2016-17 Absolute Memorabilia

• G-001 NO 2016-17 SOURCE PRINTS THE WORD "Memorabilia" IN THIS PRODUCT'S NAME, SO A BYTE-FOR-BYTE MATCHER KEYED ON EITHER SOURCE'S STRING WILL MISS THIS RELEASE. release.csv carries the reference/brands.csv string "Panini Absolute Memorabilia Basketball" verbatim, per the settled owner decision; nothing was minted and neither reference/brands.csv nor pipeline/brands.py was edited. What the sources actually print: tcdb's parent heading is "2016-17 Panini Absolute" and, asked directly, the word "Memorabilia" appears NOWHERE on that page (https://www.tcdb.com/ViewSet.cfm/sid/138222/2016-17-Panini-Absolute); tcdb's Inserts.cfm names all 52 child sets "2016-17 Panini Absolute - ..." (https://www.tcdb.com/Inserts.cfm/sid/138222/2016-17-Panini-Absolute); cardboardconnection's H1 is "2016-17 Panini Absolute Basketball Cards" and its only use of the word is the editorial sentence "Acknowledging the memorabilia side of the brand, 2016-17 Panini Absolute Basketball contains several familiar insert lines dedicated to relic pieces." (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball); sportscardchecklist names all 52 of its entries "2016-17 Panini Absolute ..." (https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-absolute/trading-card-checklists-and-product-information). I ALSO ADDED A THIRD SEASON TO THE MEASUREMENT, which specs/2016-17_SCOPE_MEMO.md section 2b explicitly says the ledger's account is too narrow to cover: tcdb's 2015-16 parent at https://www.tcdb.com/ViewSet.cfm/sid/118430/probe renders "2015-16 Panini Absolute", "Total Cards: 200", "Release Date: February 5, 2016" and no occurrence of "Memorabilia" either. So the drift is now measured on tcdb in 2015-16, 2016-17 and (per the memo) 2017-18, and on two further sources in 2016-17. WHAT A CONSUMER MUST DO: resolve this release by line_prefix or brand id, never by matching a source's printed product string, and expect both "Panini Absolute" and "Panini Absolute Basketball Cards" as aliases of the brands.csv name.
• G-002 CHANNEL IS UNKNOWN ON ALL 28 PARALLEL ROWS AND ALL 3 BASE TIERS, AND THE ONE CHANNEL-ADJACENT SENTENCE IN THE RELEASE COULD NOT BE ATTACHED TO ANY CARD. cardboardconnection states "Some autograph cards only found through the Panini Rewards program." (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball). It names no card, no set and no tier, and "Panini Rewards" is not a value in the channel enum - the nearest, panini-direct, would assert a distribution fact about rows the sentence does not identify. I therefore declined to assign it anywhere, which per EXTRACTION_SPEC 2.3 means channel=UNKNOWN with channel_basis and channel_confidence EMPTY. I also looked for the mechanical route to those cards and did not find it: NO EXCH TOKEN WAS RENDERED ON ANY OF THE 52 SIDS SWEPT, so the exchange or redemption rows the sentence implies are not identifiable from tcdb's flag column on this product. The other channel-adjacent line, "Product Configuration: 5cards per pack, 4 packs per box, 10 boxes per case", is a pack configuration and not a trade channel. WHAT WOULD CLOSE THIS: a Panini Rewards redemption list or a sell sheet naming the sets, or an EXCH token on a later tcdb render.
• G-003 THE UNNUMBERED HALF OF THE BASE TIER SPLIT IS DISCRIMINATIVE-TEST EVIDENCE, NOT A SOURCE STATEMENT, AND release.csv CANNOT SAY SO. base_print_run is varies-by-tier with base_print_run_basis=sourced, because migrations/002 permits basis=assumed only beside base_print_run=unnumbered and 14 of 14 varies-by-tier releases in the corpus write sourced. The two halves of the split are NOT equally evidenced and the row cannot express that. SOURCED IN WORDS: the 999 on cards 101-200, from cardboardconnection's base section labels "Retired - #/999" and "Rookies - #/999" (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball) and from tcdb rendering SN999 on every one of those 100 rows (https://www.tcdb.com/Checklist.cfm/sid/138222/2016-17-Panini-Absolute ?PageIndex=2 and https://www.tcdb.com/Gallery.cfm/sid/138222/2016-17-Panini-Absolute ?PageIndex=18, =19, =20 with &sort=0). NOT SOURCED IN WORDS: the unnumbered on cards 1-100. No source says the base cards carry no serial number. What supports it is a within-sid positive discriminative test in both directions on two endpoints - the same renderer shows an empty flag column on 1-100 and SN999 on 101-200 - which EXTRACTION_SPEC 2.1 classifies as strong indirect evidence and names assumed. A print_run_basis column on base_card_tiers, or permitting basis=assumed beside varies-by-tier, would carry this; until then it lives here.
• G-004 THE FIRST BASE TIER HAS NO SOURCE-GIVEN NAME AND "Base" IS BORROWED FOR IT. base_card_tiers.csv names the three tiers Base (1-100), Retired (101-160) and Rookies (161-200). Retired and Rookies are both sources' own words: cardboardconnection's base section renders the labels "Retired - #/999" and "Rookies - #/999" (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball) and sportscardchecklist files "2016-17 Panini Absolute Retired (60 Cards)" and "2016-17 Panini Absolute Rookies (40 Cards)" (https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-absolute/trading-card-checklists-and-product-information). THE FIRST HUNDRED CARDS ARE NAMED BY NEITHER: cardboardconnection's heading "Base" sits over all 200 cards, and sportscardchecklist's entry for exactly those 100 is the bare "2016-17 Panini Absolute (100 Cards)". So "Base" is cardboardconnection's word applied to the untitled remainder, chosen because it is the only candidate any source prints and because inventing "Veterans" or "Active" would be minting a label. WHY IT MATTERS MECHANICALLY: reference/caveats.csv CAV-043 resolves cards.tier_id by an exact base_cards.card_tier to base_card_tiers.tier_name string match and a mismatch loads NULL and gates green, so all 100 base_cards rows carry card_tier=Base byte for byte and a later rename must change both files in one commit. Note also that "Base" as a tier_name is not the same string as the sets.csv base row's set_name "Base Set"; they are deliberately different and neither is derived from the other.
• G-005 THIS RELEASE'S BASE STRUCTURE CONTRADICTS EVERY ABSOLUTE SEASON ALREADY IN THE CATALOGUE, AND THE DIVERGENCE IS REAL RATHER THAN A CONVENTION DISPUTE. data/absolute-bkb-2018-19, data/absolute-bkb-2019-20 and data/absolute-bkb-2020-21 each carry base_set_size=100, base_print_run=unnumbered, base_print_run_basis=assumed and NO base_card_tiers.csv. This release is base_set_size=200 with three tiers and 100 of its base cards serial numbered to 999. Nothing was inherited from those three seasons; they were read only for what questions to ask, as the brief requires. The divergence is a property of the products - 2016-17 Absolute genuinely ships a 60-card Retired block and a 40-card Rookies block inside its base numbering, each at /999, and the later seasons ship a flat 100 - and it is filed here so that a consumer comparing base_set_size across the line sees a product change and not a build inconsistency. SEPARATELY AND WORTH A SECOND LOOK LATER: the 2018-19 release carries base-set parallels named 10th Anniversary and Maroon and set-scope tiers named Level 2 to Level 5, where this season's ladder is Spectrum Gold / Spectrum Black at base scope and Prime / Patch / Tags / Jumbo Prime / Jumbo Patch at set scope. The parallel vocabulary of the line changes completely between 2016-17 and 2018-19, so a resolver keyed on tier names across Absolute seasons will not match.
• G-006 THE Draft Day Ink CONTAMINATION FINDING LANDS HERE AND IS CONFIRMED, AND I CAN EXTEND IT BY ONE SOURCE. The totallycertified-bkb-2016-17 build found cardboardconnection and sportscardchecklist both listing a 15-card Draft Day Ink set under Totally Certified and proved it is an Absolute set. VERIFIED INDEPENDENTLY HERE, ON MY OWN PARENT. tcdb files a set named exactly "Draft Day Ink" at sid 142363 in the "Insert Sets (50)" section of this parent's own Inserts.cfm (https://www.tcdb.com/Inserts.cfm/sid/138222/2016-17-Panini-Absolute), and that sid's page heading reads "2016-17 Panini Absolute - Draft Day Ink" (https://www.tcdb.com/Checklist.cfm/sid/142363/2016-17-Panini-Absolute-Draft-Day-Ink). MY OWN ENUMERATION AGREES WITH cardboardconnection'S COUNT AND RUN, which is the condition the brief set before repeating them: the page renders 15 rows numbered 1-15 contiguously with "AU, SN25" on every one, and cardboardconnection's Absolute page states "Draft Day Ink Set Checklist" - "15 cards. Serial Numbered #/25." plus the prose "Also focusing on the first-year players, the Draft Day Ink inserts are hard-signed, as well." (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball). So the set is filed in this release's sets.csv as an auto set of 15 cards. THE EXTENSION: sportscardchecklist lists "2016-17 Panini Absolute Draft Day Ink (15 Cards)" on its ABSOLUTE index too (https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-absolute/trading-card-checklists-and-product-information), so that site carries the same 15-card set under BOTH products. Its Absolute filing is correct and its Totally Certified filing is the contamination; a name-keyed search on that site returns the set twice and a consumer deduplicating by name alone will pick arbitrarily. The 15 rows themselves are a CHASE_CHECKLIST_SPEC job and no insert_cards.csv was written in this pass.
• G-007 FIVE PARALLEL ROWS CARRY print_run=UNKNOWN AND serial_numbered=UNKNOWN, AND THE CENSUS THE SCHEMA CANNOT HOLD IS HERE. All five are scope_type=set, where UNKNOWN is legal, so reference/caveats.csv CAV-041 did not have to delete anything in this release. FOUR ARE PER-CARD VARIABLE, enumerated over every row and not sampled. (1) Prime, scope Frequent Flyer Material Autographs, sid 142378: 17 rows over numbers 1-9, 12-18, 20; tokens AU MEM SN10 on 15 rows and AU MEM SN5 on cards 5 and 14. (2) Prime, scope Frequent Flyer Materials, sid 142380: 56 rows over 1-15, 18-29, 31-37, 39-60; MEM SN10 on 54 rows, MEM SN5 on card 46 Kobe Bryant, MEM SN2 on card 51 Julius Erving. (3) Prime, scope Iconic Materials, sid 142358: 28 rows over 1-19, 22-30; MEM SN10 on 26 rows, MEM SN5 on card 1 Kobe Bryant, MEM SN8 on card 23. (4) Prime, scope Team Tandems, sid 142388: 19 rows over 1-19; SN25 on 18 rows and SN5 on card 18 Larry Johnson / Alonzo Mourning. THE FIFTH IS A COVERAGE PROBLEM, NOT A VARIABILITY ONE: Prime, scope Team Quads, sid 142386 renders 5 rows and populates the flag column on exactly one - card 1 reads MEM SN5 and cards 2 to 5 render an empty flag column, identically on Checklist.cfm and on Gallery.cfm?PageIndex=1&sort=0. Writing 5 would be a one-row sample and writing unnumbered would rest on an empty flag column that CAV-042 says is never evidence of absence, so both columns are UNKNOWN. IN EVERY ONE OF THE FIVE THE MODAL TOKEN WOULD HAVE WRITTEN A CLEAN SOURCED-LOOKING INTEGER - 10, 10, 10, 25 and 5 - and four of the five outliers sit on the most valuable cards in their tier, which is the shape reference/caveats.csv CAV-050 describes. A print_run_basis column, or a per-card run table, would convert this gap into recorded fact.
• G-008 A SET'S OWN PRINT RUN HAS NOWHERE TO LIVE, AND ON THIS RELEASE THAT HIDES 19 FIGURES AND 5 VARIABILITY FINDINGS. EXTRACTION_SPEC 2.2 is explicit that sets.csv cannot hold a print run and that the figure must not go in card_count. Every one of this release's 24 insert and autograph sets is serial numbered except Glass, and the figure is a property of the set rather than of a parallel tier, so none of them is expressible as a parallels.csv row. UNIFORM, ENUMERATED OVER EVERY ROW, FIGURE RECORDED IN THE SET'S OWN notes ONLY: Draft Day Ink 25, Frequent Flyer Material Autographs 75, Frequent Flyer Materials 149, Freshman Flyer Jersey Autographs 75, Freshman Flyer Jumbo Jerseys 75, NBA Stars Materials 149, Rookie Autographs 99, Team Tandems 149, Team Trios 49, Tools of the Trade Autographed Materials 75, Tools of the Trade Basketballs Prime Signatures 15, Tools of the Trade Jumbo Rookie Materials 149, Tools of the Trade Jumbo Rookie Materials Signatures 49, Tools of the Trade Rookie Materials Dual Relic 149, Quad Relic 125, Six Swatch Relic 75, Triple Relic 149. Team Quads is 25 per cardboardconnection with tcdb rendering SN25 on 1 of its 5 rows. PER-CARD VARIABLE, SO THERE IS NOT EVEN A FIGURE TO HIDE: Heroes Autographs (AU SN60 on 12 rows, AU SN75 on 13), Iconic Autographs (AU SN60 on 10, AU SN75 on 29), Marks of Fame (AU SN60 on 10, AU SN75 on 20), Heroes Materials (MEM SN149, SN99, SN49 and SN90 - the SN90 on card 16 Steve Nash), Iconic Materials (MEM SN149 on 27, SN49 on cards 12 and 17, SN99 on card 20). Glass alone carries no serial token on any of its 25 rows. A set_print_run column, or permitting a parallels row at scope_type=set with the set's own name, would hold all of this.
• G-009 A SUMMARISING FETCH OF cardboardconnection INVENTED A COUNT AND A PRINT RUN THAT ITS OWN PAGE DOES NOT CARRY, AND THE NARROW RE-ASK CAUGHT BOTH. This is reference/caveats.csv CAV-047's shape on an editorial page rather than on tcdb's Inserts.cfm, and it is recorded because the invented values were plausible and would have been written. The FIRST fetch of https://www.cardboardconnection.com/2016-17-panini-absolute-basketball, asked for a numbered list of insert sets with counts and print runs, returned among others "Heroes Materials: 40 cards (varying print runs)", "Heroes Autographs: 27 cards (varying print runs #/60-#/75)" and "Iconic Materials: 30 cards, #/149" - all OUTSIDE quotation marks. The SECOND fetch of the same page, asked to return each entry's own text character for character inside quotation marks, returned "Heroes Materials Set Checklist" - "38 cards.", "Heroes Autographs Set Checklist" - "27 cards." and "Iconic Materials Set Checklist" - "30 cards.", with NO print-run sentence on any of the three. So the 40, the varying-print-run gloss, the #/60-#/75 range and the flat #/149 were the fetch's and not the page's. CONSEQUENCES ACTED ON: card_count for Heroes Materials is written as 38, which is my own count of its rendered rows and also tcdb's header and sportscardchecklist's figure; no cardboardconnection corroboration is claimed for any variable tier; and the Iconic Materials variability finding is recorded as CONTRADICTING a flat 149 that in fact no source states. WHAT A LATER PASS SHOULD TAKE FROM THIS: every cardboardconnection figure cited anywhere in this release comes from the verbatim re-ask, not the first fetch.
• G-010 Errors.cfm RETURNED NOTHING AND NO CONCLUSION IS DRAWN FROM IT. https://www.tcdb.com/Errors.cfm/sid/138222/2016-17-Panini-Absolute renders the heading "2016-17 Panini Absolute" and the sentence "No errors or variations have been added for this set yet." with zero rows. Per reference/caveats.csv CAV-020 that sentence is returned identically for sets that provably have errata and for sets that appear not to, so it discriminates nothing. THIS RELEASE THEREFORE MAKES NO CLAIM THAT IT HAS NO VARIATION LAYER. What does rest on real evidence is the base run itself: 200 rows over 200 distinct numbers, no number repeated, enumerated on two endpoints with both overlaps agreeing and with Gallery.cfm?PageIndex=21 rendering zero tiles. No b-suffixed or c-suffixed row was rendered anywhere in the run on either endpoint, and no UER or VAR token appeared on any of the 52 sids swept - but per CAV-020 that is unsupported rather than disproved, because a variation set renders as suffixed rows that a contiguous base enumeration is not designed to surface.
• G-011 THE SOURCE BASE IS THREE LINEAGES FOR COUNTS AND TWO FOR PRINT RUNS, AND ONE WHOLE LAYER OF THE RELEASE RESTS ON tcdb ALONE. USED: tcdb - 61 page fetches, covering the parent (138222), both base parallel sids and all 50 insert sids Inserts.cfm lists, plus 19 ViewSet.cfm boundary and gap probes and one Errors.cfm read; cardboardconnection's PRODUCT page https://www.cardboardconnection.com/2016-17-panini-absolute-basketball; and sportscardchecklist's product index https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-absolute/trading-card-checklists-and-product-information, reached from a search result and never from a constructed path per reference/caveats.csv CAV-037. INDEPENDENCE ASKED PER PRODUCT, as CAV-037 as corrected requires, and answered by looking for checkable divergences: cardboardconnection carries the Team Quads #/25 figure that tcdb renders on only 1 of 5 rows, and it OMITS four sets that both tcdb and sportscardchecklist carry (Tools of the Trade Rookie Materials Dual, Triple, Quad and Six Swatch Relic), so it is not a copy of tcdb; sportscardchecklist carves the base run into three entries and the Spectrum tiers into six where tcdb files one parent and two sids, and it carries those same four omitted sets, so it is not a copy of cardboardconnection either on this product - although it does reproduce cardboardconnection's Draft Day Ink double-filing, which is a shared-upstream signature pointing at one Panini sell sheet beneath both. I therefore treat all three as usable for COUNTS and write source_confidence=multi only on the two base-set parallel rows, where both tcdb and cardboardconnection state the figure. RESTING ON tcdb ALONE: every one of the 26 set-scope parallel print runs, because cardboardconnection's verbatim 20-entry list names no Prime, Patch, Tags or Jumbo tier at all and sportscardchecklist states no print run anywhere on its index; all five per-card variability findings; and every set's row count as a contiguous enumeration. NOT REACHED: beckett.com is confirmed unreachable per CAV-033 and was not budgeted for - a search surfaced the title "2016-17 Panini Absolute Basketball Checklist and Details" at a beckett.com/news URL and the page was never opened, so the title is cited as a title and not as evidence. checklistcenter.com surfaced no page for this product in the searches run, and per CAV-037 no slug was constructed, so its absence is evidence of nothing. breakninja and gogts.net surfaced in search results and were not opened. blog.paniniamerica.net was not swept for this product.
• G-012 NO FLOOR IS CLAIMED FOR THIS LINE AND NONE SHOULD BE INFERRED, AND THE LEDGER'S DEPTH QUOTE IS NOT REPRODUCIBLE ON THIS SEASON'S PAGE. I opened the predecessor myself rather than inheriting it: https://www.tcdb.com/ViewSet.cfm/sid/118430/probe renders "2015-16 Panini Absolute", "Total Cards: 200" and "Release Date: February 5, 2016", so at least one season runs below 2016-17. specs/2016-17_SCOPE_MEMO.md section 3a records the line as absent from tcdb's 2014-15 index, and per that memo's section 3b Studio lesson ABSENCE FROM AN ADJACENT INDEX IS NOT A FLOOR - a floor needs a positive statement. I looked for one and did not find it: asked directly for any sentence containing the word "years" or any statement about how long the Absolute line has existed, cardboardconnection's 2016-17 product page returned NOT PRESENT (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball). So the "10 years" quote the existing ledger carries for this line, which would imply a debut around 2009-10, is NOT on the 2016-17 page and was not reproduced here; whatever season it came from, it is not a 2016-17 source. INCIDENTAL AND NOT CHASED, as the brief instructed: the 2015-16 parent's own 200-card header matches this season's, so the 200-card Retired-plus-Rookies shape may well run below 2016-17 too, and whoever builds 2015-16 should expect a tiered base run rather than the flat 100 the 2018-19 and later seasons carry.
• G-013 pipeline/fix_schema.py's FAMILIES LIST HAS NO PATTERN FOR EITHER OF THIS RELEASE'S TWO PARALLEL VOCABULARIES, SO ALL 28 ROWS DERIVE TO "other". parallel_family was derived by AST-lifting FAMILIES and family_of out of pipeline/fix_schema.py so the repo keeps one definition, and NO ENTRY WAS ADDED to that file, per reference/caveats.csv CAV-039. family_of returns "other" for all seven distinct names in this release: Spectrum Black, Spectrum Gold, Prime, Patch, Tags, Jumbo Prime and Jumbo Patch. TWO SEPARATE TAXONOMY QUESTIONS SIT BEHIND THAT. FIRST, Spectrum is this line's base-ladder finish name in 2016-17 and nothing in FAMILIES matches it; whether it deserves a family cannot be decided from one release, and minting one here would split it from any 2015-16 or 2017-18 Absolute rows a later pass adds. SECOND, Prime, Patch, Tags, Jumbo Prime and Jumbo Patch are a patch-quality and swatch-type ladder, not visual finishes, and that is exactly the unresolved question reference/caveats.csv CAV-021 names - Prime and Super Prime already collapse into "other" across 217 corpus rows, and nattreasures-bkb-2018-19 deliberately did not mint a Prime family for the same reason. This release adds 26 more rows to that collapsed population. I CHECKED FOR THE ORDERED-SUBSTRING HAZARD AND IT DOES NOT FIRE HERE: none of the seven names contains any FAMILIES pattern as a substring, so none of them is a Pink-inside-Ink or Cosmo-inside-Cosmos case, and no guard entry is needed.
• G-014 INTERIOR NUMBER HOLES IN NINE SETS WERE NOT CHECKED AGAINST A PARALLEL PAGE IN EVERY CASE, SO NONE OF THEM IS RECORDED AS AN UNISSUED NUMBER. reference/caveats.csv CAV-044 is explicit that a hole in a page's numbering is not evidence of an unissued number until a parallel page of the same product has been opened. The holes, all from my own enumeration of the rendered rows: Frequent Flyer Materials lacks 38 against a 1-60 span; Frequent Flyer Material Autographs lacks 11 and 19 against 1-20; Freshman Flyer Jumbo Jerseys lacks 10 against 1-40; Heroes Materials lacks 13 and 15 against 1-40; Heroes Autographs lacks 1, 2 and 20 against 1-30; Iconic Autographs lacks 23 against 1-40; Iconic Materials Prime lacks 20 and 21 against 1-30; NBA Stars Materials Prime lacks 28 against 1-30; Team Trios Prime lacks 9 against 1-10; Tools of the Trade Basketballs Prime Signatures lacks 7 and 10 against 1-33; the Six Swatch Relic Prime and Patch tiers lack 3 while the base set and the Tags tier render all 33. WHERE THE CHECK WAS ACTUALLY RUN IT PASSED: card 38 is absent from Frequent Flyer Materials Prime as well as from the base set, and card 10 is absent from Freshman Flyer Jumbo Jerseys Prime as well, so those two holes are corroborated across two sids. The others rest on one page each. In every case the row count equals the page's own Total Cards, so these are not truncation artifacts; they are either genuinely unissued numbers or tcdb rows nobody has entered, and this build does not claim to know which. NOTE that the BASE RUN has no hole at all - 200 rows over 200 contiguous numbers - so reference/caveats.csv CAV-049's non-contiguity case does not arise here.
• G-015 BOTH BOUNDARIES OF BOTH SID BLOCKS WERE PROVED AND EVERY NUMERIC GAP INSIDE THE LISTING WAS PROBED, AND THE PROBES FOUND TWO CROSS-SPORT TRAPS WORTH RECORDING. This product owns 52 child sids in two neighbourhoods, 142357-142390 (19 sids) and 142683-142715 (33 sids, contiguous), which is exactly the 50 Insert Sets plus 2 Parallel Sets that Inserts.cfm's own section headings state (https://www.tcdb.com/Inserts.cfm/sid/138222/2016-17-Panini-Absolute). The parent, 138222, sits about 4,135 sids below the nearest child, reproducing specs/2016-17_SCOPE_MEMO.md section 7.1 on a fifth parent. ALL 19 PROBES WERE RUN WITH ViewSet.cfm/sid/N/probe AND THE HOME-PAGE SIGNATURE WAS DISTINGUISHED EXPLICITLY ON EACH. Block one lower bound 142356 is "2016-17 Parkhurst - Protectors of the Net", Total Cards 8 - a hockey set. Block one upper bound 142391 renders the tcdb HOME PAGE (heading "Welcome!", "Cards Listed: 35,440,084"), the documented non-existent-sid signature. Interior gaps: 142359 "2011-12 Upper Deck Ultimate Collection - 2009-10 Upper Deck Ultimate Collection Update I: Ultimate Signatures"; 142360 "1983-84 Vachon - Panels"; 142364 to 142369 all render the home page, six non-existent sids in a row; 142371 "2013-14 SP Authentic - 2009-10 SP Authentic Update: Sign of the Times"; 142373 "2002-03 Upper Deck Inspirations - UD Promos"; 142374 "1993 Kentucky Collectables Jamal Mashburn"; 142375 "1989 Star Johnson City Cardinals - Platinum"; 142376 "1995-96 SplitSecond Saint John Flames (AHL)"; 142377 "1997-98 Sheffield Steelers (BHL)". TRAP ONE: 142372 is "2016 Panini Phoenix - Red" with "Total Cards: 200" - a 2016 Panini FOOTBALL set sitting inside my own sid block whose header figure is identical to my parent's. TRAP TWO, AND IT IS THE SAME SHAPE THE SCOPE MEMO FOUND ABOVE PRESTIGE: both boundaries of block two are 2016 Panini CERTIFIED football sets - 142682 is "2016 Panini Certified - Epix Mirror Purple", Total Cards 20, and 142716 is "2016 Panini Certified - Famed Rides", Total Cards 20. certified is a reference/brands.csv line and the memo's section 4a labels it confusable-with-another-product for this season on the Totally Certified collision; this is a SECOND, independent instance of that confusion, from a different sport, immediately adjacent to an NBA product's sid block in both directions. Anyone sweeping outward from this block lands on pages that read as plausible Panini Certified releases.
• G-016 MARKETPLACE CORROBORATION WAS NOT ATTEMPTED AND NO CARD-LEVEL CHASE CHECKLIST WAS WRITTEN. No insert_cards.csv exists for this release; the 24 insert and autograph sets have set-level rows only, which is what EXTRACTION_SPEC scopes this pass to. Nothing in this build needed a marketplace title, because no plain integer was written onto an autograph set - every card number written here is a base card, enumerated contiguously 1-200 on two tcdb endpoints - so the required-before-writing-integers check was never triggered. It follows that this release has NO evidence either way about whether marketplace corroboration works on 2016-17 Absolute, which specs/2016-17_SCOPE_MEMO.md section 9 lists as unmeasured for the whole season. A later chase pass should expect letter codes on these sets and should calibrate per set type: tcdb renders plain 1..N numbering on every one of the 50 insert sids read here, including all ten autograph sets, which per CHASE_CHECKLIST_SPEC is itself a renumbering smell that must be tested against a seller's free-text title before any of those numbers is written as a card_number.
• G-017 scarcity_tier IS EMPTY ON ALL 28 PARALLEL ROWS AND stated_odds IS EMPTY EVERYWHERE, BOTH DELIBERATELY. Per reference/caveats.csv CAV-005 an EMPTY scarcity_tier means the scarcity pass never covered the row, which is accurate - this build placed no row on a scarcity ladder - and is different from UNKNOWN, which would mean it looked and could not place it. Per CAV-015 a blank stated_odds on a Panini row means unpublished rather than unresearched: no odds statement of any kind was found on tcdb, cardboardconnection or sportscardchecklist for this product, and Panini publishes no numeric odds for these releases. The only distribution figure anywhere is cardboardconnection's "Product Configuration: 5cards per pack, 4 packs per box, 10 boxes per case" (https://www.cardboardconnection.com/2016-17-panini-absolute-basketball), which is not an odds statement.
• G-018 rookie_flag=no ON BASE CARDS 1-160 RESTS ON A DISCRIMINATIVE TEST RATHER THAN ON AN EXPLICIT NON-ROOKIE DESIGNATION. EXTRACTION_SPEC 2.4 permits yes or no only where the source explicitly shows an RC designation, and reference/caveats.csv CAV-003 records that most Panini checklist sources print no RC designation at all. HERE tcdb DOES: cards 161-200 render "RC, SN999" on both Checklist.cfm page 2 and Gallery.cfm PageIndex=18, =19 and =20 under &sort=0, and cards 1-160 render no RC token on either endpoint - so the renderer demonstrably emits RC where it applies and withholds it elsewhere within the same sid, which is the positive discriminative test in both directions. On that basis 40 rows carry yes and 160 carry no. A reader who wants only explicitly-designated values should treat the 160 no values as test-derived rather than printed. The 40 yes rows are also independently consistent with the 2016 draft class and with cardboardconnection's and sportscardchecklist's Rookies block of exactly 40 cards.
• G-019 TRUNCATION CENSUS FROM THIS RELEASE, OFFERED AS AN ADDITION TO reference/caveats.csv CAV-046 RATHER THAN AS A REMEDY TO PLAN AGAINST. Every figure is my own count of rendered rows against the page's own number span. 100-row Checklist.cfm pages: parent page 1 gave 73 rows over a 1-73 span (clean tail cut); parent page 2 gave 77 clean rows over 101-177 and then cut MID-ROW inside card 178; Spectrum Black page 1 gave 80 over 1-80 and page 2 gave 72 over 101-172; Spectrum Gold page 1 gave 75 clean rows plus a mid-row cut inside card 76 and page 2 gave 73 over 101-173. So five of five 100-row pages here were clean tail cuts with row count equal to number span - NOT the interior-drop shape CAV-046 records on optic and the Hoops parent - and the census additions are 72, 73, 73, 75 and 80 of 100, all well above the floor of 25. THE MID-ROW CUT IS THE FINDING WORTH CARRYING: on Spectrum Gold the truncation landed inside card 76 and the fetch reported that row's flag column as EMPTY while reporting SN10 on the 75 rows above it, which reads exactly like one unnumbered card inside a numbered tier. Gallery.cfm?PageIndex=8&sort=0 renders card 76 Solomon Hill as SN10. A mid-row cut therefore produces a FALSE EMPTY FLAG rather than a missing row, and comparing row count against number span does NOT catch it - the span was right. Checking the last rendered row on a second endpoint does. Separately, ?sort=0 was used on every Gallery.cfm fetch as a cache-bust and no Gallery page in this release ever returned fewer than its full ten tiles, so the dropped-tile defect did not reproduce here.
• G-020 sportscardchecklist CARVES THIS RELEASE DIFFERENTLY FROM tcdb IN TWO PLACES AND THE DIVERGENCE IS RECORDED RATHER THAN RESOLVED SILENTLY. FIRST, THE BASE RUN. tcdb files one parent of 200 rows; sportscardchecklist files three entries, "2016-17 Panini Absolute (100 Cards)", "2016-17 Panini Absolute Retired (60 Cards)" and "2016-17 Panini Absolute Rookies (40 Cards)" (https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-absolute/trading-card-checklists-and-product-information). 100 + 60 + 40 = 200, so the two describe the same cards, and its carve independently confirms the block boundaries this build read off tcdb's flag column. I wrote ONE sets.csv base row of 200 because tcdb files one parent, both Spectrum sids render all 200 numbers, and cardboardconnection counts "200 cards." - and I expressed the carve as base_card_tiers rows, which is where reference/caveats.csv CAV-020 says a stratified base run belongs. SECOND, THE SPECTRUM TIERS. tcdb files two sids, each rendering all 200 rows; sportscardchecklist files six entries - Spectrum Black at 100, Retired Spectrum Black at 60, Rookies Spectrum Black at 40, and the same three for Gold. Each triple sums to 200. I wrote TWO base-set parallel rows rather than six, because the tier that exists on tcdb covers the whole base run and scope_type=base-set is the accurate description of a tier that spans every base number; six rows would also have needed block scope, and migrations/003_parallels.sql gives every block row scope_set_id NULL so block rows must be globally unique by name, which Retired Spectrum Black and Rookies Spectrum Black would have satisfied but at the cost of asserting a three-way split no other source makes. WHAT A LATER PASS MUST CHECK IF IT DISAGREES: whether the Retired and Rookies Spectrum cards are physically numbered to a different run from the 1-100 Spectrum cards. tcdb says they are not - SN1 on all 200 Black rows and SN10 on all 200 Gold rows, enumerated row by row - so on the evidence available the six-entry carve is a display convention and not a scarcity fact.

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