 · 6 · How to check it yourself · “…52 moved. The quartile table. Sort the 103 by noise_sim_text…” · not bound to a source on the page yet
  274. 26 · 6 · How to check it yourself · “…and split into rank blocks of 26, 26, 25 and 26.…” · not bound to a source on the page yet
  275. 26 · 6 · How to check it yourself · “…and split into rank blocks of 26, 26, 25 and 26.…” · not bound to a source on the page yet
  276. 25 · 6 · How to check it yourself · “…and split into rank blocks of 26, 26, 25 and 26. Splitting…” · not bound to a source on the page yet
  277. 26 · 6 · How to check it yourself · “…into rank blocks of 26, 26, 25 and 26. Splitting the…” · not bound to a source on the page yet
  278. 88.5 · 6 · How to check it yourself · “…an answer: two other conventions give 88.5 and 100.0 for the…” · not bound to a source on the page yet
  279. 100.0 · 6 · How to check it yourself · “…two other conventions give 88.5 and 100.0 for the upper two…” · not bound to a source on the page yet
  280. 84.0 · 6 · How to check it yourself · “…and 100.0 for the upper two bands, or 84.0 and 92.3 for the…” · not bound to a source on the page yet
  281. 92.3 · 6 · How to check it yourself · “…for the upper two bands, or 84.0 and 92.3 for the middle…” · not bound to a source on the page yet
  282. 2 · 6 · How to check it yourself · “…quartile values printed in section 2 does not reproduce…” · not bound to a source on the page yet
  283. 5 · 6 · How to check it yourself · “…the value-binned shares in section 5 instead, because…” · not bound to a source on the page yet
  284. 1 September 2026 · 6 · How to check it yourself · “…README now carries the rule. Corrected 1 September 2026. The within-cell…” · not a measurement
  285. 957 · 6 · How to check it yourself · “…CSV, not the cells CSV. Group the 957 rows by intent_id…” · not bound to a source on the page yet
  286. 957 · 6 · How to check it yourself · “…whose rows all share one gap_days. The 957 rows cover 123…” · not bound to a source on the page yet
  287. 123 · 6 · How to check it yourself · “…share one gap_days. The 957 rows cover 123 cells rather than…” · not bound to a source on the page yet
  288. 111 · 6 · How to check it yourself · “…rows cover 123 cells rather than the 111, because a cell…” · not bound to a source on the page yet
  289. 15 · 6 · How to check it yourself · “…whose question text was identical; 15 fall out for…” · not bound to a source on the page yet
  290. 108 · 6 · How to check it yourself · “…for holding a single distinct gap, and 108 remain, of which 8…” · not bound to a source on the page yet
  291. 8 · 6 · How to check it yourself · “…distinct gap, and 108 remain, of which 8 sit outside the…” · not bound to a source on the page yet
  292. 111 · 6 · How to check it yourself · “…108 remain, of which 8 sit outside the 111. In each of the…” · not bound to a source on the page yet
  293. 108 · 6 · How to check it yourself · “…the 111. In each of the remaining 108, take the row with…” · not bound to a source on the page yet
  294. −0.0465 · 6 · How to check it yourself · “…one’s, rounded to four places. Median −0.0465, with 87 negative…” · not bound to a source on the page yet
  295. 87 · 6 · How to check it yourself · “…to four places. Median −0.0465, with 87 negative and 21…” · not bound to a source on the page yet
  296. 21 · 6 · How to check it yourself · “…Median −0.0465, with 87 negative and 21 positive. ⚠ The…” · not bound to a source on the page yet
  297. 37 · 6 · How to check it yourself · “…⚠ The tie-break is not decoration: 37 of the 108 cells…” · not bound to a source on the page yet
  298. 108 · 6 · How to check it yourself · “…tie-break is not decoration: 37 of the 108 cells have one,…” · not bound to a source on the page yet
  299. −0.042 · 6 · How to check it yourself · “…reasonable rules give medians from −0.042 to −0.066. Six do…” · not bound to a source on the page yet
  300. −0.066 · 6 · How to check it yourself · “…rules give medians from −0.042 to −0.066. Six do not. The…” · not bound to a source on the page yet
  301. 111 · 6 · How to check it yourself · “…in either CSV, which between them hold 111 rows and 957 rows…” · not bound to a source on the page yet
  302. 957 · 6 · How to check it yourself · “…which between them hold 111 rows and 957 rows and no…” · not bound to a source on the page yet
  303. 3 September 2026 · 6 · How to check it yourself · “…rest of this section offers. Corrected 3 September 2026. Where the CSV and…” · not a measurement
  304. 31 October 2026 · 6 · How to check it yourself · “…the trail they leave out. Last revised 31 October 2026.…” · not a measurement

↑ top

What changed here

468 entries in all. None of them is tied to a source yet. All 468: 187 are not measurements (dates, section numbers, ids, rule numbers, word counts), 281 are not bound to a source on the page yet and 0 are drawings, one entry each.

  1. 1 October 2026 · Opening · “…of the story. This site went live on 1 October 2026. Entries dated…” · not a measurement
  2. 1 · Opening · “…it was in preview. Map of this page 1 What gets an entry…” · not a measurement
  3. 294 words · Opening · “…of this page 1 What gets an entry 294 words 2 The record…” · not a measurement
  4. 2 · Opening · “…1 What gets an entry 294 words 2 The record 14,165…” · not a measurement
  5. 14,165 words · Opening · “…an entry 294 words 2 The record 14,165 words 3 How to check…” · not a measurement
  6. 3 · Opening · “…words 2 The record 14,165 words 3 How to check any…” · not a measurement
  7. 683 words · Opening · “…words 3 How to check any of it 683 words Pages of this site…” · not a measurement
  8. 1 · 1 · What gets an entry · “…1 · What gets an entry…” · not a measurement
  9. 1 September 2026 · 1 · What gets an entry · “…and what is not in it It begins on 1 September 2026. Changes before…” · not a measurement
  10. 2026 · 1 · What gets an entry · “…to the silence above. In September 2026 a pass over these…” · not bound to a source on the page yet
  11. 2 · 2 · The record · “…2 · The record 31…” · not a measurement
  12. 31 October 2026 · 2 · The record · “…2 · The record 31 October 2026 The first step of…” · not a measurement
  13. 10.5281/zenodo.23780961 · 2 · The record · “…the study is archived on Zenodo (doi:10.5281/zenodo.23780961) and the plain…” · not a measurement
  14. 31 October 2026 · 2 · The record · “…is repeated where the step is stated. 31 October 2026 The dates on which…” · not a measurement
  15. 4 July · 2 · The record · “…questions frozen and registered on 4 July; the first wave…” · not a measurement
  16. 4 July · 2 · The record · “…wave after ten questions registered on 4 July and all eighteen…” · not a measurement
  17. 9 July · 2 · The record · “…on 4 July and all eighteen by 9 July; the first wave…” · not a measurement
  18. 5 July · 2 · The record · “…wrongly written questions corrected on 5 July The capture ledger…” · not a measurement
  19. 4 July · 2 · The record · “…It shows ten questions registered on 4 July, eighteen on 9…” · not a measurement
  20. 9 July · 2 · The record · “…registered on 4 July, eighteen on 9 July, six engines in…” · not a measurement
  21. 10 August · 2 · The record · “…wave and Meta AI and Mistral from 10 August. The earlier…” · not a measurement
  22. 3 September 2026 · 2 · The record · “…than the record holds. The entry of 3 September 2026 below repeats the…” · not a measurement
  23. 28 October 2026 · 2 · The record · “…account and is left as it was written. 28 October 2026 Part of the 3…” · not a measurement
  24. 3 September · 2 · The record · “…written. 28 October 2026 Part of the 3 September change is…” · not a measurement
  25. 001 · 2 · The record · “…there as Makefield study MF-2026-001. before no…” · not bound to a source on the page yet
  26. 001 · 2 · The record · “…“Listed as Makefield study MF-2026-001” On 3 September…” · not bound to a source on the page yet
  27. 3 September · 2 · The record · “…as Makefield study MF-2026-001” On 3 September the attribution…” · not a measurement
  28. 19 October · 2 · The record · “…still reads “Denes Csaszar”. Since 19 October the study page’s…” · not a measurement
  29. 3 September · 2 · The record · “…see that too. The two corrections of 3 September that stopped pages…” · not a measurement
  30. 28 October 2026 · 2 · The record · “…are. The dataset files are unchanged. 28 October 2026 The plan on the…” · not a measurement
  31. 19 October 2026 · 2 · The record · “…alongside the interfaces (decided 19 October 2026). The plan had…” · not a measurement
  32. 19 October 2026 · 2 · The record · “…the change where the plan is stated. 19 October 2026 The record page,…” · not a measurement
  33. 9 · 2 · The record · “…the paraphrase study, in section 9 and marked as a…” · not bound to a source on the page yet
  34. 19 October 2026 · 2 · The record · “…status depending on the page. 19 October 2026 On the English and…” · not a measurement
  35. 19 October 2026 · 2 · The record · “…legal pages are translations of them. 19 October 2026 The study page…” · not a measurement
  36. 5 · 2 · The record · “…study page gains a figure in section 5, after the…” · not bound to a source on the page yet
  37. 19 October 2026 · 2 · The record · “…if the drawing and the file disagree. 19 October 2026 The study page’s…” · not a measurement
  38. 001 · 2 · The record · “…The study is MF-2026-001 in the Makefield…” · not bound to a source on the page yet
  39. 19 October 2026 · 2 · The record · “…than describing it a second time. 19 October 2026 The structured…” · not a measurement
  40. 1609 · 2 · The record · “…orcid.org/0009-0002-1002-1609 Until now the…” · not bound to a source on the page yet
  41. 19 October 2026 · 2 · The record · “…an organisation is not the person. 19 October 2026 The footer of the…” · not a measurement
  42. 1 October · 2 · The record · “…still link the English ones. The 1 October entry says the…” · not a measurement
  43. 19 October 2026 · 2 · The record · “…stays as written; this one narrows it. 19 October 2026 Ten sentences on…” · not a measurement
  44. 5,000 · 2 · The record · “…it” after “p below the smallest value 5,000 replicates can…” · not bound to a source on the page yet
  45. 22 August · 2 · The record · “…The registration sentence is about 22 August, and read today…” · not a measurement
  46. 19 October 2026 · 2 · The record · “…the floor test is still unpublished. 19 October 2026 The privacy notice…” · not a measurement
  47. 10 September 2026 · 2 · The record · “…October 2026 The privacy notice gave 10 September 2026 as its date, in…” · not a measurement
  48. 22 September 2026 · 2 · The record · “…and in its translations. It now gives 22 September 2026. before „Diese…” · not a measurement
  49. 10 · 2 · The record · “…before „Diese Erklärung hat den Stand 10. September 2026.“…” · not bound to a source on the page yet
  50. 2026 · 2 · The record · “…Erklärung hat den Stand 10. September 2026.“ after „Diese…” · not bound to a source on the page yet
  51. 22 · 2 · The record · “…after „Diese Erklärung hat den Stand 22. September 2026.“…” · not bound to a source on the page yet
  52. 2026 · 2 · The record · “…Erklärung hat den Stand 22. September 2026.“ Section 3 was…” · not bound to a source on the page yet
  53. 3 · 2 · The record · “…den Stand 22. September 2026.“ Section 3 was changed on 22…” · not bound to a source on the page yet
  54. 22 September · 2 · The record · “…2026.“ Section 3 was changed on 22 September, when the…” · not a measurement
  55. 17 September · 2 · The record · “…bound, and it names a value read on 17 September. A notice dated…” · not a measurement
  56. 1 September · 2 · The record · “…trusted, which is the argument of the 1 September entry about…” · not a measurement
  57. 19 October 2026 · 2 · The record · “…the German date and move with it. 19 October 2026 The legal notice…” · not a measurement
  58. 19 October 2026 · 2 · The record · “…nothing the German page does not. 19 October 2026 A statement on how…” · not a measurement
  59. 3 October 2026 · 2 · The record · “…on the record, study and method pages. 3 October 2026 The record page,…” · not a measurement
  60. 08 · 2 · The record · “…in Hungarian, at /hu, /hu/drift-2026-08, /hu/modszer and…” · not bound to a source on the page yet
  61. 1 October 2026 · 2 · The record · “…to it from a Hungarian page says so. 1 October 2026 Four pages said…” · not a measurement
  62. 22 August · 2 · The record · “…said the floor test was written on 22 August “after the data…” · not a measurement
  63. 25 August · 2 · The record · “…the seventh was first captured on 25 August. before “the floor…” · not a measurement
  64. 22 August · 2 · The record · “…“the floor test itself was written on 22 August, after the data…” · not a measurement
  65. 25 August · 2 · The record · “…dates the last wave’s first capture to 25 August, three days after…” · not a measurement
  66. 1 October 2026 · 2 · The record · “…build seat while checking the dates. 1 October 2026 The record page,…” · not a measurement
  67. 08 · 2 · The record · “…in German, at /de, /de/drift-2026-08, /de/methode and…” · not bound to a source on the page yet
  68. 1 October 2026 · 2 · The record · “…link to it from a German page says so. 1 October 2026 The record page…” · not a measurement
  69. 1 October 2026 · 2 · The record · “…gave them nothing more to go on. 1 October 2026 The download…” · not a measurement
  70. 6 · 2 · The record · “…page and the study page, and section 6 of the study page…” · not bound to a source on the page yet
  71. 1 October 2026 · 2 · The record · “…nothing else to recompute them from. 1 October 2026 The record page…” · not a measurement
  72. 1 October 2026 · 2 · The record · “…reader while translating the page. 1 October 2026 The journey page…” · not a measurement
  73. 1 October 2026 · 2 · The record · “…the headline; it does not replace it. 1 October 2026 The journey page…” · not a measurement
  74. 1 October 2026 · 2 · The record · “…the two pages now say the same thing. 1 October 2026 The journey page…” · not a measurement
  75. 1 October 2026 · 2 · The record · “…two answers to the same question. 1 October 2026 The journey page…” · not a measurement
  76. 17 September · 2 · The record · “…sentence the record page replaced on 17 September. before “No second…” · not a measurement
  77. 17 September · 2 · The record · “…is the reason given in the entry of 17 September below, and that…” · not a measurement
  78. 1 October 2026 · 2 · The record · “…entry changed the record page only. 1 October 2026 A sentence on the…” · not a measurement
  79. 1 October 2026 · 2 · The record · “…had the two the wrong way round. 1 October 2026 The record page…” · not a measurement
  80. 1 October 2026 · 2 · The record · “…ratification, a plan and not a result. 1 October 2026 The legal notice…” · not a measurement
  81. 1 October 2026 · 2 · The record · “…a statement the German page does not. 1 October 2026 The study page…” · not a measurement
  82. 0.288 · 2 · The record · “…“in the direction that flatters it”; “0.288 and 0.135, against…” · not bound to a source on the page yet
  83. 0.135 · 2 · The record · “…that flatters it”; “0.288 and 0.135, against a floor…” · not bound to a source on the page yet
  84. 0.50 · 2 · The record · “…“0.288 and 0.135, against a floor of 0.50 to 0.61” after “so…” · not bound to a source on the page yet
  85. 0.61 · 2 · The record · “…and 0.135, against a floor of 0.50 to 0.61” after “so no…” · not bound to a source on the page yet
  86. 0.288 · 2 · The record · “…“in the direction that raises it”; “0.288 for cosmetic…” · not bound to a source on the page yet
  87. 0.135 · 2 · The record · “…it”; “0.288 for cosmetic rewording and 0.135 for rewording that…” · not bound to a source on the page yet
  88. 0.50 · 2 · The record · “…adds a constraint, against a floor of 0.50 to 0.61” A reader…” · not bound to a source on the page yet
  89. 0.61 · 2 · The record · “…constraint, against a floor of 0.50 to 0.61” A reader of the…” · not bound to a source on the page yet
  90. 1 October 2026 · 2 · The record · “…prints both with their classes. 1 October 2026 The record page…” · not a measurement
  91. 25 September 2026 · 2 · The record · “…for the choice is what can be said. 25 September 2026 The method page…” · not a measurement
  92. 25 September 2026 · 2 · The record · “…claim that the fix moves the outcome. 25 September 2026 Three disclosures…” · not a measurement
  93. 5 · 2 · The record · “…of a paragraph deep in section 5 of the study page;…” · not bound to a source on the page yet
  94. 4 July 2026 · 2 · The record · “…“What this does not show”, carrying 4 July 2026 and 22 August and…” · not a measurement
  95. 22 August · 2 · The record · “…not show”, carrying 4 July 2026 and 22 August and the word…” · not a measurement
  96. 30 August 2026 · 2 · The record · “…plan in this practice is dated 30 August 2026, itself still a…” · not a measurement
  97. 24 September 2026 · 2 · The record · “…with two dates a reader can check. 24 September 2026 The study page’s…” · not a measurement
  98. 0.29 · 2 · The record · “…three answer engines. before Jaccard 0.29 to 0.50 after…” · not bound to a source on the page yet
  99. 0.50 · 2 · The record · “…answer engines. before Jaccard 0.29 to 0.50 after Jaccard 0.29…” · not bound to a source on the page yet
  100. 0.29 · 2 · The record · “…Jaccard 0.29 to 0.50 after Jaccard 0.29 to 0.50 across…” · not bound to a source on the page yet
  101. 0.50 · 2 · The record · “…0.29 to 0.50 after Jaccard 0.29 to 0.50 across three…” · not bound to a source on the page yet
  102. 5.1 · 2 · The record · “…is not a correction of it. Section 5.1 of the cited paper…” · not bound to a source on the page yet
  103. 0.29 · 2 · The record · “…medians one engine at a time, 0.29 to 0.31, 0.33 to…” · not bound to a source on the page yet
  104. 0.31 · 2 · The record · “…medians one engine at a time, 0.29 to 0.31, 0.33 to 0.40 and…” · not bound to a source on the page yet
  105. 0.33 · 2 · The record · “…one engine at a time, 0.29 to 0.31, 0.33 to 0.40 and 0.50,…” · not bound to a source on the page yet
  106. 0.40 · 2 · The record · “…at a time, 0.29 to 0.31, 0.33 to 0.40 and 0.50, and 0.29…” · not bound to a source on the page yet
  107. 0.50 · 2 · The record · “…a time, 0.29 to 0.31, 0.33 to 0.40 and 0.50, and 0.29 to 0.50…” · not bound to a source on the page yet
  108. 0.29 · 2 · The record · “…to 0.31, 0.33 to 0.40 and 0.50, and 0.29 to 0.50 is the…” · not bound to a source on the page yet
  109. 0.50 · 2 · The record · “…0.33 to 0.40 and 0.50, and 0.29 to 0.50 is the correct…” · not bound to a source on the page yet
  110. 0.21 · 2 · The record · “…come from different engines and sit 0.21 apart on the same…” · not bound to a source on the page yet
  111. 22 September 2026 · 2 · The record · “…number stays. It now says what it is. 22 September 2026 The privacy page…” · not a measurement
  112. 59 · 2 · The record · “…timestamp was available and roughly 59 hours earlier. The…” · not bound to a source on the page yet
  113. 29 · 2 · The record · “…des Kontos angenommen, spätestens am 29. März 2026”;…” · not bound to a source on the page yet
  114. 2026 · 2 · The record · “…angenommen, spätestens am 29. März 2026”; “„Spätestens am…” · not bound to a source on the page yet
  115. 29 · 2 · The record · “…am 29. März 2026”; “„Spätestens am 29. März 2026“ steht…” · not bound to a source on the page yet
  116. 2026 · 2 · The record · “…März 2026”; “„Spätestens am 29. März 2026“ steht dort, weil…” · not bound to a source on the page yet
  117. 2026-03-26 · 2 · The record · “…war” after “Das Konto wurde am 2026-03-26 23:01:52 UTC…” · not a measurement
  118. 23 · 2 · The record · “…after “Das Konto wurde am 2026-03-26 23:01:52 UTC…” · not bound to a source on the page yet
  119. 01 · 2 · The record · “…“Das Konto wurde am 2026-03-26 23:01:52 UTC angelegt,…” · not bound to a source on the page yet
  120. 52 · 2 · The record · “…“Das Konto wurde am 2026-03-26 23:01:52 UTC angelegt, das…” · not bound to a source on the page yet
  121. 27 · 2 · The record · “…23:01:52 UTC angelegt, das ist der 27. März 2026, 00:01…” · not bound to a source on the page yet
  122. 2026 · 2 · The record · “…UTC angelegt, das ist der 27. März 2026, 00:01 MEZ”, with…” · not bound to a source on the page yet
  123. 00 · 2 · The record · “…angelegt, das ist der 27. März 2026, 00:01 MEZ”, with the…” · not bound to a source on the page yet
  124. 01 · 2 · The record · “…das ist der 27. März 2026, 00:01 MEZ”, with the…” · not bound to a source on the page yet
  125. 28 · 2 · The record · “…date; and “kein gesondert nach Art. 28 DSGVO…” · not bound to a source on the page yet
  126. 23 · 2 · The record · “…the zone are printed with it because 23:01:52 UTC on the…” · not bound to a source on the page yet
  127. 01 · 2 · The record · “…zone are printed with it because 23:01:52 UTC on the 26th…” · not bound to a source on the page yet
  128. 52 · 2 · The record · “…zone are printed with it because 23:01:52 UTC on the 26th is…” · not bound to a source on the page yet
  129. 00 · 2 · The record · “…it because 23:01:52 UTC on the 26th is 00:01 CET on the…” · not bound to a source on the page yet
  130. 01 · 2 · The record · “…because 23:01:52 UTC on the 26th is 00:01 CET on the 27th: a…” · not bound to a source on the page yet
  131. 2 September · 2 · The record · “…contradicts it; what was missing on 2 September was a separately…” · not a measurement
  132. 28 · 2 · The record · “…was a separately concluded Art. 28 contract and a…” · not bound to a source on the page yet
  133. 22 September 2026 · 2 · The record · “…here, and is not named on the page. 22 September 2026 A section listing…” · not a measurement
  134. 10 · 2 · The record · “…mirrors it. The colophon moved from 10 to 9. before…” · not bound to a source on the page yet
  135. 9 · 2 · The record · “…it. The colophon moved from 10 to 9. before “Earlier…” · not bound to a source on the page yet
  136. 9 · 2 · The record · “…to be captured.”, under the heading “9 · Earlier work” after…” · not a measurement
  137. 17 September 2026 · 2 · The record · “…failure this record exists to catch. 17 September 2026 The study page…” · not a measurement
  138. 3 September 2026 · 2 · The record · “…paragraph changed. before “Corrected 3 September 2026. What follows is…” · not a measurement
  139. 3 September 2026 · 2 · The record · “…between-wave term…” after “Corrected 3 September 2026. What matters is…” · not a measurement
  140. 17 September 2026 · 2 · The record · “…day carrying its own mark, “Corrected 17 September 2026” The clause was…” · not a measurement
  141. 17 September 2026 · 2 · The record · “…a stale claim does the most damage. 17 September 2026 Four dated…” · not a measurement
  142. 3 September 2026 · 2 · The record · “…mark on the record page, “Corrected 3 September 2026”, closing a…” · not a measurement
  143. 17 September · 2 · The record · “…whose first half had been rewritten on 17 September; and no mark at…” · not a measurement
  144. 17 September 2026 · 2 · The record · “…it belongs to; and “Corrected 17 September 2026” on the journey…” · not a measurement
  145. 17 September 2026 · 2 · The record · “…which is how the dates came apart. 17 September 2026 The record page…” · not a measurement
  146. section 1 · 2 · The record · “…counting a column. The two subsets in section 1 are not rows…” · not a measurement
  147. at 2 or more · 2 · The record · “…is the cells CSV’s within_pairs column at 2 or more.” 52 of 65 was…” · not a measurement
  148. 52 · 2 · The record · “…within_pairs column at 2 or more.” 52 of 65 was added to…” · not bound to a source on the page yet
  149. 65 · 2 · The record · “…column at 2 or more.” 52 of 65 was added to the…” · not bound to a source on the page yet
  150. 17 September 2026 · 2 · The record · “…reader one clause rather than a page. 17 September 2026 1,889 came off the…” · not a measurement
  151. 1,889 · 2 · The record · “…rather than a page. 17 September 2026 1,889 came off the…” · not bound to a source on the page yet
  152. 2,000 · 2 · The record · “…populations. before “it now runs 2,000 permutations over…” · not bound to a source on the page yet
  153. 1,889 · 2 · The record · “…“it now runs 2,000 permutations over 1,889 cells in under…” · not bound to a source on the page yet
  154. 1,889 · 2 · The record · “…than the version it replaced. That 1,889 is a working count…” · not bound to a source on the page yet
  155. 3 September · 2 · The record · “…after neither sentence The entry dated 3 September below qualified…” · not a measurement
  156. 1,889 · 2 · The record · “…dated 3 September below qualified 1,889 rather than…” · not bound to a source on the page yet
  157. 17 September 2026 · 2 · The record · “…the analysis code is not published. 17 September 2026 The study page’s…” · not a measurement
  158. −0.086 · 2 · The record · “…one wave only. That cell was empty. −0.086 is a figure this…” · not bound to a source on the page yet
  159. −0.086 · 2 · The record · “…row “Floor from one wave only” after −0.086, the median of…” · not bound to a source on the page yet
  160. 60 · 2 · The record · “…the median of delta_text over those 60 cells, read from…” · not bound to a source on the page yet
  161. −0.032 · 2 · The record · “…is the worst in the table, against −0.032 on the…” · not bound to a source on the page yet
  162. −0.065 · 2 · The record · “…−0.032 on the better-measured half and −0.065 over all 103, so…” · not bound to a source on the page yet
  163. 103 · 2 · The record · “…half and −0.065 over all 103, so the blank was…” · not bound to a source on the page yet
  164. 17 September 2026 · 2 · The record · “…the cell is typed back into the page. 17 September 2026 The defensible…” · not a measurement
  165. 85 · 2 · The record · “…figure stopped being a range. 85 of 103 leads…” · not bound to a source on the page yet
  166. 103 · 2 · The record · “…figure stopped being a range. 85 of 103 leads wherever it…” · not bound to a source on the page yet
  167. 103 · 2 · The record · “…two better-measured subsets of those 103 stand beside it as…” · not bound to a source on the page yet
  168. 52 · 2 · The record · “…beside it as subsets. One of them, 52 of the 65 cells…” · not bound to a source on the page yet
  169. 65 · 2 · The record · “…it as subsets. One of them, 52 of the 65 cells whose floor…” · not bound to a source on the page yet
  170. 76.7 · 2 · The record · “…“the defensible figure is a range, 76.7 to 82.5 percent,…” · not bound to a source on the page yet
  171. 82.5 · 2 · The record · “…defensible figure is a range, 76.7 to 82.5 percent, and which…” · not bound to a source on the page yet
  172. 82.5 · 2 · The record · “…how much measurement a cell carries”; “82.5 percent sits at…” · not bound to a source on the page yet
  173. 76.7 · 2 · The record · “…of a range whose pessimistic end is 76.7”; “the same way…” · not bound to a source on the page yet
  174. 76.7 · 2 · The record · “…“the same way the headline is read as 76.7 to 82.5 percent”…” · not bound to a source on the page yet
  175. 82.5 · 2 · The record · “…way the headline is read as 76.7 to 82.5 percent” after 85…” · not bound to a source on the page yet
  176. 85 · 2 · The record · “…is read as 76.7 to 82.5 percent” after 85 of 103 as the…” · not bound to a source on the page yet
  177. 103 · 2 · The record · “…as 76.7 to 82.5 percent” after 85 of 103 as the headline,…” · not bound to a source on the page yet
  178. 33 · 2 · The record · “…after 85 of 103 as the headline, with 33 of the 43 cells…” · not bound to a source on the page yet
  179. 43 · 2 · The record · “…of 103 as the headline, with 33 of the 43 cells whose floor…” · not bound to a source on the page yet
  180. 52 · 2 · The record · “…floor is measured from both waves and 52 of the 65 whose…” · not bound to a source on the page yet
  181. 65 · 2 · The record · “…measured from both waves and 52 of the 65 whose floor rests…” · not bound to a source on the page yet
  182. 76.7 · 2 · The record · “…it rather than bracketing it; and 76.7 named as the share…” · not bound to a source on the page yet
  183. 33 · 2 · The record · “…A subgroup estimate is not a bound. 33 of 43 is the same…” · not bound to a source on the page yet
  184. 43 · 2 · The record · “…estimate is not a bound. 33 of 43 is the same count…” · not bound to a source on the page yet
  185. 76.7 · 2 · The record · “…where the truth stops, and printing 76.7 to 82.5 as a range…” · not bound to a source on the page yet
  186. 82.5 · 2 · The record · “…the truth stops, and printing 76.7 to 82.5 as a range told a…” · not bound to a source on the page yet
  187. 52 · 2 · The record · “…which the design does not support. 52 of 65 was…” · not bound to a source on the page yet
  188. 65 · 2 · The record · “…the design does not support. 52 of 65 was recomputed…” · not bound to a source on the page yet
  189. 103 · 2 · The record · “…rather than carried across: of the 103 cells with a…” · not bound to a source on the page yet
  190. 38 · 2 · The record · “…across: of the 103 cells with a floor, 38 rest on a single…” · not bound to a source on the page yet
  191. 65 · 2 · The record · “…on a single repeat pair, which leaves 65, and 52 of those…” · not bound to a source on the page yet
  192. 52 · 2 · The record · “…repeat pair, which leaves 65, and 52 of those moved. It…” · not bound to a source on the page yet
  193. 85 · 2 · The record · “…the file before every build, as 85 of 103 and 33 of…” · not bound to a source on the page yet
  194. 103 · 2 · The record · “…the file before every build, as 85 of 103 and 33 of 43…” · not bound to a source on the page yet
  195. 33 · 2 · The record · “…before every build, as 85 of 103 and 33 of 43 already…” · not bound to a source on the page yet
  196. 43 · 2 · The record · “…every build, as 85 of 103 and 33 of 43 already were. 17…” · not bound to a source on the page yet
  197. 17 September 2026 · 2 · The record · “…85 of 103 and 33 of 43 already were. 17 September 2026 The study stopped…” · not a measurement
  198. 17 September 2026 · 2 · The record · “…what the figures are allowed to mean. 17 September 2026 The study stopped…” · not a measurement
  199. 3 September · 2 · The record · “…the account-state entry below, dated 3 September, where it is now…” · not a measurement
  200. 17 September 2026 · 2 · The record · “…reason; this entry is its correction. 17 September 2026 The table of…” · not a measurement
  201. 17 September 2026 · 2 · The record · “…opposite of what the file says it is. 17 September 2026 The record page…” · not a measurement
  202. −0.042 · 2 · The record · “…in both places The spread runs −0.042 to −0.066 and the…” · not bound to a source on the page yet
  203. −0.066 · 2 · The record · “…both places The spread runs −0.042 to −0.066 and the interval…” · not bound to a source on the page yet
  204. −0.058 · 2 · The record · “…runs −0.042 to −0.066 and the interval −0.058 to −0.031. What is…” · not bound to a source on the page yet
  205. −0.031 · 2 · The record · “…to −0.066 and the interval −0.058 to −0.031. What is true is…” · not bound to a source on the page yet
  206. −0.066 · 2 · The record · “…the spread lies outside the interval: −0.066 sits below −0.058,…” · not bound to a source on the page yet
  207. −0.058 · 2 · The record · “…the interval: −0.066 sits below −0.058, while −0.042 sits…” · not bound to a source on the page yet
  208. −0.042 · 2 · The record · “…−0.066 sits below −0.058, while −0.042 sits inside it. So…” · not bound to a source on the page yet
  209. 17 September 2026 · 2 · The record · “…the comparison between them was not. 17 September 2026 Half of a sentence…” · not a measurement
  210. 88.5 · 2 · The record · “…before “two other conventions give 88.5 and 100.0, or 84.0…” · not bound to a source on the page yet
  211. 100.0 · 2 · The record · “…“two other conventions give 88.5 and 100.0, or 84.0 and 92.3,…” · not bound to a source on the page yet
  212. 84.0 · 2 · The record · “…conventions give 88.5 and 100.0, or 84.0 and 92.3, for the…” · not bound to a source on the page yet
  213. 92.3 · 2 · The record · “…give 88.5 and 100.0, or 84.0 and 92.3, for the upper two…” · not bound to a source on the page yet
  214. 88.5 · 2 · The record · “…92.3, for the upper two bands” after “88.5 and 100.0 for the…” · not bound to a source on the page yet
  215. 100.0 · 2 · The record · “…the upper two bands” after “88.5 and 100.0 for the upper two…” · not bound to a source on the page yet
  216. 84.0 · 2 · The record · “…and 100.0 for the upper two bands, or 84.0 and 92.3 for the…” · not bound to a source on the page yet
  217. 92.3 · 2 · The record · “…for the upper two bands, or 84.0 and 92.3 for the middle…” · not bound to a source on the page yet
  218. 84.0 · 2 · The record · “…the convention that produces them, 84.0 and 92.3 are the…” · not bound to a source on the page yet
  219. 92.3 · 2 · The record · “…that produces them, 84.0 and 92.3 are the shares of…” · not bound to a source on the page yet
  220. 17 September 2026 · 2 · The record · “…this class of error survives a check. 17 September 2026 The seven capture…” · not a measurement
  221. 17 September 2026 · 2 · The record · “…every time the archive grows. 17 September 2026 The capture window…” · not a measurement
  222. 4 July · 2 · The record · “…before “seven capture waves between 4 July and 25 August…” · not a measurement
  223. 25 August 2026 · 2 · The record · “…capture waves between 4 July and 25 August 2026” after “seven…” · not a measurement
  224. 4 July · 2 · The record · “…whose first captures fall between 4 July and 25 August…” · not a measurement
  225. 25 August 2026 · 2 · The record · “…first captures fall between 4 July and 25 August 2026” Three of the…” · not a measurement
  226. 27 August · 2 · The record · “…several days and the last ran on into 27 August, which the study’s…” · not a measurement
  227. 17 September 2026 · 2 · The record · “…the record and journey pages as well. 17 September 2026 The study page…” · not a measurement
  228. 20260705 · 2 · The record · “…a cell: the fourth, makefield-cycle0-20260705” after “the…” · not a measurement
  229. 20260705 · 2 · The record · “…the sentence lists makefield-cycle0-20260705 second of the…” · not a measurement
  230. 17 September 2026 · 2 · The record · “…above the table that contradicts it. 17 September 2026 The download card…” · not a measurement
  231. 9 · 2 · The record · “…it no longer has. before contents: “9 sections” after…” · not bound to a source on the page yet
  232. 24 · 2 · The record · “…“9 sections” after contents: “24 top-level keys”…” · not bound to a source on the page yet
  233. 24 · 2 · The record · “…keys” The published summary file has 24 keys at its top…” · not bound to a source on the page yet
  234. 17 September 2026 · 2 · The record · “…again without the build saying so. 17 September 2026 The record page’s…” · not a measurement
  235. 17 September 2026 · 2 · The record · “…can weigh against the published files. 17 September 2026 The method page…” · not a measurement
  236. 17 September 2026 · 2 · The record · “…rather than filling it in from memory. 17 September 2026 The journey page…” · not a measurement
  237. 30 August 2026 · 2 · The record · “…pages use The registration file dated 30 August 2026 carries a status…” · not a measurement
  238. 17 September 2026 · 2 · The record · “…on the page carrying the least detail. 17 September 2026 The record page’s…” · not a measurement
  239. 10 September · 2 · The record · “…the thing that checks it. I decided on 10 September that they come off…” · not a measurement
  240. 17 September 2026 · 2 · The record · “…on the page depended on the section. 17 September 2026 A promise that the…” · not a measurement
  241. 17 September 2026 · 2 · The record · “…is corrected in every copy or in none. 17 September 2026 The colophon said…” · not a measurement
  242. 17 September 2026 · 2 · The record · “…that a reader checks by looking up. 17 September 2026 The colophon…” · not a measurement
  243. 17 September 2026 · 2 · The record · “…is unchanged and is on the same page. 17 September 2026 The first…” · not a measurement
  244. 29 August 2026 · 2 · The record · “…of its eleven checks when it ran on 29 August 2026, a read that has…” · not a measurement
  245. 6 · 2 · The record · “…the verifier result left in section 6 where it already…” · not bound to a source on the page yet
  246. 6 · 2 · The record · “…and not softened. It is in section 6, with its date and…” · not bound to a source on the page yet
  247. 3 September · 2 · The record · “…corrected in the same position on 3 September, for its missing…” · not a measurement
  248. 17 September 2026 · 2 · The record · “…to be edited for what it puts first. 17 September 2026 Two clauses left…” · not a measurement
  249. 1 · 2 · The record · “…values, which is the correction of 1 and 3 September;…” · not bound to a source on the page yet
  250. 3 September · 2 · The record · “…which is the correction of 1 and 3 September; the journey page…” · not a measurement
  251. 17 September 2026 · 2 · The record · “…and the name this record lists it by. 17 September 2026 Three pointers…” · not a measurement
  252. 17 September 2026 · 2 · The record · “…the record page a pointer that works. 17 September 2026 This page stated…” · not a measurement
  253. 119 · 2 · The record · “…marked false where it stands, with the 119 tied groups and…” · not bound to a source on the page yet
  254. 20 · 2 · The record · “…with the 119 tied groups and the 20 holding three rows…” · not bound to a source on the page yet
  255. 10 September 2026 · 2 · The record · “…were found by a review of this site on 10 September 2026, and neither is a…” · not a measurement
  256. 10 September · 2 · The record · “…way: the placeholder was filled on 10 September and that has its…” · not a measurement
  257. 10 September 2026 · 2 · The record · “…and that has its own entry above. 10 September 2026 The privacy notice…” · not a measurement
  258. 29 March 2026 · 2 · The record · “…account was set up and no later than 29 March 2026, current version…” · not a measurement
  259. 6.4 · 2 · The record · “…than 29 March 2026, current version 6.4 of 3 April 2026;…” · not bound to a source on the page yet
  260. 3 April 2026 · 2 · The record · “…29 March 2026, current version 6.4 of 3 April 2026; transfers to the…” · not a measurement
  261. 10 July 2023 · 2 · The record · “…the Commission’s adequacy decision of 10 July 2023, with the…” · not a measurement
  262. 6.2 · 2 · The record · “…contractual clauses under section 6.2 of the addendum if…” · not bound to a source on the page yet
  263. 10 September 2026 · 2 · The record · “…lapses; the certification checked on 10 September 2026 in the US…” · not a measurement
  264. 29 March 2026 · 2 · The record · “…in the official list. “No later than 29 March 2026” goes exactly as…” · not a measurement
  265. 3 September 2026 · 2 · The record · “…day, not that it was created that day. 3 September 2026 The Impressum has…” · not a measurement
  266. 9 · 2 · The record · “…of its own after Britzkestraße 9, 12347 Berlin, on…” · not bound to a source on the page yet
  267. 12347 · 2 · The record · “…of its own after Britzkestraße 9, 12347 Berlin, on both…” · not bound to a source on the page yet
  268. 171 · 2 · The record · “…holds a Zustellungsvollmacht under § 171 ZPO, and no court…” · not bound to a source on the page yet
  269. 28 · 2 · The record · “…left on this site that day, the Art. 28 line; it was…” · not bound to a source on the page yet
  270. 10 September 2026 · 2 · The record · “…the Art. 28 line; it was filled on 10 September 2026 and has the entry…” · not a measurement
  271. 3 September 2026 · 2 · The record · “…2026 and has the entry above. 3 September 2026 The 108 cells…” · not a measurement
  272. 108 · 2 · The record · “…the entry above. 3 September 2026 The 108 cells behind the…” · not bound to a source on the page yet
  273. 103 · 2 · The record · “…figure on this site were never the 103, and are not a…” · not bound to a source on the page yet
  274. 111 · 2 · The record · “…the 103, and are not a subset of the 111 either. Two of…” · not bound to a source on the page yet
  275. 108 · 2 · The record · “…effect can produce it”, with 108 printed on three…” · not bound to a source on the page yet
  276. 108 · 2 · The record · “…effect can produce it”, with the 108 traced to the 123…” · not bound to a source on the page yet
  277. 123 · 2 · The record · “…it”, with the 108 traced to the 123 cells the…” · not bound to a source on the page yet
  278. 100 · 2 · The record · “…123 cells the gap-pairs file covers, 100 of them inside the…” · not bound to a source on the page yet
  279. 111 · 2 · The record · “…file covers, 100 of them inside the 111 and 8 outside, and…” · not bound to a source on the page yet
  280. 8 · 2 · The record · “…covers, 100 of them inside the 111 and 8 outside, and the…” · not bound to a source on the page yet
  281. 3 September · 2 · The record · “…reason that file was published on 3 September. It covers 123…” · not a measurement
  282. 123 · 2 · The record · “…published on 3 September. It covers 123 cells, because a…” · not bound to a source on the page yet
  283. 108 · 2 · The record · “…one distinct gap and fall out, leaving 108. Restricted to the…” · not bound to a source on the page yet
  284. 100 · 2 · The record · “…out, leaving 108. Restricted to the 100 that are also in…” · not bound to a source on the page yet
  285. −0.042 · 2 · The record · “…also in the floor test, the median is −0.042 with 79 of 100…” · not bound to a source on the page yet
  286. 79 · 2 · The record · “…floor test, the median is −0.042 with 79 of 100 negative.…” · not bound to a source on the page yet
  287. 100 · 2 · The record · “…test, the median is −0.042 with 79 of 100 negative. Two…” · not bound to a source on the page yet
  288. −0.046 · 2 · The record · “…is untrue of them. Dropping both gives −0.046 over 106, with 85…” · not bound to a source on the page yet
  289. 106 · 2 · The record · “…them. Dropping both gives −0.046 over 106, with 85 negative…” · not bound to a source on the page yet
  290. 85 · 2 · The record · “…both gives −0.046 over 106, with 85 negative and 21…” · not bound to a source on the page yet
  291. 21 · 2 · The record · “…−0.046 over 106, with 85 negative and 21 positive. The…” · not bound to a source on the page yet
  292. 3 September 2026 · 2 · The record · “…was the clause doing the persuading. 3 September 2026 The study’s…” · not a measurement
  293. 1 September · 2 · The record · “…shares, and the recipe retracted on 1 September was still live in…” · not a measurement
  294. 0.200 · 2 · The record · “…before a Range column reading “up to 0.200 / 0.200 to 0.266 /…” · not bound to a source on the page yet
  295. 0.200 · 2 · The record · “…a Range column reading “up to 0.200 / 0.200 to 0.266 / 0.266…” · not bound to a source on the page yet
  296. 0.266 · 2 · The record · “…column reading “up to 0.200 / 0.200 to 0.266 / 0.266 to 0.321 /…” · not bound to a source on the page yet
  297. 0.266 · 2 · The record · “…“up to 0.200 / 0.200 to 0.266 / 0.266 to 0.321 / above…” · not bound to a source on the page yet
  298. 0.321 · 2 · The record · “…to 0.200 / 0.200 to 0.266 / 0.266 to 0.321 / above 0.321”;…” · not bound to a source on the page yet
  299. 0.321 · 2 · The record · “…to 0.266 / 0.266 to 0.321 / above 0.321”; and “split at…” · not bound to a source on the page yet
  300. 0.0357 · 2 · The record · “…the observed span of each rank block, 0.0357 to 0.1985 and so…” · not bound to a source on the page yet
  301. 0.1985 · 2 · The record · “…span of each rank block, 0.0357 to 0.1985 and so on; and…” · not bound to a source on the page yet
  302. 57.7 · 2 · The record · “…by those printed boundary values gives 57.7, 81.5, 95.8 and…” · not bound to a source on the page yet
  303. 81.5 · 2 · The record · “…printed boundary values gives 57.7, 81.5, 95.8 and 96.2.…” · not bound to a source on the page yet
  304. 95.8 · 2 · The record · “…boundary values gives 57.7, 81.5, 95.8 and 96.2. The…” · not bound to a source on the page yet
  305. 96.2 · 2 · The record · “…values gives 57.7, 81.5, 95.8 and 96.2. The dataset…” · not bound to a source on the page yet
  306. 3 · 2 · The record · “…page also sent a reader to section 3 for three values…” · not bound to a source on the page yet
  307. 2 · 2 · The record · “…3 for three values printed in section 2. 3 September 2026…” · not bound to a source on the page yet
  308. 3 September 2026 · 2 · The record · “…for three values printed in section 2. 3 September 2026 The record page…” · not a measurement
  309. 26 · 2 · The record · “…counts, and the verifier table showed 26 twice with no word…” · not bound to a source on the page yet
  310. 26 · 2 · The record · “…with no population; two rows reading 26 after four rows…” · not bound to a source on the page yet
  311. 111 · 2 · The record · “…the engine list stated as all 111 comparable cells…” · not bound to a source on the page yet
  312. 103 · 2 · The record · “…pointed at the study’s list over the 103; and a line saying…” · not bound to a source on the page yet
  313. 111 · 2 · The record · “…front of them. The engine list sums to 111 and the study’s…” · not bound to a source on the page yet
  314. 103 · 2 · The record · “…sums to 111 and the study’s sums to 103, so the two looked…” · not bound to a source on the page yet
  315. 3 September 2026 · 2 · The record · “…rather than that it is a coincidence. 3 September 2026 The journey page…” · not a measurement
  316. 4 · 2 · The record · “…to what the page does, with station 4’s two figure…” · not bound to a source on the page yet
  317. 2 · 2 · The record · “…into an interpretation mark Rules 2 and 5 answer the…” · not bound to a source on the page yet
  318. 5 · 2 · The record · “…an interpretation mark Rules 2 and 5 answer the…” · not bound to a source on the page yet
  319. 4 · 2 · The record · “…what it implies, both named in station 4. Rules 1, 3, 4 and…” · not bound to a source on the page yet
  320. 1 · 2 · The record · “…both named in station 4. Rules 1, 3, 4 and 6 have…” · not bound to a source on the page yet
  321. 3 · 2 · The record · “…both named in station 4. Rules 1, 3, 4 and 6 have no…” · not bound to a source on the page yet
  322. 4 · 2 · The record · “…both named in station 4. Rules 1, 3, 4 and 6 have no…” · not bound to a source on the page yet
  323. 6 · 2 · The record · “…named in station 4. Rules 1, 3, 4 and 6 have no…” · not bound to a source on the page yet
  324. 2136073e · 2 · The record · “…said fails, present tense, about build 2136073e, which the record…” · not a measurement
  325. 3 September 2026 · 2 · The record · “…claim and ending on a conclusion. 3 September 2026 Three sentences on…” · not a measurement
  326. 4 · 2 · The record · “…is the published rows. And section 4 of the method page…” · not bound to a source on the page yet
  327. 3 September 2026 · 2 · The record · “…contradicted the page it sits on. 3 September 2026 A parenthetical…” · not a measurement
  328. 2000 · 2 · The record · “…and Apple)”; “Written by hand”; “2000 permutations over…” · not bound to a source on the page yet
  329. 1889 · 2 · The record · “…by hand”; “2000 permutations over 1889 cells”; “My…” · not bound to a source on the page yet
  330. 1,889 · 2 · The record · “…DACH”; a colophon without the claim; 1,889 named as a working…” · not bound to a source on the page yet
  331. 1,889 · 2 · The record · “…a colophon whose other claims can. The 1,889 read as a fourth…” · not bound to a source on the page yet
  332. 111 · 2 · The record · “…read as a fourth population beside 111, 103 and 108 and…” · not bound to a source on the page yet
  333. 103 · 2 · The record · “…as a fourth population beside 111, 103 and 108 and is…” · not bound to a source on the page yet
  334. 108 · 2 · The record · “…fourth population beside 111, 103 and 108 and is none of…” · not bound to a source on the page yet
  335. 3 September 2026 · 2 · The record · “…files that came out that morning. 3 September 2026 Account state was…” · not a measurement
  336. 31 August · 2 · The record · “…of error as the sign-in correction of 31 August, which is why it…” · not a measurement
  337. 3 September 2026 · 2 · The record · “…the headline that contradicted it. 3 September 2026 The reason given…” · not a measurement
  338. 4 July · 2 · The record · “…way to show from here that the tool on 4 July was the tool on 25…” · not a measurement
  339. 25 August · 2 · The record · “…the tool on 4 July was the tool on 25 August A review ruled the…” · not a measurement
  340. 3 September 2026 · 2 · The record · “…which is real and now stated as one. 3 September 2026 A third party’s…” · not a measurement
  341. 4 July 2026 · 2 · The record · “…was caught on the day of that wave, 4 July 2026, and both…” · not a measurement
  342. 3 September 2026 · 2 · The record · “…both of which used the correct domain. 3 September 2026 The reason the…” · not a measurement
  343. 38 · 2 · The record · “…it rather than against memory. Of the 38 runs in the…” · not bound to a source on the page yet
  344. 3 September 2026 · 2 · The record · “…archive and why none of it can go out. 3 September 2026 A third pass over…” · not a measurement
  345. 119 · 2 · The record · “…touches is exactly two rows”, with the 119 tied groups and…” · not bound to a source on the page yet
  346. 20 · 2 · The record · “…with the 119 tied groups and the 20 holding three rows…” · not bound to a source on the page yet
  347. 119 · 2 · The record · “…universal was false. The file holds 119 tied gap-day…” · not bound to a source on the page yet
  348. 20 · 2 · The record · “…tied gap-day groups within cells and 20 of them hold three…” · not bound to a source on the page yet
  349. 37 · 2 · The record · “…rows; the claim is true only of the 37 groups at a cell’s…” · not bound to a source on the page yet
  350. 3 September 2026 · 2 · The record · “…finished rather than improved further. 3 September 2026 The paragraph that…” · not a measurement
  351. -0.0658 · 2 · The record · “…before “the median of the tied rows, -0.0658 and 93”; and the…” · not bound to a source on the page yet
  352. 93 · 2 · The record · “…median of the tied rows, -0.0658 and 93”; and the record…” · not bound to a source on the page yet
  353. −0.047 · 2 · The record · “…and 93”; and the record page listing −0.047 with a 95 percent…” · not bound to a source on the page yet
  354. 95 · 2 · The record · “…the record page listing −0.047 with a 95 percent interval…” · not bound to a source on the page yet
  355. -0.0582 · 2 · The record · “…“the mean of the two tied values, -0.0582 and 90”, with the…” · not bound to a source on the page yet
  356. 90 · 2 · The record · “…of the two tied values, -0.0582 and 90”, with the reason…” · not bound to a source on the page yet
  357. −0.042 · 2 · The record · “…printed; and the record page carrying −0.042 to −0.066 and 83…” · not bound to a source on the page yet
  358. −0.066 · 2 · The record · “…and the record page carrying −0.042 to −0.066 and 83 to 93 of…” · not bound to a source on the page yet
  359. 83 · 2 · The record · “…page carrying −0.042 to −0.066 and 83 to 93 of 108 Every…” · not bound to a source on the page yet
  360. 93 · 2 · The record · “…carrying −0.042 to −0.066 and 83 to 93 of 108 Every tie…” · not bound to a source on the page yet
  361. 108 · 2 · The record · “…−0.042 to −0.066 and 83 to 93 of 108 Every tie in the…” · not bound to a source on the page yet
  362. 119 · 2 · The record · “…above corrects it: the file holds 119 tied gap-day…” · not bound to a source on the page yet
  363. 20 · 2 · The record · “…file holds 119 tied gap-day groups and 20 of them hold three…” · not bound to a source on the page yet
  364. 37 · 2 · The record · “…so the universal is true only of the 37 groups the rule…” · not bound to a source on the page yet
  365. 1 September · 2 · The record · “…is the quartile-band failure of 1 September one level down.…” · not a measurement
  366. 3 September 2026 · 2 · The record · “…point estimate are unchanged by both. 3 September 2026 A third file was…” · not a measurement
  367. 08284347 · 2 · The record · “…three files; floor-test-gap-pairs-08284347.csv, 957 rows, one…” · not a measurement
  368. 957 · 2 · The record · “…floor-test-gap-pairs-08284347.csv, 957 rows, one per wave…” · not bound to a source on the page yet
  369. 37 · 2 · The record · “…wave pairs at the same gap in days and 37 of the 108 do, so…” · not bound to a source on the page yet
  370. 108 · 2 · The record · “…at the same gap in days and 37 of the 108 do, so five…” · not bound to a source on the page yet
  371. -0.042 · 2 · The record · “…tie-breaks give medians from -0.042 to -0.066 and…” · not bound to a source on the page yet
  372. -0.066 · 2 · The record · “…tie-breaks give medians from -0.042 to -0.066 and counts from 83…” · not bound to a source on the page yet
  373. 83 · 2 · The record · “…from -0.042 to -0.066 and counts from 83 to 93 of 108. The…” · not bound to a source on the page yet
  374. 93 · 2 · The record · “…-0.042 to -0.066 and counts from 83 to 93 of 108. The…” · not bound to a source on the page yet
  375. 108 · 2 · The record · “…to -0.066 and counts from 83 to 93 of 108. The direction…” · not bound to a source on the page yet
  376. 3 September 2026 · 2 · The record · “…the second most conservative of them. 3 September 2026 A review ran over…” · not a measurement
  377. aca27635 · 2 · The record · “…of client material”; summary sha256 aca27635… after…” · not a measurement
  378. cf21d8a9 · 2 · The record · “…earlier client work”; summary sha256 cf21d8a9…, 11,684 bytes The…” · not a measurement
  379. 11,684 · 2 · The record · “…work”; summary sha256 cf21d8a9…, 11,684 bytes The gate…” · not bound to a source on the page yet
  380. 3 September 2026 · 2 · The record · “…the zero without the reason for it. 3 September 2026 The study page’s…” · not a measurement
  381. 3 · 2 · The record · “…count of twelve excluded cells Section 3 of this page…” · not bound to a source on the page yet
  382. 3 September · 2 · The record · “…for this page’s own list and has since 3 September; the study page’s…” · not a measurement
  383. 17 September 2026 · 2 · The record · “…same file. This entry is written on 17 September 2026, two weeks late,…” · not a measurement
  384. 2 September 2026 · 2 · The record · “…note and this list did not have it. 2 September 2026 The method page…” · not a measurement
  385. 0.0007 · 2 · The record · “…nowhere. before the two differ by 0.0007, about forty times…” · not bound to a source on the page yet
  386. 0.062 · 2 · The record · “…after the interval on that median is 0.062 wide, about ninety…” · not bound to a source on the page yet
  387. 0.031 · 2 · The record · “…is the factor against the half-width, 0.031. The interval as…” · not bound to a source on the page yet
  388. −0.092 · 2 · The record · “…prints it on every other surface is −0.092 to −0.030, which…” · not bound to a source on the page yet
  389. −0.030 · 2 · The record · “…it on every other surface is −0.092 to −0.030, which is 0.062…” · not bound to a source on the page yet
  390. 0.062 · 2 · The record · “…surface is −0.092 to −0.030, which is 0.062 wide, and 0.062…” · not bound to a source on the page yet
  391. 0.062 · 2 · The record · “…to −0.030, which is 0.062 wide, and 0.062 over 0.0007 is…” · not bound to a source on the page yet
  392. 0.0007 · 2 · The record · “…which is 0.062 wide, and 0.062 over 0.0007 is about ninety.…” · not bound to a source on the page yet
  393. 17 September 2026 · 2 · The record · “…factor. This entry is written on 17 September 2026, two weeks late,…” · not a measurement
  394. 2 September 2026 · 2 · The record · “…the same reason as the entry above it. 2 September 2026 The carve-out…” · not a measurement
  395. 1 September · 2 · The record · “…are in neither file The narrowing on 1 September fixed the word…” · not a measurement
  396. 111 · 2 · The record · “…and absent from the CSV, which holds 111 rows of which 8…” · not bound to a source on the page yet
  397. 8 · 2 · The record · “…the CSV, which holds 111 rows of which 8 have no floor. 1…” · not bound to a source on the page yet
  398. 1 September 2026 · 2 · The record · “…111 rows of which 8 have no floor. 1 September 2026 The request…” · not a measurement
  399. 0 · 2 · The record · “…it checks against. Requests so far: 0. after the same…” · not bound to a source on the page yet
  400. 1 September 2026 · 2 · The record · “…the paragraph above says it does not. 1 September 2026 The claim that…” · not a measurement
  401. 85 · 2 · The record · “…The per-engine cell counts, the 85 of 103, the 33 of…” · not bound to a source on the page yet
  402. 103 · 2 · The record · “…The per-engine cell counts, the 85 of 103, the 33 of 43 and…” · not bound to a source on the page yet
  403. 33 · 2 · The record · “…cell counts, the 85 of 103, the 33 of 43 and the…” · not bound to a source on the page yet
  404. 43 · 2 · The record · “…cell counts, the 85 of 103, the 33 of 43 and the median…” · not bound to a source on the page yet
  405. 1 September 2026 · 2 · The record · “…before the change record started. 1 September 2026 The study page…” · not a measurement
  406. 14 · 2 · The record · “…table described as gap_days banded at 14 and 40 and the…” · not bound to a source on the page yet
  407. 40 · 2 · The record · “…described as gap_days banded at 14 and 40 and the quartile…” · not bound to a source on the page yet
  408. 103 · 2 · The record · “…with the population stated as the 103 cells that have a…” · not bound to a source on the page yet
  409. 111 · 2 · The record · “…the gap-table recipe over all 111 rows gets 28 of 40…” · not bound to a source on the page yet
  410. 28 · 2 · The record · “…recipe over all 111 rows gets 28 of 40 and 49 of 59…” · not bound to a source on the page yet
  411. 40 · 2 · The record · “…recipe over all 111 rows gets 28 of 40 and 49 of 59…” · not bound to a source on the page yet
  412. 49 · 2 · The record · “…over all 111 rows gets 28 of 40 and 49 of 59 instead of…” · not bound to a source on the page yet
  413. 59 · 2 · The record · “…all 111 rows gets 28 of 40 and 49 of 59 instead of 28 of…” · not bound to a source on the page yet
  414. 28 · 2 · The record · “…gets 28 of 40 and 49 of 59 instead of 28 of 35 and 49 of…” · not bound to a source on the page yet
  415. 35 · 2 · The record · “…28 of 40 and 49 of 59 instead of 28 of 35 and 49 of 56, and…” · not bound to a source on the page yet
  416. 49 · 2 · The record · “…and 49 of 59 instead of 28 of 35 and 49 of 56, and the…” · not bound to a source on the page yet
  417. 56 · 2 · The record · “…49 of 59 instead of 28 of 35 and 49 of 56, and the quartile…” · not bound to a source on the page yet
  418. 1 September 2026 · 2 · The record · “…did not determine an answer at all. 1 September 2026 The dataset README…” · not a measurement
  419. 26 · 2 · The record · “…quartiles after equal rank blocks of 26, 26, 25 and 26…” · not bound to a source on the page yet
  420. 26 · 2 · The record · “…after equal rank blocks of 26, 26, 25 and 26 over…” · not bound to a source on the page yet
  421. 25 · 2 · The record · “…after equal rank blocks of 26, 26, 25 and 26 over the…” · not bound to a source on the page yet
  422. 26 · 2 · The record · “…equal rank blocks of 26, 26, 25 and 26 over the cells…” · not bound to a source on the page yet
  423. 1 September 2026 · 2 · The record · “…does not reproduce the shares at all. 1 September 2026 A revision date…” · not a measurement
  424. 1 September 2026 · 2 · The record · “…without going to the repository. 1 September 2026 The record page’s…” · not a measurement
  425. 31 August 2026 · 2 · The record · “…the page itself. before dateModified 31 August 2026, while the…” · not a measurement
  426. 1 September · 2 · The record · “…the colophon and the sitemap said 1 September after one date per…” · not a measurement
  427. 1 September 2026 · 2 · The record · “…that disagree are worse than none. 1 September 2026 The journey page…” · not a measurement
  428. 31 August 2026 · 2 · The record · “…that way the first time. It did not. 31 August 2026 Every claim about…” · not a measurement
  429. 31 August 2026 · 2 · The record · “…modes. An inference is not a record. 31 August 2026 The quartile…” · not a measurement
  430. 56 · 2 · The record · “…were restated at one decimal. before 56, 84, 92 and 96…” · not bound to a source on the page yet
  431. 84 · 2 · The record · “…restated at one decimal. before 56, 84, 92 and 96 percent…” · not bound to a source on the page yet
  432. 92 · 2 · The record · “…at one decimal. before 56, 84, 92 and 96 percent…” · not bound to a source on the page yet
  433. 96 · 2 · The record · “…at one decimal. before 56, 84, 92 and 96 percent after…” · not bound to a source on the page yet
  434. 57.7 · 2 · The record · “…before 56, 84, 92 and 96 percent after 57.7, 84.6, 92.0 and…” · not bound to a source on the page yet
  435. 84.6 · 2 · The record · “…56, 84, 92 and 96 percent after 57.7, 84.6, 92.0 and 96.2…” · not bound to a source on the page yet
  436. 92.0 · 2 · The record · “…92 and 96 percent after 57.7, 84.6, 92.0 and 96.2 percent…” · not bound to a source on the page yet
  437. 96.2 · 2 · The record · “…96 percent after 57.7, 84.6, 92.0 and 96.2 percent The…” · not bound to a source on the page yet
  438. 3 · 3 · How to check any of it · “…3 · How to check any of…” · not a measurement
  439. 103 · 3 · How to check any of it · “…Recomputable from the two CSVs: the 103 cells and the 85…” · not bound to a source on the page yet
  440. 85 · 3 · How to check any of it · “…the two CSVs: the 103 cells and the 85 that moved, the…” · not bound to a source on the page yet
  441. 18 · 3 · How to check any of it · “…moved, the per-engine cell counts, the 18 questions and the…” · not bound to a source on the page yet
  442. 111 · 3 · How to check any of it · “…cell counts, the 18 questions and the 111 rows, the noise…” · not bound to a source on the page yet
  443. 1 September 2026 · 3 · How to check any of it · “…one wave and from both. Checked on 1 September 2026, and again on 2…” · not a measurement
  444. 2 September · 3 · How to check any of it · “…on 1 September 2026, and again on 2 September, by recomputing…” · not a measurement
  445. 3 September · 3 · How to check any of it · “…than by trusting the summary. And from 3 September the within-cell…” · not a measurement
  446. −0.0465 · 3 · How to check any of it · “…CSV published that day: median −0.0465 over 108 cells, 87…” · not bound to a source on the page yet
  447. 108 · 3 · How to check any of it · “…that day: median −0.0465 over 108 cells, 87 negative…” · not bound to a source on the page yet
  448. 87 · 3 · How to check any of it · “…day: median −0.0465 over 108 cells, 87 negative and 21…” · not bound to a source on the page yet
  449. 21 · 3 · How to check any of it · “…over 108 cells, 87 negative and 21 positive,…” · not bound to a source on the page yet
  450. 2 September · 3 · How to check any of it · “…both corrections are worth seeing. On 2 September it replaced a list…” · not a measurement
  451. 3 September · 3 · How to check any of it · “…disagreed about what was checkable. On 3 September it went from six…” · not a measurement
  452. 1 September 2026 · 3 · How to check any of it · “…enough to repeat it. Requests as of 1 September 2026: 0. That count is…” · not a measurement
  453. 0 · 3 · How to check any of it · “…it. Requests as of 1 September 2026: 0. That count is…” · not bound to a source on the page yet
  454. 57.7 · 3 · How to check any of it · “…quarters. Two rules give the published 57.7, 84.6, 92.0 and…” · not bound to a source on the page yet
  455. 84.6 · 3 · How to check any of it · “…Two rules give the published 57.7, 84.6, 92.0 and 96.2;…” · not bound to a source on the page yet
  456. 92.0 · 3 · How to check any of it · “…rules give the published 57.7, 84.6, 92.0 and 96.2; two…” · not bound to a source on the page yet
  457. 96.2 · 3 · How to check any of it · “…the published 57.7, 84.6, 92.0 and 96.2; two others give…” · not bound to a source on the page yet
  458. 88.5 · 3 · How to check any of it · “…84.6, 92.0 and 96.2; two others give 88.5 and 100.0 for the…” · not bound to a source on the page yet
  459. 100.0 · 3 · How to check any of it · “…and 96.2; two others give 88.5 and 100.0 for the upper two…” · not bound to a source on the page yet
  460. 84.0 · 3 · How to check any of it · “…and 100.0 for the upper two bands, or 84.0 and 92.3 for the…” · not bound to a source on the page yet
  461. 92.3 · 3 · How to check any of it · “…for the upper two bands, or 84.0 and 92.3 for the middle…” · not bound to a source on the page yet
  462. 2 September 2026 · 3 · How to check any of it · “…into the dataset README. ⚠ Corrected 2 September 2026. The second…” · not a measurement
  463. 81.5 · 3 · How to check any of it · “…2026. The second alternative read 81.5 and 95.8, which is…” · not bound to a source on the page yet
  464. 95.8 · 3 · How to check any of it · “…The second alternative read 81.5 and 95.8, which is the…” · not bound to a source on the page yet
  465. 0.200 · 3 · How to check any of it · “…values the pages themselves print, 0.200, 0.266 and 0.321,…” · not bound to a source on the page yet
  466. 0.266 · 3 · How to check any of it · “…the pages themselves print, 0.200, 0.266 and 0.321, does…” · not bound to a source on the page yet
  467. 0.321 · 3 · How to check any of it · “…themselves print, 0.200, 0.266 and 0.321, does not…” · not bound to a source on the page yet
  468. 31 October 2026 · 3 · How to check any of it · “…says only what moved. Last revised 31 October 2026.…” · not a measurement

↑ top

Every figure the site keeps in one place

Every key in lib/figures.ts, the one file the site’s figures are read from, and the pages that print it as a sourced number.

↑ top