[daggerok]

What

• ETF-H6, WisdomTree part (review WT-5): wisdomtree.com prints "Since Inception" in the Average Annual row as the CUMULATIVE figure for funds younger than one year (NTSD, launched 2026-03-19: 20.18 as of 2026-09-30), and it was published as siAnn and returns.monthEnd.sinceInception
  • new withoutYoungSinceInception nulls it below 365 days between the catalog inception and the table date, on the Market Price, NAV and index rows of the month-end and quarter-end tables (unknown inception keeps the figure)
  • mergeOfficialReturns no longer fills that gap from the Yahoo estimate: it is dated weeks later (WQTM got -9.59 to 2026-10-23 under the official 2026-09-30 date in the first scoped run)
• Derived siAnn was measured from the first close of whatever was loaded (a HISTORY_RANGE=5y window on an old fund gave siAnn = cagr5y). It now needs a series that starts within 10 days of Yahoo's first trade date (catalog inception as fallback) and at least one year, same rule as Invesco fix(updater): kept data reads its own meta.json, siAnn never annualized below one year or from a window Invesco#81
• WT-7: the derived 1-month window used Date.UTC(y, m - 1, d), so Oct 31 started at "Sep 31" = Oct 1; it is clamped to the end of the shorter month now (only used when the official 1-month figure is missing)
• README: the siAnn line documents both rules
• Not here, noted for the shared work: WT-2 (N-PORT tickers filled by name on bonds and FX forwards) belongs to ETF-C3 ticker hygiene; WT-3 (000000000 CUSIP drops the ISIN) and WT-8 (missing pctVal written as "0") belong to the shared N-PORT reader ETF-H2

Published values that change

• Next refresh of each fund, no data committed here. Funds younger than one year at the 2026-09-30 table lose a cumulative siAnn: GDT -14.29, NTSD 20.18, WAMA 7.76, WDIG -22.65, WDRN -13.68, WIMA 8.12, WQTM 7.59, WSPC -9.78, WTLS 20.12 -> null
• Scoped real run TICKERS=NTSD,WQTM,WDAF,DGRW REQUEST_SLEEP=0: NTSD 20.18 -> null, WQTM 7.59 -> null (NAV 7.49 and index 7.58 rows too), WDAF -0.75 (383 days) and DGRW 12.94 unchanged; exit 0, 2 updated, 2 unchanged, 0 failed; the only data changes were the since-inception fields and generatedAt; data discarded (git checkout -- api/)

Checks

• bun test: 57 pass, 0 fail, exit 0; the same run with the config controls exported also exits 0
• New tests in metrics 2 fail on origin/main (2 fail): a 5-year series from inception gives 14.87 while the same series without its first 3 years or without an inception gives null, Oct 31 -> Sep 30 gives a 1-month 10 (was -45), USFR's table keeps 1.99 for a 2014 or exactly-one-year fund and nulls all three rows for NTSD's 2026-03-19 inception, and the merge does not refill it from Yahoo
• bun build --target=bun scripts/update-data.ts --outfile=/dev/null and src/main.tsx: ok
• .claude/tools/etfs/tc/check.sh: 64 lines before and after, no new errors
• check-readme.ts: ok; git diff --check: clean