2017-18 Ascension

• G-001 RELEASE DATE LINEAGE. 2017-11-22 is written, but NOT from the string most sources lead with. cardboardconnection carries only "Estimated Release Date: 11/22/2017", and EXTRACTION_SPEC section 3.2 says a source that labels its own date an estimate has told you it does not know; that string was refused. The value rests on two other renderings. (1) tcdb's parent set page renders the unqualified line "Release Date: November 22, 2017", read back on two different slugs (https://www.tcdb.com/ViewSet.cfm/sid/159372/2017-18-Panini-Ascension and https://www.tcdb.com/ViewSet.cfm/sid/159372/Ascension-2017-18-Panini-Basketball-Cards), with the word "Estimated" appearing nowhere on the page -- the same class of evidence the prestige-bkb-2017-18 build wrote its date from. (2) gogts.net, byline "Nov 21, 2017", prints "Panini brings their latest stylish and colorful NBA product to market this Wednesday, November 22nd, 2017." (https://gogts.net/2017-18-panini-ascension-nba-basketball-cards-checklist/) -- off-lineage from cardboardconnection, and the same source and the same one-day-ahead shape that the encased-bkb-2017-18 build accepted for Encased. November 22 2017 was in fact a Wednesday, so the gogts line is internally consistent. THE RESIDUAL RISK, STATED PLAINLY: unlike prestige, where tcdb DISAGREED with the estimate (Feb 28 against an estimated Feb 21) and the disagreement was itself the proof of independence, here both unqualified renderings agree exactly with the estimate. Agreement with a forecast is also what shared lineage produces, so tcdb's field may have been populated from the same sell-sheet date. What defeats the objection only partly: gogts published one day before the stated date rather than three months ahead like the blowoutbuzz preview that the prestige build correctly refused, and by then a street date is known rather than estimated. A first-party Panini source, or a dated post-release retail or grading record, would settle it. Nobody has found one.
• G-002 base_print_run=unnumbered rests on a discriminative test, not on a statement, so base_print_run_basis is "assumed" and not "sourced". No source anywhere says the base cards carry no serial number. The evidence is within-parent and runs in both directions: all 140 base rows render no serial token across https://www.tcdb.com/Checklist.cfm/sid/159372/2017-18-Panini-Ascension?PageIndex=1, https://www.tcdb.com/Checklist.cfm/sid/159372/2017-18-Panini-Ascension?PageIndex=2 and Gallery.cfm pages 7 to 20, while the five colour sids of the same parent render SN1, SN25, SN50, SN75, SN125 and SN129 and the autograph sids render AU together with SN tokens, so the renderer demonstrably prints the flag column for this product. EXTRACTION_SPEC section 2.1 is explicit that this shape is "assumed".
• G-003 tcdb RENDERS 60 BASE CARDS WITH A NON-INTEGER NUMBER AND base_cards.csv WRITES THE INTEGER SLOT. On Checklist.cfm the parent renders numbers 81 to 140 as a and b pairs -- "81a | Chris Paul | Houston Rockets | NONE" beside "81b | Chris Paul | New Orleans Hornets | VAR", and "101a | Markelle Fultz | Philadelphia 76ees"-style pairs through 140b. cards.card_number is INTEGER NOT NULL, so base_cards.csv carries the 140 integer slots 1 to 140 and the 60 b rows are not in it; they are the two variation layers, which are filed in sets.csv instead. The suffix is tcdb's disambiguator rather than a character known to be printed on the card: Gallery.cfm renders the same rows as plain "81" twice, and no source quotes a printed suffix. Precise finding to carry forward: which of the three forms (81, 81a, or something else) the card itself shows is unresolved. Evidence: https://www.tcdb.com/Checklist.cfm/sid/159372/2017-18-Panini-Ascension?PageIndex=2 https://www.tcdb.com/Gallery.cfm/sid/159372/2017-18-Panini-Ascension?PageIndex=9 https://www.tcdb.com/Errors.cfm/sid/159372/2017-18-Panini-Ascension . NOTE for a future pass: Gallery.cfm here strips the a/b suffix but DOES render the VAR token, which is the opposite of what the 2017-18 scope memo section 7.7 reports for National Treasures ("the variation token invisible"). Both observations may be correct per-parent; neither should be generalised.
• G-004 ONE CATALOGUED TIER AT TWO PRINT RUNS, WRITTEN AS TWO ROWS BECAUSE THE SCHEMA CANNOT HOLD ONE. tcdb catalogues Blue as a single sid, 162997, "Total Cards: 140", whose Inserts.cfm Notes cell reads "Serial numbered to 125". A sweep of all 180 of its rows shows numbers 1-100 render SN125 while numbers 101-140 render SN129 on a and b rows alike, so that prose is true of 100 of its 140 numbers and false of 40. cardboardconnection prints the same asymmetry independently -- "PARALLEL CARDS: Blue #/125, Red #/75, Green #/25, Black 1/1." under Base Set and "PARALLEL CARDS: Blue #/129, Red #/75, Purple #/50, Green #/25, Black 1/1." under Base Rookies -- so /129 is a real figure and not a transcription slip. This is caveat CAV-036's third shape (prose true of only part of its own set, first measured on the chronicles-bkb-2017-18 parent) reproduced on a second line, and neither a format check nor the both-directions blank-annotation test would have caught it, because the cell IS annotated and the set IS numbered. Writing the modal token would have produced a confident false 125 on 40 numbers; writing UNKNOWN was not available either, because unknown_run_only_when_scoped forbids print_run=UNKNOWN on a base-set-scope row. The tier is therefore two block-scope rows, "Blue" against tier Base Set at 125 and "Rookies Blue" against tier Rookies at 129. Evidence: https://www.tcdb.com/ViewSet.cfm/sid/162997/2017-18-Panini-Ascension-Blue https://www.tcdb.com/Checklist.cfm/sid/162997/2017-18-Panini-Ascension-Blue?PageIndex=2 https://www.tcdb.com/Inserts.cfm/sid/159372/2017-18-Panini-Ascension https://www.cardboardconnection.com/2017-18-panini-ascension-basketball-cards/amp . See also G-005.
• G-005 THE PARALLEL NAME "Rookies Blue" IS A DISAMBIGUATOR THIS BUILD COMPOSED AND NO SOURCE PRINTS IT. Every source prints the tier simply as "Blue": tcdb has exactly one Blue sid (162997), sportscardchecklist exactly one "2017-18 Panini Ascension Blue (140 Cards)" entry, and cardboardconnection the bare word "Blue" under both of its PARALLEL CARDS headings. EXTRACTION_SPEC section 2.3 requires every block-scope row in a release to be globally unique by NAME, because a block row's scope_set_id is NULL unconditionally and parallel_unique_in_scope therefore keys them all into one group; two rows named "Blue" would collide at load. The spec prefers a source's own longer form over an invented disambiguator, and there is none here -- checked against tcdb, cardboardconnection, sportscardchecklist, breakninja (which carries no set or parallel list at all, only a team-sorted checklist) and gogts.net (which states no parallel names or print runs). The chosen string composes cardboardconnection's own label for the block ("40-card Rookies subset", "Base Rookies Set Checklist") with the printed colour, so no word in it is invented, but the string as a whole is not printed anywhere. A RESOLVER MUST NOT TREAT IT AS A PRINTED NAME. If the schema later gains a way to express one tier at two runs, or if a source turns up that names the rookie tier distinctly, this row should be revisited together with G-004.
• G-006 PER-CARD VARIABLE PRINT RUN, RECORDED HERE BECAUSE THE CSV CANNOT CARRY IT. Autographs Green (tcdb sid 163003, "Total Cards: 67") is PROVABLY serial numbered on every one of its 67 rows and the run differs by card, so EXTRACTION_SPEC section 2.3's forced pairing collapses it to UNKNOWN/UNKNOWN, which is indistinguishable from having found nothing. The full census, swept over all 67 rows rather than sampled: SN25 on the majority, SN10 on cards 2, 3, 4, 5, 12, 18, 36, 53, 65, 68, 70, 77, 80, 83, 89 and 93; SN5 on cards 7, 15, 34, 44, 63, 88, 90 and 91; SN3 on cards 33, 55 and 92. Four distinct tokens. Evidence: https://www.tcdb.com/ViewSet.cfm/sid/163003/2017-18-Panini-Ascension-Autographs-Green . NOTE ON HOW NEARLY THIS WENT WRONG: this sid's Inserts.cfm Notes cell reads "3.0% of images added", an image-completion percentage carrying no serial information, while the base-set Green tier of the same product is a flat /25 and tcdb's own Notes cell says so. An agent calibrating "Green" on the Parallel Sets section and carrying it into the Insert Sets section would have written a confident false 25 on 67 rows.
• G-007 A SET'S OWN PRINT RUN THAT NO COLUMN CAN HOLD. Autographs (tcdb sid 163001, "Total Cards: 68") is serial numbered on every row with a run that differs by card: the 68 rows carry THIRTY-FOUR distinct serial tokens -- SN5, SN6, SN8, SN10, SN14, SN15, SN17, SN20, SN25, SN26, SN28, SN30, SN35, SN38, SN44, SN50, SN53, SN60, SN68, SN71, SN75, SN77, SN87, SN99, SN100, SN106, SN111, SN125, SN142, SN144, SN149, SN178, SN180 and SN199. sets.csv has no print_run column, EXTRACTION_SPEC section 2.2 forbids putting the figure in card_count, and it is not a parallels.csv row because it is the base tier of the set rather than a parallel of it. It is therefore in the sets.csv notes in words and here as the full census. No editorial source states a run for this set at all: cardboardconnection gives it only "Base Autographs Set Checklist 68 cards." Evidence: https://www.tcdb.com/ViewSet.cfm/sid/163001/2017-18-Panini-Ascension-Autographs https://www.cardboardconnection.com/2017-18-panini-ascension-basketball-cards/amp
• G-008 CHANNEL IS UNKNOWN ON ALL 17 PARALLEL ROWS AND THE BASIS IS DELIBERATELY EMPTY. The only channel evidence that exists for this product is product-level: cardboardconnection says "each hobby box has one signed card to go along with a virtual rainbow of colorful designs." and "Each hobby box is slated to have: 1 Autograph", and the words retail, Target, Walmart, blaster, mega, exclusive, SRP and price appear nowhere on that page (https://www.cardboardconnection.com/2017-18-panini-ascension-basketball-cards). Nothing states a channel for any individual tier. EXTRACTION_SPEC section 2.3 records an incident in which an agent was right to decline to inherit exactly this kind of product-level hobby note onto tiers it did not cover, and the database enforces (channel = 'UNKNOWN') = (channel_basis IS NULL), so channel_basis is empty rather than "unsupported". channel_confidence is also left empty per the same section. WHAT WOULD CLOSE THIS CHEAPLY: a single source stating whether this product shipped in any configuration other than hobby boxes would let a channel pass promote all 17 rows at once. Note that no retail configuration was found, which is not the same as one not existing.
• G-009 BASE SET SIZE 140 AGAINST A SOURCE THAT PRESENTS IT AS 100 PLUS 40. Two independent lineages state 140 directly -- tcdb's parent "Total Cards: 140" and sportscardchecklist "2017-18 Panini Ascension (140 Cards)" -- and the parent enumerates as a contiguous 1 to 140. cardboardconnection instead heads two checklists, "Base Set Checklist 100 cards." and "Base Rookies Set Checklist 40 cards.", which is a presentation split rather than a size disagreement: its own prose reads "Anchoring the 2017-18 Panini Ascension Basketball checklist is the dynamic base set and 40-card Rookies subset." The two blocks are also given different parallel ladders by that page (Purple exists on the rookies block only), which is why base_card_tiers.csv carries them as two tiers rather than one. EXTRACTION_SPEC section 5.3's exclusion rule was considered and does NOT apply: cards 101-140 carry the RC token and nothing else, with no AU or MEM token anywhere in the parent, so they are base rookie cards rather than an autograph block numbered past the base run. THE SAME CALL WAS MADE ON THIS SEASON'S SIBLING BUILD AND AGREES: prestige-bkb-2017-18 sets base_set_size=200 including its RC-flagged cards 151-200, and courtkings-bkb-2017-18 sets 232 including four 33-card RC rookie tiers. No release in data/ disagrees with the call made here.
• G-010 TWO VARIATION LAYERS FILED AS "insert" ROWS BECAUSE THE set_type ENUM HAS NO VARIATION VALUE, AND THEY CARRY DUPLICATE BASE NUMBERS. Base Veteran Throwback Variations (20 cards, base numbers 81-100) and Base Rookies Variation (40 cards, base numbers 101-140) have no tcdb sid of their own; Errors.cfm on the parent returns all 60 in one fetch with per-row annotations (https://www.tcdb.com/Errors.cfm/sid/159372/2017-18-Panini-Ascension). They are filed rather than gapped because EXTRACTION_SPEC section 5.5's repeated lesson is that filing is what makes a block visible and filterable while a gap makes it invisible to every consumer, and because two further lineages carry each as its own checklist with the same count (cardboardconnection "Base Veteran Throwback Variations Set Checklist 20 cards." / "Base Rookies Variation Set Checklist 40 cards."; sportscardchecklist "Veteran Throwback (20 Cards)" / "Variation (40 Cards)"). THREE THINGS A REVIEWER SHOULD WEIGH. (1) set_type=insert is the least wrong of three legal values, not a claim that these are pack inserts. (2) Their card numbers duplicate base numbers 81-140, so (set_name, card_number) rather than card_number alone is the only safe key across this release. (3) The names used are cardboardconnection's fuller forms; sportscardchecklist's shorter forms are recorded here so a resolver can match either.
• G-011 A LINEAGE IS ONE TIER SHORT, AND IT IS THE ONE THAT IS USUALLY STRONGEST ON PRINT RUNS. cardboardconnection's checklist carries "Base Autographs Set Checklist 68 cards." and "Base Autographs Green Set Checklist 67 cards." and NO Autographs Black heading at all, while tcdb catalogues sid 163002 "2017-18 Panini Ascension - Autographs Black" with "Total Cards: 68" and SN1 on all 68 rows, and sportscardchecklist lists "2017-18 Panini Ascension Autographs Black (68 Cards)". Two lineages against one. The tier is written from the two that carry it. Evidence: https://www.cardboardconnection.com/2017-18-panini-ascension-basketball-cards/amp https://www.sportscardchecklist.com/sport-basketball/year-2017-18/index-panini-ascension/trading-card-checklists-and-product-information https://www.tcdb.com/ViewSet.cfm/sid/163002/2017-18-Panini-Ascension-Autographs-Black . Worth noting for the season: cardboardconnection carries a fact the others lack on the SAME product (the Blue /129 asymmetry, G-004) while lacking one they have here, so on this product the two are demonstrably not one opinion in either direction.
• G-012 NO PACK ODDS EXIST FOR THIS RELEASE FROM ANY REACHABLE SOURCE, so every stated_odds field in sets.csv and parallels.csv is empty. cardboardconnection carries only a product configuration ("5 cards per pack, 12 packs per box, 12 boxes per case") and a per-box statement ("Each box should contain one base RC cards and one Rookie Variation."), neither of which is an odds statement. beckett.com, which is where odds for this era usually live, 302s to https://maintenance.beckett.com/ on this product's own news URL (https://www.beckett.com/news/get-ready-burst-color-2017-18-panini-ascension-basketball/), reproducing the 2017-18 scope memo's Beckett finding on a third URL. No checklistcenter page for this product surfaced through a web search, so unlike most 2017-18 products this one has no cardboardconnection/checklistcenter pair at all.
• G-013 MEASURED NEGATIVE, FOR THE ORCHESTRATOR TO PLACE: THIS LINE APPEARS TO HAVE NO PREDECESSOR AND NO SUCCESSOR, AND THE CLAIM WAS TESTED RATHER THAN REPEATED. cardboardconnection asserts "Hoping to hitch a ride on the surging influx of NBA talent, 2017-18 Panini Ascension Basketball is a new brand in the Panini portfolio." Four independent index probes agree with it. (1) tcdb's 2016-17 season index renders 79 entries across 13 sections and no Ascension entry, while provably rendering "2016-17 Panini Prestige" (sid 146982), "2016-17 Panini Revolution" (sid 140756) and "2016-17 Panini Excalibur" (sid 149101), so the page is not truncating in a way that would hide one (https://www.tcdb.com/ViewAll.cfm/sp/Basketball/year/2016). (2) tcdb's 2015-16 index renders 57 entries and no Ascension, while rendering "2015-16 Panini Prestige" (sid 119634) and "2015-16 Panini Complete" (sid 120516) (https://www.tcdb.com/ViewAll.cfm/sp/Basketball/year/2015). (3) cardboardconnection's OWN 2016-17 season index, reached through a search result rather than a constructed path per caveat CAV-037, lists 47 products and no Ascension, while rendering Prestige, Studio and Excalibur (https://www.cardboardconnection.com/sports-cards-sets/nba-basketball-cards/2016-2017-basketball-cards). (4) tcdb's 2018-19 index carries no Ascension either, so there is no successor season (https://www.tcdb.com/ViewAll.cfm/sp/Basketball/year/2018). THE LIMITS OF THIS, STATED: all four are editorial or catalogue indexes, beckett is unreachable, and nobody swept Panini's own archive or looked for a solicited-but-cancelled predecessor. This release does NOT write reference/data_line_coverage.csv and is not permitted to; the finding is recorded here for whoever does.
• G-014 SET-NAME DIVERGENCES RECORDED RATHER THAN RESOLVED, so a resolver can match either form. tcdb's headings and sportscardchecklist both write "New Frontiers Die Cuts" and "Overdrive Die Cuts" while cardboardconnection writes "New Frontiers Die-Cuts" and "Overdrive Die-Cuts" with a hyphen. cardboardconnection and tcdb write "The Thrill of Victory" while sportscardchecklist drops the article to "Thrill of Victory". tcdb calls the base autograph set "Autographs" and its rookie set "Rookie Ascent Autographs" while cardboardconnection prefixes "Base" to the former and sportscardchecklist shortens the latter to "Rookie Ascent". In each case sets.csv carries the form two of the three lineages print. No name was changed to tidy it and no set was merged.
• G-015 A ONE-CARD ROSTER DIFFERENCE INSIDE THE AUTOGRAPH LADDER, EXPLAINED RATHER THAN LEFT AS A SHORTFALL. Autographs and Autographs Black both render 68 rows against a header of 68; Autographs Green renders 67 against a header of 67. The single base number present in the 68-row tiers and absent from Green is 9, Blake Griffin, whose base-tier row reads "9 | Blake Griffin | AU, SN8" -- a run of 8, which is lower than Green's own lowest token of 3 only in the sense that the two ladders are not nested. The difference is therefore a real roster difference and not a truncated fetch: all three tiers were rendered complete against their own headers. cardboardconnection and sportscardchecklist both independently state 68 and 67 for the two tiers.
• G-016 TWO VARIATION ANNOTATIONS RENDER AS A BARE PREFIX WITH AN EMPTY DESCRIPTOR. Errors.cfm on the parent returns 60 rows, 58 of which carry a photo descriptor after the token ("VAR: Shooting", "VAR: Dunking", "VAR: Clippers Not Visible on Jersey", "VAR: Throwback Team"). Two carry the token and nothing after the colon: card 137b OG Anunoby and card 139b Sterling Brown both render the annotation as "VAR:" alone. The rows themselves are real and are counted in Base Rookies Variation; only the descriptor is missing, and it was NOT filled in from the neighbouring rows, which would be derivation. Evidence: https://www.tcdb.com/Errors.cfm/sid/159372/2017-18-Panini-Ascension
• G-017 EVIDENCED ABSENCE, NOT AN UNRESOLVED HOLE: THE VETERAN THROWBACK VARIATION LAYER HAS NO COLOUR PARALLEL. The 40-card Base Rookies Variation layer has a full five-colour ladder, carried in parallels.csv as five set-scope rows, because each colour sid renders its rookie numbers as an a row and a b row with the same serial token. The 20-card Base Veteran Throwback Variations layer has none: all five colour sids render numbers 1-100 as plain single rows with no b row, and each begins its b rows at 101a. That was observed on all five sids independently -- Black, Blue, Green and Red each start Checklist.cfm page 2 at row 101a, which fixes numbers 1-100 at one row each by row arithmetic, and Purple renders no number below 101 at all. sportscardchecklist's 28-entry index likewise lists Variation Black, Blue, Green, Purple and Red but no Veteran Throwback colour entry. Recorded as a gap only because the absence is a fact worth being able to cite, not because anything is unresolved.
• G-018 A DISPUTED LINE IN specs/2017-18_SCOPE_MEMO.md SECTION 7.7 CAN BE CLOSED, AND THIS BUILD FELL OVER THE EVIDENCE INCIDENTALLY. That section records Phase 0's claim that sid 182809 is 2018-19 Hoops, the flawless-bkb-2017-18 build's report that it could not reach 2018-19 Hoops at that sid, and the orchestrator's decision to leave the disagreement undecided. tcdb's 2018-19 season index renders the entry "2018-19 Hoops" with sid 182809 in its own link (https://www.tcdb.com/ViewAll.cfm/sp/Basketball/year/2018), which is Phase 0's reading. That page was fetched here for an unrelated reason (checking whether Ascension has a successor season), so this is one rendering rather than a deliberate re-probe, and the memo's own standard -- re-fetch before deciding -- should still be applied by whoever updates it. Recorded so the observation is not lost.
• G-019 SOURCE-LINEAGE ACCOUNTING FOR THE WHOLE PARALLEL LADDER, stated because source_confidence counts INDEPENDENT SOURCES rather than how sure the agent feels. Six of the seventeen rows are "multi" and eleven are "single". The six are the base-run colour tiers (Black, Green, Red, Blue at 125, Rookies Blue at 129, Purple at 50), where cardboardconnection states each figure in a line returned inside quotation marks on a narrow verbatim re-fetch ("PARALLEL CARDS: Blue #/125, Red #/75, Green #/25, Black 1/1." and "PARALLEL CARDS: Blue #/129, Red #/75, Purple #/50, Green #/25, Black 1/1.") and tcdb's row-level token sweep is the independent second leg. The eleven "single" rows are the four Rookie Ascent Autographs colour tiers, both Autographs tiers and the five Base Rookies Variation tiers: for all of those, tcdb is the ONLY source that states a print run, and its Inserts.cfm Notes cell plus its row sweep are two endpoints of one source rather than two sources. sportscardchecklist supplies counts for them but never runs. TWO FURTHER LIMITS ON THIS RELEASE'S SOURCING. (1) An early summarising fetch of the cardboardconnection page DID report a Rookie Ascent Autographs colour ladder with runs; it was not inside quotation marks, a narrow verbatim re-fetch of the same page did not reproduce it, and it was therefore refused rather than used -- which is the reason those four rows are "single" rather than "multi". (2) This product has only ONE editorial lineage, not the season's usual two: no checklistcenter page for it surfaces through a web search, and beckett 302s to maintenance. So where the 2017-18 scope memo's section 7.6 debate about cardboardconnection/checklistcenter independence applies elsewhere, here there is nothing to weigh -- sportscardchecklist and gogts.net are the off-lineage sources instead, and neither states a print run.

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