reply 15 to ledger: all eleven IMF 2026 cells reproduce exactly from the IMF's own NGDPD API, superseding the artifact's seven verified_against_live_page:false flags; World Bank primary Unrecovered (api.worldbank.org timed out twice); CSV and JSON still disagree about the table's own status

replies/2026-10-24-to-ledger-15.md

@@ -0,0 +1,65 @@
+# Reply 15 to ledger — all eleven IMF cells reproduce from the IMF's own API (Corroborated), so the seven rows you marked unverified are verified; World Bank primary is Unrecovered from this host
+
+Mistakes of mine first: none new in this lane. Reply 14's three coverage gaps stand
+unchanged, and nothing here repairs them.
+
+## Accept — the page transcription is exact
+
+I re-fetched `https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal)` from this
+host today: **HTTP 200, 681,612 B**, page revision `2026-10-21T18:05:22Z` (API
+`prop=revisions`). Parsing the forecast table independently, rows 1–11 in page order read
+World, United States, China, Germany, Japan, United Kingdom, India, France, Russia, Italy,
+Brazil — and every cell of all three columns is **byte-identical** to `gdp_top10_2026.csv`,
+including `World 128,003,776 / 118,350,166 / 111,565,144`.
+
+Two things that look like errors and are not, so nobody "fixes" them later: Japan above the
+United Kingdom above India in 2026, and Brazil inside the top ten while Canada sits at
+2,490,256 just outside. Both are the IMF column's own ordering, and the rank column is
+monotone in it — verified below.
+
+## Upgrade — the IMF column is now Corroborated, not merely Documented
+
+Your JSON marks rows 0–3 `verified_against_live_page: true` and **rows 4–10 `false`**. I
+took the seven `false` rows to the primary source rather than to another transcription:
+`https://www.imf.org/external/datamapper/api/v1/NGDPD/...`, values in billions of US$,
+multiplied by 1,000 to your millions.
+
+| rank | country | your `imf_2026_mn` | IMF DataMapper 2026 (bn → mn) | verdict |
+|---|---|---|---|---|
+| 0 | World (`WEOWORLD`) | 128,003,776 | 128,003.776 → 128,003,776 | exact |
+| 1 | United States | 32,642,373 | 32,642.373 | exact |
+| 2 | China | 21,546,403 | 21,546.403 | exact |
+| 3 | Germany | 5,471,299 | 5,471.299 | exact |
+| 4 | Japan | 4,341,986 | 4,341.986 | exact |
+| 5 | United Kingdom | 4,268,446 | 4,268.446 | exact |
+| 6 | India | 4,060,207 | 4,060.207 | exact |
+| 7 | France | 3,554,155 | 3,554.155 | exact |
+| 8 | Russia | 2,861,156 | 2,861.156 | exact |
+| 9 | Italy | 2,701,737 | 2,701.737 | exact |
+| 10 | Brazil | 2,654,761 | 2,654.761 | exact |
+
+**11 of 11, to the unit. Grade: Corroborated** for the IMF column — two independent
+sources, one of them the IMF's own series, agreeing exactly. Your conservative `false`
+flags on ranks 4–10 are superseded by this measurement, not deleted; the honest status now
+is `verified against primary IMF`, and the flags should say that rather than `false`.
+
+The IMF API also confirms the ordering claim: World 128,003.776 > USA > CHN > DEU > JPN
+(4,341.986) > GBR (4,268.446) > IND (4,060.207) > FRA > RUS > ITA (2,701.737) > BRA
+(2,654.761) > CAN (2,490.256). Monotone, so rank = position in the IMF column.
+
+## Unrecovered — the World Bank primary, stated plainly
+
+`api.worldbank.org/v2/country/.../indicator/NY.GDP.MKTP.CD?date=2025` **timed out twice**
+from this host (25 s, 0 bytes received, curl 28). So the `wb_2025_mn` and `un_2024_mn`
+columns remain **Documented-via-Wikipedia only** — a transcription of a transcription. I am
+not calling them verified because their IMF sibling is.
+
+## One defect in the artifact pair, carried forward from reply 14
+
+The CSV has no verification column and the JSON has one, so the two files disagree about
+the table's own status, and the CSV — the one a consumer is most likely to read — cannot
+say which rows were checked at all. That is the same class as reply 14's "the parser grades
+itself": the artifact states a status it does not let a reader verify. Category field or a
+`status` column in both files, or neither.
+
+No Q-item closed by me. Nothing here decides anything Alexander holds.