2017-18 NBA Hoops Premium Box Set

• G-001 EXTRACTION_SPEC 5.2 BASE-CARD IDENTITY COLLISION, NAMING BOTH RELEASES: hoopspremium-bkb-2017-18 (this release) and paninihoops-bkb-2017-18 describe the same 300 physical card faces under two release ids - this one the #/199 holofoil factory-set printing, that one the pack-issued unnumbered printing. Built in full per 5.2 rather than deduplicated, referenced or double-counted: the 300 rows here were enumerated from this product's own sid (160898) and were never copied from the other directory. THE OVERLAP, MEASURED WITH csv.DictReader RATHER THAN ASSERTED: by card number 300 of 300, and positionally (row index against row index) 300 of 300, on card_number, player_name and team_or_school; zero card numbers appear on one side only; and rookie_flag, card_type and card_tier also match 300 of 300. Both figures are reported because they can diverge badly - the 2020-21 pair of this line reads 175 of 270 positionally against 269 of 270 by number, because one row numbered 3b shifts everything after it. THE COPY CHECK OF EXTRACTION_SPEC 4.2 THEREFORE FIRED AT MAXIMUM AND WAS INVESTIGATED RATHER THAN DELETED: a complete overlap is exactly what a full base parallel looks like as well as what a repackaged base set looks like, so it does not settle product versus parallel - what settles that is the owner decision in specs/2017-18_SCOPE_MEMO.md section 8a and the release's own chase set. A resolver pass should decide which side of the pair keeps the shared objects. Sources: https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=1 ; https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=2 ; https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=3 ; https://www.tcdb.com/Gallery.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=8 through 30 ; data/paninihoops-bkb-2017-18/base_cards.csv
• G-002 parallels.csv IS INTENTIONALLY EMPTY - HEADER ONLY, ZERO DATA ROWS - AND THE ABSENCE IS EVIDENCED FROM POSITIVE ENUMERATIONS RATHER THAN FROM SILENCE. (1) tcdb's inventory of the parent product was read in full (https://www.tcdb.com/Inserts.cfm/sid/153685/2017-18-Hoops): its "Parallel Sets (16)" section contains exactly one row whose name contains "Premium" - "Premium", sid 160898, which IS this release's base set - and no tier scoped to it (no Premium Silver, Premium Gold or the like); the sixteen parallel rows are Artist Proof, Artist Proof Black, Blue, Blue Checkerboard, Gold, Green, Orange, Orange Explosion, Platinum, Premium, Red, Red Back, Red Checkerboard, Silver, Teal and Teal Explosion, all of which are parallels OF THE PACK PRODUCT and are carried by data/paninihoops-bkb-2017-18/parallels.csv. (2) sportscardchecklist's 2017-18 Panini Hoops index (https://www.sportscardchecklist.com/sport-basketball/year-2017-18/index-panini-hoops/trading-card-checklists-and-product-information) lists exactly three entries containing "Premium": the Premium Box Set itself (290 Cards), the Kobe Bryant Career Tribute Premium Box Set (10 Cards) - the same object's Kobe block under that site's unit convention - and the Kobe Bryant Career Tribute Autographs Premium Box Set (10 Cards), which is this release's chase set and is a sets.csv row here. (3) cardboardconnection (https://www.cardboardconnection.com/2017-18-panini-nba-hoops-basketball) prints "Premium Box Set #/199" as one line of the base-set parallel list and gives no sub-tier of it. So no source reachable in this pass describes a parallel OF this product, on three lineages. This is a zero-row file plus this row, which is a finding; a missing file would have been a coverage hole.
• G-003 RELEASE DATE IS UNRESOLVED AND IS WRITTEN AS UNKNOWN RATHER THAN GUESSED OR INHERITED. Precisely what was tried and what each source said. (1) cardboardconnection has NO product page for this product in this season - it is the only season of this line without one - and its 2017-18 Panini NBA Hoops page (https://www.cardboardconnection.com/2017-18-panini-nba-hoops-basketball) prints "Release Date: 10/25/2017" for the PACK product, mentions no separate Premium page and no separate Premium date, asked directly for both. That 10/25/2017 is the figure data/paninihoops-bkb-2017-18/release.csv carries and it is NOT evidence for this product. (2) sportscardchecklist's index for the product states no date anywhere, asked directly (https://www.sportscardchecklist.com/sport-basketball/year-2017-18/index-panini-hoops/trading-card-checklists-and-product-information). (3) tcdb renders no date field at all: the set page's metadata lines are "Total Cards: 300", "Rating: 0.0 (0 ratings)" and "Notes: Factory set; SN199; holofoil cardstock", with no Parent Set line and no release line (https://www.tcdb.com/ViewSet.cfm/sid/160898/2017-18-Hoops-Premium). (4) A Beckett article for exactly this product exists by TITLE in search results - "2017-18 Hoops Premium Box Set Checklist, Info, Kobe Bryant Autographs" at https://www.beckett.com/news/2017-18-hoops-premium-factory-set/ - and the host 302s: the fetch returned "Status: 302 Found" with a Location header of https://maintenance.beckett.com/ , the same failure this project has recorded on every pass. A search-result title is not the page, so nothing was read from it. (5) Two web searches for a Panini America launch post found none; the 2018-19 season of this line rests its date, price and channel on exactly such a post, and specs/2017-18_SCOPE_MEMO.md section 5b already records that that first-party evidence has no 2017-18 counterpart. WHAT WOULD RESOLVE IT: the Beckett article when that host returns, or a Panini blog archive capture.
• G-004 STRUCTURAL DIVERGENCE BETWEEN THIS CATALOGUE AND tcdb, STATED RATHER THAN RESOLVED SILENTLY, AND VERIFIED ON BOTH ENDPOINTS THIS PASS. tcdb gives this product NO parent sid of its own: the season index (https://www.tcdb.com/ViewAll.cfm/sp/Basketball/year/2017) lists exactly two entries whose titles contain "Hoops" - "2017-18 Hoops" (sid 153685) and "2017-18 Hoops NBA 2K18 Legend Edition Shaquille O'Neal" (sid 175517) - and nothing titled as a Premium or Premium Box Set product, asked explicitly. The object is filed instead as subset sid 160898 inside the 2017-18 Hoops parent's "Parallel Sets (16)" section (https://www.tcdb.com/Inserts.cfm/sid/153685/2017-18-Hoops), and its chase set as sid 160903 inside the same parent's "Insert Sets (45)" section. So the most structural source in this pass files this product as a PARALLEL. It is catalogued here as a RELEASE on the owner decision recorded in specs/2017-18_SCOPE_MEMO.md section 8a dated 2026-09-27, whose named discriminator - a chase set of its own - was re-verified here. Two consequences worth recording for a later pass: a season scope bounded from ViewAll.cfm under-counts by at least this release and the Donruss Optic Premium Box Set, as section 5a of that memo already notes; and tcdb files the 2018-19 and later seasons of this line the same way, so this is the line's normal shape rather than a 2017-18 anomaly.
• G-005 THE PARENT CHECKLIST CARRIES A 301ST ROW THAT THIS PRODUCT DOES NOT, AND THAT DIFFERENCE IS A FINDING RATHER THAN AN ERROR ON EITHER SIDE. Both halves were measured in this pass. The parent renders 301 rows over 300 distinct numbers: https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=31 renders exactly one tile, card number "1", subject "Champions", Golden State Warriors, and its page 30 renders 291-300 and reports that a link to page 31 exists. This product renders 300: https://www.tcdb.com/Gallery.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=30 renders 291-300 and reports no page 31 exists, and https://www.tcdb.com/Gallery.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=31&sort=0 renders no tiles with the pagination topping out at page 30. Card index and card number therefore agree throughout this set, confirmed at four page boundaries (page 8 = 71-80, page 18 = 171-180, page 29 = 281-290, page 30 = 291-300) and at both 100-row Checklist.cfm boundaries (page 2 begins at 101, page 3 at 201). data/paninihoops-bkb-2017-18 files the Champions card as a one-card sets.csv row because cards.card_number is UNIQUE per release and card 1 is already Joel Embiid; no equivalent row is written here, because this product does not render it. Consequence for a consumer: the two releases' base runs are identical, and the pack product carries one extra object that the factory set does not.
• G-006 ROOKIE FLAG RESTS ON THE PARENT'S RENDERINGS, NOT ON THIS PRODUCT'S OWN, AND A POSITIVE CONTROL WAS RUN BEFORE TREATING THIS PRODUCT'S SILENCE AS UNINFORMATIVE. tcdb renders no RC designation anywhere on this product: https://www.tcdb.com/Gallery.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=29 renders "SN199" on tiles 281-290 and no "RC" on any of them, while https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=29 renders "RC" on all ten of the same card numbers. The renderer therefore suppresses RC on this sid while still rendering the serial flag there - the same behaviour hoopspremium-bkb-2018-19 and -2019-20 recorded - so the absence of RC here carries no information. rookie_flag was read off the PARENT for the identical card numbers with full 1-300 coverage and no range assumed: https://www.tcdb.com/Checklist.cfm/sid/153685/2017-18-Hoops?PageIndex=1 rendered rows 1-80 with "RC" on 56 only; https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=9 (81-90) and https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=10 (91-100) rendered no RC; https://www.tcdb.com/Checklist.cfm/sid/153685/2017-18-Hoops?PageIndex=2 rendered 101-184 with "RC" on 113 and 153 only; https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=19 (181-190) and https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=20 (191-200) rendered none; https://www.tcdb.com/Checklist.cfm/sid/153685/2017-18-Hoops?PageIndex=3 rendered 201-278 with "RC" on the contiguous run 250-278; https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=28 rendered "RC" on all ten of 271-280, giving 279 and 280; https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=29 rendered "RC" on all ten of 281-290; and https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=30 rendered no flag token at all on the ten Kobe cards 291-300. The union is 44 cards: 56, 113, 153 and 250-290. That is a within-page positive discriminative test in both directions (RC present on some rows and absent on their neighbours on the same page), which is why no is written rather than UNKNOWN. THE WEAK RANGES, STATED: cards 84-100 and 185-200 were reached by Gallery.cfm ONLY, once each; no Checklist.cfm fetch reached them, and data/paninihoops-bkb-2017-18 gap G-004 records that this same Gallery page 9 dropped a VAR token between two readings, so that endpoint's flag column is known to be lossy on this parent. The 44-card result is identical to what data/paninihoops-bkb-2017-18/base_cards.csv carries, which is corroboration between two independent readings of the same pages rather than a second source. No second source publishes per-card RC designations for this checklist.
• G-007 A SET PRINT RUN THAT sets.csv HAS NO COLUMN FOR, RECORDED HERE IN FULL (EXTRACTION_SPEC 2.2). Kobe Bryant Career Tribute Autographs Premium Box is uniformly serial numbered to 20: tcdb renders the literal tokens "AU, SN20" on all ten of its rows, counted one by one (https://www.tcdb.com/ViewSet.cfm/sid/160903/2017-18-Hoops-Kobe-Bryant-Career-Tribute-Autographs-Premium-Box), and cardboardconnection prints "AUTOGRAPH PARALLELS: Autographs Premium Box #/20." (https://www.cardboardconnection.com/2017-18-panini-nba-hoops-basketball). The figure is NOT written in card_count, which means how many cards and is 10 here; putting 20 there would be indistinguishable from a count and is the trap EXTRACTION_SPEC 5.4 documents (2021-22 Hoops Tribute Premium Box Set Autographs was recorded as card_count 10 because someone read a print run as a count). No parallels.csv row was written for it either: the set is not a parallel tier of this release's base set, and it is already a sets.csv row, which EXTRACTION_SPEC 2.3 says is the more accurate resolution when the thing has a set row. The distinct-token set across all ten rows is the single token pair "AU, SN20", enumerated rather than sampled, so this set does not have the per-card variable run that EXTRACTION_SPEC 3.2 warns about.
• G-008 A SET ROW THIS RELEASE AND data/paninihoops-bkb-2017-18 BOTH CARRY UNDER THE SAME NAME, RECORDED SO THE CARDS ARE NOT COUNTED TWICE. data/paninihoops-bkb-2017-18/sets.csv already carries "Kobe Bryant Career Tribute Autographs Premium Box", set_type auto, card_count 10 - the same ten signed cards - because that build enumerated the parent's whole insert inventory before the classification question was decided. It is written here as well, because the owner decision of 2026-09-27 makes this set the discriminator that defines this release and instructs that the cards be counted once, on the release. THAT OTHER DIRECTORY WAS NOT EDITED by this build: one agent, one release, and rows are never deleted here. A resolver pass should decide which side keeps the row. The same shape exists one season later: data/paninihoops-bkb-2018-19 and data/hoopspremium-bkb-2018-19 both carry that season's equivalent autograph set, and that release's gap G-001 says the same thing.
• G-009 VARIATION SETS: ABSENCE IS NOT ESTABLISHED AND IS NOT CLAIMED, AND ONE OPEN QUESTION ON THE PARENT WAS PROBED ON THIS SID. https://www.tcdb.com/Errors.cfm/sid/160898/2017-18-Hoops-Premium returns "No errors or variations have been added for this set yet." That silence is worthless (reference/caveats.csv CAV-020: the endpoint returns that identical sentence on sets that provably carry UER rows), so no conclusion is drawn from it. What the clean base run actually rests on is the contiguous 1-300 enumeration described in release.csv, plus two facts about the suffix question: Gallery.cfm strips a letter suffix, so its tiles cannot answer it, and Checklist.cfm can - https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=1&sort=0 was fetched under a second sort precisely to reach further down the page, rendered rows 1-83, and carries NO card number with a letter suffix anywhere in those 83 rows. That also speaks to the one variation question data/paninihoops-bkb-2017-18 left open in its gap G-004, a "VAR" token seen once on the parent's card 83 (George Hill) and absent on a second reading: on THIS sid card 83 renders as "83 George Hill SN199" with no VAR and no suffix, on the Checklist endpoint, and https://www.tcdb.com/Gallery.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=9 rendered the same tile with SN199 only. A token not seen is evidence of nothing, so this neither confirms nor refutes the parent's positive; it is recorded so the next pass knows this endpoint was reached. WHAT THE PAGE ARITHMETIC DOES AND DOES NOT EXCLUDE: Gallery.cfm paginates by card INDEX, so an extra row anywhere would shift every later page's index-to-number mapping; page 8 renders indices 71-80 as numbers 71-80, page 10 indices 91-100 as 91-100, pages 18 to 20 indices 171-200 as 171-200, pages 29 and 30 indices 281-300 as 281-300, and thirty pages of ten end exactly on card 300. An extra suffixed ROW is therefore excluded throughout the set by arithmetic. What is NOT excluded is a suffix ON one of the rows in 84-100 or 185-200, which Checklist.cfm never reached on this sid and which Gallery.cfm would strip.
• G-010 TRUNCATION CENSUS FROM THIS BUILD, CONTRIBUTED TO THE SEASON'S RUNNING MEASUREMENT RATHER THAN PLANNED AGAINST. Checklist.cfm pages are 100 server-side rows and the fetch layer delivered: 80 of 100 on https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=1 (row 80 cut mid-row), 83 of 100 on the same page under https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=1&sort=0 - a differently-sorted second fetch that reached three rows further - 80 of 100 on https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=2 (rows 101-180), and 86 of 100 on https://www.tcdb.com/Checklist.cfm/sid/160898/2017-18-Hoops-Premium?PageIndex=3 (rows 201-286, row 286 cut mid-row). On the parent sid in the same pass: 80 of 100, 84 of 100 and 78 of 100. Seven measurements, 78 to 86, all above specs/2017-18_SCOPE_MEMO.md section 7.2's revised floor of 40 and all below 100, so top-ups were needed on every page and the 60 per cent budget was not exhausted. Gallery.cfm delivered ten tiles on every one of the nine pages fetched on this sid and five on the parent, with no dropped tile observed; each page was read once except page 31, which was read with an added &sort=0 to bust the per-URL error cache. Total endpoint fetches for this release, counted one by one rather than estimated: 15 on sid 160898 (ViewSet, Checklist pages 1, 2, 3, Checklist page 1 again under sort=0, Errors, and Gallery pages 8, 9, 10, 18, 19, 20, 29, 30, 31), 1 on sid 160903, 12 on the parent sid 153685 (Inserts, Checklist pages 1, 2, 3, and Gallery pages 9, 10, 19, 20, 28, 29, 30, 31), 1 on the season index, 4 off-tcdb page fetches and 4 web searches.
• G-011 MARKETPLACE CORROBORATION WAS ATTEMPTED AND NOTHING WAS WRITTEN FROM IT. (1) A comc card page for exactly this product's autograph set surfaced in search with the title "2017-18 Panini NBA Hoops - [Base] - Premium Box Set Autographs #296 - Kobe Bryant Career Tribute - Kobe Bryant /20 [PSA 8 NM-MT]" and a path segment of 296, which would corroborate both the numbering and the /20 run - but the page itself returns HTTP 403 through this environment's proxy (https://www.comc.com/Cards/Basketball/2017-18/Panini_NBA_Hoops_-_Base_-_Kobe_Bryant_Career_Tribute_Premium_Box_Set_Autographs_Autographed/296/Kobe_Bryant_Career_Tribute_-_Kobe_Bryant/12699952/Graded/PSA/8), and a search-result title is not the listing (CHASE_CHECKLIST_SPEC: a result snippet rendered one card as NWJ-CBS where the page read NWJ-CSX). Nothing was written from it. (2) Two eBay-scoped searches returned no free-text seller title naming this product; the titles that came back were blank or eBay catalogue boilerplate. (3) data/paninihoops-bkb-2017-18 gap G-018 records the one title that was found for this treatment in the previous pass and rejected it correctly: "2017-18 Panini NBA Hoops - Kobe Bryant Career Tribute Kobe Bryant #293 Premium Box Set /199" is in eBay's own catalogue form and a clean contamination fingerprint does not clear that shape. So the #/199 base print run and the #/20 autograph run rest on tcdb row tokens plus editorial statements, with no card-in-hand corroboration - which is weaker than the 2017-18 marketplace corroboration the scope memo reports for National Treasures, and is stated rather than implied.
• G-012 SOURCE INDEPENDENCE, STATED PER CLAIM RATHER THAN ASSUMED. Three lineages were reached: tcdb, cardboardconnection and sportscardchecklist. reference/caveats.csv CAV-037 as corrected on 2026-09-25 says to read cardboardconnection and checklistcenter as ONE lineage by default and to look for a checkable divergence per product; no checklistcenter page for 2017-18 Panini NBA Hoops was surfaced by a site-scoped search in this pass, and per CAV-037 no URL was constructed and no 404 was treated as an absence claim, so that site is recorded as not-found rather than as absent - the same result data/paninihoops-bkb-2017-18 gap G-017 reached. WHAT RESTS ON tcdb ALONE: the 300-row checklist itself (player names, teams and card numbers), the per-row SN199 tokens, the per-card RC designations behind rookie_flag, and the existence of sid 160903 as a set page. WHAT IS CORROBORATED ACROSS LINEAGES: base_set_size 300 (tcdb's header and enumeration; cardboardconnection's "300 cards, including 10 Kobe Bryant Tribute cards"; sportscardchecklist's 290 + 10); the #/199 base print run (tcdb's Notes cell and row tokens; cardboardconnection's "Premium Box Set #/199"); and the autograph set's existence, size and Premium-Box belonging (all three). The Kobe block's boundaries as 291-300 are named in words by no source - they rest on the ten rendered rows plus cardboardconnection's count of ten, which is the same limit data/paninihoops-bkb-2017-18 gap G-020 records.
• G-013 THIS RELEASE AND data/paninihoops-bkb-2017-18/parallels.csv NOW DESCRIBE THE SAME OBJECT IN TWO PLACES, BY DESIGN, AND A CONSUMER MUST BE ABLE TO SEE IT. That file carries a base-set parallel row named "Premium", print_run 199, channel factory-set, whose visual_notes now end with a dated RECLASSIFIED 2026-09-27 marker recording the owner decision and pointing at this release; rows are never deleted here, so a resolver matching the name tcdb prints under its Parallel Sets heading still finds it. The cards are counted once, on this release: base_cards.csv here holds the 300 rows and this release's base_print_run is the 199. NOTE THE APPARENT CONTRADICTION THAT IS NOT ONE: data/paninihoops-bkb-2017-18/release.csv carries base_print_run unnumbered with basis assumed, while this release carries 199 with basis sourced. Both are correct about different objects - the pack printing carries no serial and the factory-set printing is numbered to 199 - and the two rows are about the same card faces, which is why it reads as a contradiction at a glance. That build's gap G-002 anticipated this release by name and said this row should be revisited if it were ever created; it was not edited by this build.
• G-014 WHAT specs/2017-18_SCOPE_MEMO.md SECTION 5b SAID AND WHAT THIS BUILD MEASURED, ITEM BY ITEM, SINCE THAT SECTION IS THE EVIDENCE BASE FOR THIS RELEASE'S EXISTENCE. CONFIRMED: subset 160898 renders "2017-18 Hoops - Premium", Total Cards 300, SN199 on every row, Notes "Factory set; SN199; holofoil cardstock"; cardboardconnection gives it no product page and lists "Premium Box Set #/199" among the base-set parallels alongside "AUTOGRAPH PARALLELS: Autographs Premium Box #/20."; sportscardchecklist lists it as a product at 290 cards. CONFIRMED AS ALREADY CORRECTED BY THE paninihoops-bkb-2017-18 BUILD ON 2026-09-23: the 290-versus-300 is a UNIT and not a size disagreement - that site prints 290 for the pack product's base set too and counts the ten Kobe cards separately in fifteen places - so all three lineages agree the thing is 300; and card 291 is a BASE card rather than an insert, which this build re-confirmed from this product's own page 30 rendering 291-300 as its own base tiles. CONFIRMED AND NOW THE BACKBONE OF THE RELEASE: the chase set, sid 160903, ten cards 291-300, "AU, SN20" on every row (https://www.tcdb.com/ViewSet.cfm/sid/160903/2017-18-Hoops-Kobe-Bryant-Career-Tribute-Autographs-Premium-Box). STILL UNRESOLVED exactly as section 5b says: the release date. NOT RE-TESTED HERE: section 5b's statement that the Premium treatment does not reach any ordinary insert set, which the paninihoops build established from a 61-row Inserts.cfm read and an 82-entry sportscardchecklist index; this build read the parent's Parallel Sets section in full and asked for every row containing "Premium" across the whole page, which returned only sid 160898 and the sid 160903 autograph set, and that is consistent with it without being an independent re-test of all 45 insert rows.
• G-015 INSERT AND AUTOGRAPH CARD-LEVEL CHECKLISTS ARE OUT OF SCOPE FOR THIS PASS and no insert_cards.csv was written; CHASE_CHECKLIST_SPEC governs them. The one set that would need one is Kobe Bryant Career Tribute Autographs Premium Box, and its ten card numbers are already fully rendered and recorded in that set's notes: 291-300, Kobe Bryant, Los Angeles Lakers (https://www.tcdb.com/ViewSet.cfm/sid/160903/2017-18-Hoops-Kobe-Bryant-Career-Tribute-Autographs-Premium-Box). A later chase pass should note that those numbers MIRROR base numbers 291-300, so card_number is non-unique across this release by construction, which EXTRACTION_SPEC and CHASE_CHECKLIST_SPEC both say is expected and correct.
• G-016 NO base_card_tiers.csv WAS WRITTEN AND THAT IS A DELIBERATE CALL. The base run is uniform on the only axis that file exists to express: the literal token "SN199" appears on all 300 rows and no other serial token appears anywhere in the set, so base_print_run is a flat sourced integer rather than varies-by-tier, and EXTRACTION_SPEC 2.5 makes the file required only in the varies-by-tier case. base_cards.csv's card_tier column does carry the product's internal structure - Main on 1-290 and Kobe Bryant Career Tribute on 291-300 - which is where that fact belongs. Written out because a reader comparing this release with a tiered sibling such as select-bkb-2017-18 will see no tier file and should not read it as skipped work.

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