Corrections

They are not re-priced: the series is append-only. From 2026-10-16 each new row records where its reference came from (btc_usd_source), and an input carried for 10 or more NYSE sessions is flagged on the row (stale_cap_exceeded).

Why the two rows have none: the daily run read the BTC/USD close for the newest constituent close (a Tokyo close already in for the next day), not for the row's data date, so the reference and the row were a day apart and the row recorded null. The run now reads the close for the row's data date, retries once on a rejected read, and records what it found. The BTC/USD reference is informational: the level is computed from constituent prices and share counts only and never reads it, so no level is affected. Stale cap (methodology note, 2026-10-16): an input carried for 10 or more NYSE sessions keeps its last close in the level, as the methodology says, and is listed on the row and in the heartbeat's stale_inputs; the company's membership becomes a question on the agenda of the next scheduled review, never a change outside one. Silvia (observed 2026-09-28) is the first input over the cap. No row, level, divisor, capping factor or constituent moved; the existing rows are byte-identical and the hash chain re-verifies.

The nightly runs of 2026-10-12 and 2026-10-13 committed nothing (the pre-commit test gate was red until 2026-10-14); the 2026-10-14 run recorded LBTC's 2026-10-12 reading late, but the late-session pass covers LBTC only.

No row, level, divisor, threshold, capping factor or constituent moved; the existing rows of both series are byte-identical and both hash chains re-verify. LBTC, the cap-weighted index, has no missing session: 2026-10-10 and 2026-10-11 are a weekend. Every NYSE session between a series' first and last reading that has no reading is now written once beside the rows (missing_sessions, missing_sessions_h: date, reason, recorded_at) and served with the recorder's heartbeat at /api/listed-bitcoin/health; the chart breaks the line at a missing session instead of drawing across it.

Under R6 it is a bitcoin business (Rule A, at least 50 %). Membership does not change outside a scheduled review: it is pending re-admission and re-enters at review LBTC-REV-2026Q4 (last close on the old list 2026-12-31) through the pending-admission mechanism of 2026-10-04.

On 2026-10-11 Core Scientific left the index under methodology 2.0 because its Rule A line was not yet cited (rule_a null) and its Rule B measure was 1.41 % of market capitalisation. The filing was read on 2026-10-13: Consolidated Statements of Operations, revenue lines colocation $65,424K, digital asset self-mining $229,207K, hosted mining from customers $24,388K, total $319,019K; hosting and colocation are not counted, as on the other miners' rows. No threshold, constituent, weight, capping factor, divisor or level moved; the 2.0 entry is unchanged.

The 20 % single-name cap now binds in the level through capping factors set at each review. The index is re-based: 100 on 2026-10-09.

Why: on 2026-10-09 Space Exploration Technologies and Tesla were 94.5 % of the index's market capitalisation and each held 0.1 % of its own market capitalisation in bitcoin; under 1.x the single-name cap was a display view and the level measured two non-bitcoin businesses. The publisher's decision of 2026-10-11 (A-or-B, 30 % admit, 20 % exit). Rule A evidence is each company's own annual filing, cited on its row (materiality.rule_a); Rule B is measured on each row from its holding, the reference BTC-USD close of 2026-10-09 (82,546.32) and its market capitalisation as recorded (materiality.rule_b), so a reader recomputes it. Of the 32 constituents 28 pass: 6 under Rule A alone (Coinbase Global, Hut 8, Riot Platforms, Ionic Digital, Bitdeer, Keel Infrastructure), 15 under Rule B alone, 7 under both measures (Bullish, Galaxy Digital, MARA, CleanSpark, Gemini, BitFuFu, Canaan - admitted on Rule B today, their Rule A revenue line to be cited by the next review). Four do not pass either and leave: Space Exploration Technologies (0.1 %), Tesla (0.1 %), Block (1.6 %), Core Scientific (1.4 %). Their company pages remain in the register with this reason; 1,163,627 of 1,203,813 BTC (96.7 %) stays in the index. Mechanics: nobody cites the 1.x series yet, so the publisher chose a clean base over chaining a three-week-old series with a composition it now rejects. Base date 2026-10-09 = 100 (the last close before the decision); the 1.x series, base 2026-09-22 = 100 under divisor 3686509878964.8403, is closed at its 2026-10-08 reading in the effective view and its 2026-10-09 reading stays in the record view; no 1.x figure is revised. Capping factors: Strategy 0.294608, Coinbase Global 0.369127, every other constituent 1; capped weights on the base date Strategy 20.00 %, Coinbase 20.00 %, Hut 8 11.89 %. The 17 rows admitted pending review under 1.x are re-screened under R6 at the next review (2026-12-31); on 2026-10-09 figures Bitmine Immersion (an Ethereum treasury), GameStop and Figma would not pass either rule. Methodology changes from this version take effect at a pre-announced close; this first one takes effect on publication because no external party cites the series yet.

Holdings do not enter the cap-weighted level, the weights or the divisor, so none of them moved and the published level series is unchanged. The earlier figure was correct for its own date (2026-09-25).

Holdings do not enter the cap-weighted level, the weights or the divisor, so none of them moved and the published level series is unchanged. The earlier Strategy and Strive figures were correct for their own dates (2026-09-27 and 2026-06-30).

None enters the level, the weights or the divisor before the review.

Each passed every acceptance rule on its own SEC quarterly report, read on 2026-10-04: holdings as of the report's balance-sheet date, share count from its cover page, listing from the exchange its cover registers. Under methodology 1.7 membership changes only at the quarterly review, so each has its company page now and enters at that review's close. Holdings recorded: Cipher Digital about 646 BTC, Fold Holdings 271, Exodus Movement 600, GD Culture Group 7,500, Figma about 173, HIVE Digital Technologies 190, Bit Digital 21.5, GameStop 1.0. No published figure was revised.

Each figure now carries the date the filing states it at.

Holdings do not enter the cap-weighted level, the weights or the divisor, so none of them moved. The earlier figures were correct for their own dates (Strategy as of 2026-09-13, MARA as of 2026-03-31) but were published without those dates.

No admission rule, holding, share count, weight or divisor changed. The next review date is derived from the calendar, and the version number from this log.

The 2026-09-29 and 2026-09-30 rows are withdrawn.

Until now a row took the date of the newest close among the constituents. The daily job runs after New York closes and while Tokyo trades, so Metaplanet's next-day close dated the rows of 2026-09-29 and 2026-09-30, each with 1 of 32 constituents priced on that date and 31 carried forward. From methodology 1.6 the date is the most common close date (a tie goes to the earlier day); a close already in for the next day is used as it is and counted as ahead; a date on which fewer than half the constituents closed is refused. Both rows stay in the series unchanged and are withdrawn by void rows that name them by hash. A new 2026-09-29 row, with 26 of 32 constituents priced on that date, was recorded from the closes on file. No price, share count or divisor was revised.

Closes are now refreshed daily and each day's level is recorded in the research series.

No acceptance rule, holding, share count, price or divisor was revised by this entry, so no weight or level moved. The daily research series began with the 2026-09-25 admission-state reading (level 100.00) and the first refreshed reading of 2026-09-28.

Every existing constituent was classified under the new wording; none changed status. Metaplanet's listing source was the Japan Exchange Group home page, which names no company; it is now Metaplanet's own interim results disclosure of 2026-08-13, whose cover states its Tokyo Stock Exchange listing and securities code 3350. No holding, share count or price was revised, so no weight or level moved.

The divisor was adjusted so the level did not move on admission.

Bullish, Canaan, BitFuFu, DeFi Technologies, Boyaa Interactive International, Capital B, The Smarter Web Company, OranjeBTC each passed every acceptance rule on sources read on 2026-09-28; prices are the 2026-09-25 close, and the four non-dollar prices and the Canadian one are converted at that day's closing exchange rate, recorded on each row. Before admission the twenty-four constituents summed to $3,679,046,392,812 and the divisor was 3,679,046,392,812. The eight add $7,463,486,153 at their recorded prices. The divisor was scaled by the same ratio, from 3,679,046,392,812 to 3,686,509,878,965, so the level is identical immediately before and after the admission. Not admitted: DDC Enterprise (its latest filed share count, at 30 June 2026, does not fit the September price series and a convertible preferred class is outstanding), MicroCloud Hologram and NEXON (no current bitcoin quantity in a filing read), Bitcoin Group SE (its annual report states bitcoin only as a euro value), and Semler Scientific (deregistered after its acquisition by Strive, which is already a constituent). No existing figure was revised.

The divisor was adjusted so the level did not move on admission.

Gemini Space Station, Galaxy Digital, Space Exploration Technologies (SpaceX), Next Technology Holding, Silvia (formerly ProCap Financial), Empery Digital, Ionic Digital, Keel Infrastructure, KULR Technology Group and Bitdeer Technologies Group each passed every acceptance rule on sources read on 2026-09-27; prices are the 2026-09-25 close. Before admission the fourteen constituents summed to $1,703,628,920,341 and the divisor was 1,703,628,920,341. The ten add $1,975,417,472,472 at their recorded prices. The divisor was scaled by the same ratio, from 1,703,628,920,341 to 3,679,046,392,812, so the level is identical immediately before and after the admission. American Bitcoin is not admitted: it is majority-owned by Hut 8, whose recorded holding already includes American Bitcoin's bitcoin. No existing figure was revised.

The divisor was adjusted so the level did not move on the revision.

Strategy: 349.8M (StockAnalysis, read 2026-09-18) to 384,225,751 class A and class B shares (Form 10-Q cover page, as of July 24, 2026). MARA Holdings: 381.3M (StockAnalysis) to 386,299,297 (Form 10-Q cover page, as of July 30, 2026). Metaplanet: 1.28B (StockAnalysis) to 1,281,256,899 shares excluding treasury shares (interim financial results filed with the Tokyo Stock Exchange, as of June 30, 2026). No price was changed: the three recorded prices were confirmed against exchange-session closes for 2026-09-18. The ten constituents summed to $94,004,898,035 before the revision and $99,371,855,051 after; the divisor was scaled by the same ratio, from 94,004,898,035 to 99,371,855,051.

The divisor was adjusted so the level did not move on admission.

The four had been fully sourced and held since 2026-09-24 because, weighted by market capitalisation alone, they would have been 94% of the index and Tesla 88%. Methodology 1.2 adds a 20% single-name cap, and each of the four passed every acceptance rule; each source was read again on 2026-09-25. Before admission the ten constituents summed to $99,371,855,051 and the divisor was 99,371,855,051. The four add $1,604,257,065,290 at their recorded prices. The divisor was scaled by the same ratio, from 99,371,855,051 to 1,703,628,920,341, so the level is identical immediately before and after the admission. The cap changes weights, not the level. No existing figure was revised.

The divisor was adjusted so the level did not move on admission.

Before admission the three constituents summed to $60,880,284,000 at their recorded prices and the divisor was 60,880,284,000. The seven new constituents add $33,124,614,035 at their recorded prices. The divisor was scaled by the same ratio, from 60,880,284,000 to 94,004,898,035, so the level is identical immediately before and after the admission. Adding a company changes what the index holds; it does not move the level. No existing figure was revised.

The index is set to 100 on 2026-09-22 over the three candidates that satisfy every acceptance rule.

No figure was revised. This entry exists so that the log begins at the same moment the series does, rather than at the first mistake.