2020-21 NBA Hoops Premium Box Set

• G-001 EXTRACTION_SPEC 5.2 BASE-CARD IDENTITY COLLISION, naming both releases: hoopspremium-bkb-2020-21 (this release) and paninihoops-bkb-2020-21 describe THE SAME 270 physical cards. This release is the /199 Premium Box Set printing of that checklist; the parent is the pack-issued unnumbered printing (paninihoops-bkb-2020-21 carries base_print_run=unnumbered). MEASURED, NOT ESTIMATED, against data/paninihoops-bkb-2020-21/base_cards.csv: 269 of 270 rows identical on card_number, player_name and team_or_school; 0 rows share a card number and differ on player or team; the multiset of 270 (player_name, team_or_school) pairs is IDENTICAL on both sides. The one number-level difference is this release's "3b" (Jeremy Lamb, Indiana Pacers) against that release's "98" (Jeremy Lamb, Indiana Pacers) - the same physical card under two catalogued numbers, see G-003. POSITIONAL overlap, i.e. row index against row index, is only 175 of 270, purely because the 3b row sits at index 4 and shifts rows 4-98 by one before the two files realign at card 99; positional comparison understates the checklist identity and by-card-number comparison is the correct measure here. Built per 5.2 rather than deduplicated or double-counted. Note also that the same physical object is recorded THREE more times in paninihoops-bkb-2020-21: as a base parallel row "Premium Box Set" print_run 199, as a sets.csv row "Hoops Tribute Premium Box Set Autographs" (20 cards), and in its release notes; its own open gap already says those rows may belong to a separate release record. This release IS that record, and a resolver pass should decide which side keeps them.
• G-002 parallels.csv is written with its header and ZERO data rows, which is a finding and not an unfinished file (EXTRACTION_SPEC 2.3). EVIDENCE FOR THE ABSENCE, not merely absence of evidence: https://www.tcdb.com/Inserts.cfm/sid/253171/2020-21-Hoops enumerates the parent set's complete inventory - 104 insert sets and 24 parallel sets - and "Premium Box Set" (sid 259425) appears exactly once among the 24 parallels with no sub-parallel of it of any kind (there is no "Premium Box Set Gold", no "Premium Box Set Holo", nothing). The autograph set "Premium Box Set Autographs" (sid 259513) likewise has no parallel entry. cardboardconnection's page for the product (https://www.cardboardconnection.com/2020-21-nba-hoops-premium-box-set-basketball-cards) describes only the /199 base cards and the one /10 Tribute autograph and names no parallel. This matches hoopspremium-bkb-2019-20, which is also parallel-free. It remains true that no source says in words "this box set has no parallels"; the absence is inferred from a complete enumeration rather than from a statement.
• G-003 CARD NUMBER 98 IS UNISSUED IN THIS PRODUCT AND THE CONTIGUITY CHECK FAILS BY DESIGN. tcdb renders the base run as 1, 2, 3, 3b, 4 ... 97, 99, 100 ... 270: the row immediately after card 3 carries the card-number cell "3b" with player Jeremy Lamb, Indiana Pacers and the flag token "UER", and no card numbered 98 exists. Verified on FOUR fetches, three of them independent: the Premium Box Set checklist page 1 (https://www.tcdb.com/Checklist.cfm/sid/259425/2020-21-Hoops-Premium-Box-Set?PageIndex=1), the Premium Box Set gallery page 10 (https://www.tcdb.com/Gallery.cfm/sid/259425/2020-21-Hoops-Premium-Box-Set?PageIndex=10 , last tile caption quoted verbatim as "100 - Domantas Sabonis *SN199* - Indiana Pacers", no 98), the PARENT gallery page 10 (https://www.tcdb.com/Gallery.cfm/sid/253171/2020-21-Hoops?PageIndex=10 , 10 tiles: 90-97, 99, 100) and the PARENT checklist page 1 (https://www.tcdb.com/Checklist.cfm/sid/253171/2020-21-Hoops?PageIndex=1 , which also renders "3b | Jeremy Lamb | UER"). The row arithmetic corroborates it: 270 rows = 269 plain numbers + 3b, matching tcdb's Total Cards "270" exactly, and Checklist page 2 starts at card 101 and gallery page 11 starts at card 101, which only holds if cards 1-100 occupy exactly 100 rows. CONTRADICTS EXISTING DATA: data/paninihoops-bkb-2020-21/base_cards.csv numbers this card 98 and has no 3b row, i.e. the aggregator sources that release used (beckett, checklistcenter, sportscardchecklist) carry the intended checklist slot while tcdb carries the printed number. UNRESOLVED - I did not find a source stating which number is physically printed on the card, and card_number is written verbatim as the source renders it (EXTRACTION_SPEC 2.4), so this release carries "3b" and no 98. If a later pass establishes that the card reads 98, this row and the contiguity gap should both be revisited.
• G-004 TASK-BRIEFING TRAPS 1 AND 2 TESTED AND BOTH RESOLVED IN THIS RELEASE'S FAVOUR, recorded so the next pass need not re-test them. (1) NAME: the briefing said the "Premium Stock Box Set" styling belongs to 2019-20 only and that 2020-21 should read "Premium Box Set". Confirmed: tcdb prints "2020-21 Hoops - Premium Box Set" (sid 259425) and cardboardconnection heads its page "2020-21 NBA Hoops Premium Box Set Basketball Cards Checklist". Note the sibling release hoopspremium-bkb-2019-20 already corrected the briefing on the other half of this: the 2019-20 entry in THIS line is also styled "Premium Box Set", and "Premium Stock Box Set" is a different product. (2) PREMIUM STOCK IS A DIFFERENT LINE AND IS NOT MERGED HERE. tcdb files Hoops Premium Stock under its own parent, 2019-20 sid 245171 (https://www.tcdb.com/Inserts.cfm/sid/245171/2019-20-Hoops-Premium-Stock); the line is ongoing (a beckett news page titled "2023-24 Panini Hoops Premium Stock Basketball Checklist, Team Set Lists and Details" exists). NO 2020-21 Hoops Premium Stock product was found by any search or source consulted, so there is nothing to conflate in this season; that agrees with data/paninihoops-bkb-2020-21, whose gaps already state "NBA Hoops Premium Stock is a 2019-20 SKU". reference/brands.csv still has no row for a Premium Stock line, so if one is ever built it needs a brand decision - that escalation is already open as G-003 on hoopspremium-bkb-2019-20 and is NOT re-decided here. LINE CONTINUITY: the task brief's claim that this line ran through 2024-25 is supported - cardboardconnection has "2021-22 NBA Hoops Premium Box Set Checklist, Set Info, Boxes" and "2023-24 NBA Hoops Premium Box Set Checklist, Set Info, Boxes", and checklistinsider has "2024-25 NBA Hoops Premium Box Set Basketball Checklist, Boxes". I did not open those pages, so this is search-result-title evidence only and no value was written from it.
• G-005 EXTRACTION_SPEC 5.4 UNIT CHECK AT RELEASE LEVEL - THE SAME ERROR THE 2019-20 SIBLING CAUGHT, AND IT RECURS HERE IN A DIFFERENT NUMBER. cardboardconnection calls this product a "271-card factory set" while also stating "Full base set (270 cards)". 271 is CARDS PER SEALED BOX: the 270-card base set plus the one Tribute autograph, exactly the shape the same site describes with "every sealed box of 2020-21 NBA Hoops Premium Box Set Basketball adds one Tribute Autograph". 271 is therefore written into NO field; base_set_size is 270, sourced from tcdb Total Cards "270" and from the "Full base set (270 cards)" phrase, and corroborated by counting 270 rendered rows. For the record of how this trap has behaved across the line: in 2019-20 the confusable figures were 305 and 303 and belonged to a DIFFERENT product (Premium Stock); in 2020-21 the figure is 271 and belongs to THIS product but is still not a base set size.
• G-006 release_date CONTRADICTS EXISTING DATA AND RESTS ON ONE SOURCE PLUS AN INFERRED YEAR. This release writes 2021-03-09 from cardboardconnection, "Available through Panini's website on Tuesday, March 9 at 11 AM (Eastern)" (https://www.cardboardconnection.com/2020-21-nba-hoops-premium-box-set-basketball-cards). The source states month and day but not the year; 2021 is taken from the product season, and 9 March 2021 was a Tuesday, which is consistent but is a derivation rather than a quotation. data/paninihoops-bkb-2020-21/release.csv describes the same object and states "released 2021-03-15" in its notes - six days later. Neither figure was averaged or dropped and I did not attempt to adjudicate them; tcdb prints no release date on sid 259425. UNRESOLVED.
• G-007 CHANNEL IS KNOWN IN WORDS BUT HAS NOWHERE TO LIVE IN THE SCHEMA, so it is recorded here. cardboardconnection states "Available through Panini's website", sold in a Dutch auction where "The price opens at $4,000 and incrementally drops every few minutes until it either hits $750 or sells out". In the taxonomy of EXTRACTION_SPEC 2.3 that is channel=panini-direct with channel_basis=sourced-exclusive and channel_class=promotion. release.csv has no channel column and parallels.csv has no rows, so no machine-readable channel value exists anywhere in this release and a consumer filtering by channel will not find it. Contrast hoopspremium-bkb-2019-20, whose G-005 records that NO source stated the 2019-20 channel at all - for 2020-21 the channel IS sourced and is still unrepresentable.
• G-008 TCDB HEADER DISAGREES WITH ITS OWN RENDERED ROWS ON THE AUTOGRAPH SET, AND THE ROWS WERE BELIEVED. https://www.tcdb.com/ViewSet.cfm/sid/259513/2020-21-Hoops-Premium-Box-Set-Autographs prints Total Cards "12" while rendering 20 rows, numbered 251-270 with no gaps, each flagged "AU, SN10". Counted on two endpoints (ViewSet.cfm and Checklist.cfm) and both render 20. EXTRACTION_SPEC 3.2 says Total Cards counts different things on different lines and that a disagreement is a question, never an answer; card_count=20 is written because 20 rows were counted and because data/paninihoops-bkb-2020-21/sets.csv independently records the same set as 20 cards at 251-270 from beckett. What "12" counts on that page is unexplained - it is not a distinct-number count, not a signer count I could verify, and not a per-tier line count. UNRESOLVED, and a reason to distrust Total Cards on this line generally.
• G-009 rookie_flag RESTS ON THE PARENT SET, NOT ON THIS PRODUCT'S OWN CHECKLIST, and its negative half is partly structural. tcdb renders NO "(RC)" token anywhere on the Premium Box Set's own rows or tiles - the parallel set renders "SN199" instead - which is the same behaviour hoopspremium-bkb-2019-20 recorded in its G-008: a parallel gallery is a good print-run source and a bad rookie-flag source. rookie_flag=yes for 201-250 was therefore read off the PARENT set, https://www.tcdb.com/Checklist.cfm/sid/253171/2020-21-Hoops?PageIndex=3 , which renders "RC" on every one of the 50 rows 201-250 and renders "TRIB" (not RC) on all 20 rows 251-270 - so rookie_flag=no on 251-270 is directly evidenced. For cards 1-200 the evidence is thinner: the parent checklist page 1 was fetched and shows "(none)" on rows 1-20 with the explicit answer that "RC" appears on no row of that page, the Premium Box Set checklist pages 1 and 2 render only "SN199" (and "UER" on 3b) across cards 1-77 and 101-186, and cardboardconnection's 2020-21 Panini NBA Hoops page places "Rookies" at 201-250 and "Hoops Tribute" at 251-270. No row-level token was read for cards 78-100 and 187-200 on any endpoint that renders RC; rookie_flag=no there follows from the stated block structure rather than from a rendered flag column. data/paninihoops-bkb-2020-21/base_cards.csv independently records rookie_flag=yes for exactly 201-250 and no elsewhere, so the two releases AGREE - unlike the 2019-20 pair, where hoopspremium-bkb-2019-20 G-007 records a direct rookie_flag contradiction with paninihoops-bkb-2019-20.
• G-010 card_tier DIVERGES FROM BOTH NEIGHBOURS AND THE DIVERGENCE IS RECORDED RATHER THAN SMOOTHED. This release writes card_tier="Hoops Tribute" on cards 251-270 and leaves it empty on 1-250. Basis: the tcdb PARENT checklist renders the literal flag token "TRIB" on exactly those 20 rows and "RC" on 201-250 (https://www.tcdb.com/Checklist.cfm/sid/253171/2020-21-Hoops?PageIndex=3), and cardboardconnection's 2020-21 Panini NBA Hoops page names a "Hoops Tribute" subset at 251-270 (https://www.cardboardconnection.com/2020-21-panini-nba-hoops-basketball-cards). NOTE the expansion of the token "TRIB" to the words "Hoops Tribute" comes from cardboardconnection, not from tcdb - EXTRACTION_SPEC 3.3 warns that the fetch layer invents glosses for tcdb flag tokens, so the token is recorded here as the literal string it is. TWO DIVERGENCES: (a) data/paninihoops-bkb-2020-21/base_cards.csv leaves card_tier EMPTY on all 270 rows including 251-270, although its release notes describe the same three blocks in prose; (b) the 201-250 rookie block carries a named subset label ("Rookies") on cardboardconnection but gets no card_tier here, following hoopspremium-bkb-2019-20 G-009, which left its rookie block's card_tier empty for the same reason. The base run is NOT treated as tiered for print-run purposes: the run is uniformly /199, so base_print_run is the integer 199 and not varies-by-tier, and base_card_tiers.csv is correctly absent (EXTRACTION_SPEC 2.5).
• G-011 THE TWO STRUCTURAL SOURCES FOR THIS RELEASE MAY BE CLOSER TO ONE OPINION THAN TWO, and there is no third. tcdb and cardboardconnection agree on every value written here (270 base cards, /199, a 20-card Tribute autograph at 251-270 numbered /10) and cardboardconnection's set inventories for the Hoops line track tcdb closely enough that independence cannot be assumed - the same caution hoopspremium-bkb-2019-20 raised in its G-010. beckett was NOT reachable in this pass and no beckett page was opened. Independent corroboration that does exist is marketplace only and was seen as a search-result product title rather than a fetched page: "2020-21 Panini NBA Hoops - LaMelo Ball #223 Premium Box Set /199 (RC)" (https://www.ebay.com/p/7056500331), which matches this checklist at card 223 (LaMelo Ball, Charlotte Hornets), matches the /199 print run, and matches rookie_flag=yes at 223. No value in these CSVs rests on that title alone. The independent-ish cross-check that DID carry weight is data/paninihoops-bkb-2020-21, built from beckett, checklistcenter and sportscardchecklist, which agrees with this release on all 270 players and teams and on the rookie block.
• G-012 SPEC DEFECT MEASURED IN THIS PASS, affecting EXTRACTION_SPEC 3.2. The claim that "Checklist.cfm?PageIndex=N PAGINATES AT 100 ROWS PER PAGE AND BEATS Gallery.cfm 10:1" is right about the SERVER and wrong about the FETCH LAYER. Server-side pagination really is 100 rows per page - page 2 of sid 259425 begins at card 101 and page 3 at card 201 - but the summarizing fetch delivers only the first ~77-86 rows of each page body and the remainder is never delivered, so re-asking for the tail with a narrow prompt returns "content cuts off" rather than the rows. Measured here: page 1 stopped at card 77 of 100 on two separate prompts, page 2 stopped at card 186 of 200, page 3 (70 rows, under the truncation point) came back complete. The working recipe is Checklist for the bulk plus Gallery.cfm?PageIndex=N top-ups for the last quarter of each 100-row page: this 270-card run cost 3 Checklist fetches plus 6 Gallery fetches, not 3 and not 27. A SECOND DEFECT, also measured: a Gallery.cfm fetch SILENTLY DROPPED A TILE - ?PageIndex=10 returned "9 card tiles" and omitted card 100 on two consecutive calls, and the second call was a 15-minute cache hit on the identical URL rather than an independent check; appending a harmless query parameter (&sort=0) busted the cache and returned 10 tiles including "100 - Domantas Sabonis *SN199* - Indiana Pacers". A row count that agrees with itself across two fetches of the SAME URL is one observation, not two. THIRD, for the toolkit: direct HTTPS to www.tcdb.com via curl is refused by this environment's egress proxy (403 to CONNECT), so WebFetch and its truncation are the only route and raw HTML parsing is not available.
• G-013 No pack odds and no scarcity tier are recorded for either set. This is a sealed complete factory box set with fixed contents sold as a single unit by Dutch auction, not a pack product, so per-set pack odds do not exist in the ordinary sense; no source consulted states any odds figure. stated_odds, odds_evidence_url, odds_evidence_quote and scarcity_tier are therefore empty on both sets.csv rows. Same disposition as hoopspremium-bkb-2019-20 G-012.
• G-014 license_status WAS NOT SUPPLIED BY THE TASK BRIEFING and the catalogue is already inconsistent on it, so the value written here is a convention choice and is recorded as one. EXTRACTION_SPEC 2.1 says license_status comes "as given in your task prompt"; this prompt gave product_name, season, manufacturer and league only. This release writes "licensed (NBA/NBPA)", matching its own line's prior season hoopspremium-bkb-2019-20 byte for byte. The parent product for the SAME season, data/paninihoops-bkb-2020-21/release.csv, writes "licensed (NBA)" for what is physically the same checklist under the same licence. Both values are live per the spec and nothing in either row tells a reader why they differ. No source was consulted that distinguishes them; I did not adjudicate, and the columns that actually gate the loader (licensed=yes, league=NBA) are unambiguous and identical on both sides.

If you know a source that answers one of these, tell us — name the question's label and where to look.