2016-17 Totally Certified

• G-001 CHANNEL IS UNKNOWN ON ALL 51 PARALLEL ROWS, DELIBERATELY, AND channel_basis IS EMPTY ON ALL OF THEM. The only channel-adjacent evidence found anywhere for this release is a PRODUCT-level pack configuration line on cardboardconnection, "5 cards per pack, 8 packs per box, 16 boxes per case (2 inners)", at https://www.cardboardconnection.com/2016-17-panini-totally-certified-basketball. That is a configuration, not a trade-channel or exclusivity statement; it names no tier and asserts no exclusivity. Asked directly, that page carries no hobby-only or retail-only claim. Per EXTRACTION_SPEC 2.3 a declined channel is channel=UNKNOWN with an EMPTY basis, because migrations/003_parallels.sql enforces (channel = 'UNKNOWN') = (channel_basis IS NULL); channel_confidence is likewise empty rather than carrying an out-of-vocabulary value (reference/caveats.csv CAV-038). NOTE the sibling season reached the same verdict from a different quote ("Each hobby box averages one autograph and one memorabilia card."), which does not appear on the 2016-17 page, so this is an independent arrival at the same answer and not an inheritance.
• G-002 NO FLOOR IS CLAIMED FOR THIS LINE AND NONE MAY BE INFERRED FROM THE SEASONS PRESENT. specs/2016-17_SCOPE_MEMO.md section 8a decision 3 extends the totallycertified admission from "2017-18 only" to 2016-17 and records that no floor is claimed for either. That memo section 3a measures 2015-16 Totally Certified at tcdb sid 120302 and 2014-15 at sid 99182, so at least two seasons run below this one and this build closes nothing. Per that memo section 3b (the Studio lesson) absence from adjacent season indexes would not be a floor even if it existed; a floor needs a positive statement in words and this build found none and did not look for one.
• G-003 NO base_card_tiers.csv WAS WRITTEN, AND A THIRD SOURCE CARVES THE BASE RUN IN A WAY THAT WOULD SUPPORT ONE. sportscardchecklist.com splits this product's base family into "2016-17 Panini Totally Certified (100 Cards)" plus "2016-17 Panini Totally Certified Rookies (40 Cards)", each with its own seven-colour ladder, at https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-totally-certified/trading-card-checklists-and-product-information. 100 + 40 = 140, which reconciles tcdb's header and my own enumeration exactly, so the three catalogues describe the same cards and the split WOULD tile base_set_size perfectly as tiers 1-100 and 101-140. I did not write it, for three reasons stated so a later pass can overturn them on evidence rather than taste. FIRST, base_print_run is not varies-by-tier: the print run is uniform across the whole run (unnumbered on every one of the 140 rows) and EXTRACTION_SPEC 2.5 makes the tier file required only when the run varies, so a tier file here would carry no print-run information at all. SECOND, the only source that carves the run is a single third-lineage aggregator that reference/caveats.csv CAV-040 documents as contamination-prone, and this build independently caught that same site carrying a foreign product's set on this page (see the Draft Day Ink gap below), so its decomposition is not load-bearing evidence. THIRD, a veterans-then-rookies split is the ordinary shape of most Panini base sets and this catalogue does not tier those, so writing tiers here would make this release the odd one out; CAV-020 (base_card_tiers structure) reserves the abstraction for a repeated roster or a sub-brand anthology, and neither applies - every one of the 140 numbers belongs to a different player. WHAT A LATER PASS MUST CHECK IF IT DISAGREES: check_tier_partition() in migrations/009_triggers_and_views.sql requires the tiers to tile with no gap and counts summing to base_set_size, which 100 + 40 = 140 satisfies; and CAV-043 requires every base_cards.card_tier string to match a base_card_tiers.tier_name EXACTLY, so all 140 card rows would have to be filled in the same change or they would load with tier_id NULL and gate green. base_cards.card_tier is empty on all 140 rows today, which per CAV-022 means not stratified.
• G-004 TWO SETS ARE PROVABLY PER-CARD SERIAL NUMBERED AND THIS SCHEMA CANNOT HOLD THE FIGURE. EXTRACTION_SPEC 2.2 records that sets.csv has no print_run column, and 2.3 records that the parallels table cannot express "serial numbered, run varies per card" at all. (1) Representatives Autographs, sid 142210: every one of the 30 rows renders AU plus a serial token and the DISTINCT TOKEN SET is {SN14, SN25, SN30, SN35, SN50, SN65, SN75, SN100} - eight different runs - enumerated row by row at https://www.tcdb.com/Checklist.cfm/sid/142210/2016-17-Panini-Totally-Certified-Representatives-Autographs and confirmed token for token on a second, differently-varied endpoint at https://www.tcdb.com/ViewSet.cfm/sid/142210/2016-17-Panini-Totally-Certified-Representatives-Autographs. My own count from the enumerated list: SN50 on 9 rows, SN100 on 7, SN35 on 7, SN25 on 2, SN75 on 2, and SN30, SN65 and SN14 on one row each, summing to 30. THE FETCH'S OWN TALLY DISAGREED WITH ITS OWN LIST - it printed the SN100 line twice and its numbers summed to 28 against the 30 rows it had just printed - so the list was used and the tally discarded. (2) Signed, Sealed, Delivered, sid 142227: every one of the 20 rows renders AU plus a serial token and the DISTINCT TOKEN SET is {SN35, SN75, SN99}, enumerated row by row at https://www.tcdb.com/Checklist.cfm/sid/142227/2016-17-Panini-Totally-Certified-Signed-Sealed-Delivered; my count is SN75 on 11 rows, SN35 on 5 and SN99 on 4. (3) Representatives Autographs Mirror, sid 142211, IS carried in parallels.csv and carries print_run=UNKNOWN / serial_numbered=UNKNOWN for the same reason: distinct token set {SN25, SN10, SN5} over all 30 rows, with cardboardconnection agreeing card for card that card 22 is the /5. That row is legal only because it is set-scope; per CAV-041 a base-scope variable tier has no legal row at all. This is the same expressiveness gap CAV-041 records for a sixth consecutive season. MEASURED STRUCTURE, WHICH IS THE USEFUL PART: 3 of 3 autograph tiers whose rows disagree are autograph sets, and 0 of the 48 colour tiers of non-autograph sets vary. That reproduces the generalisation specs/2016-17_SCOPE_MEMO.md section 7.5 states - "variability concentrates in the base tiers of autograph sets" - on a line that had never shown it, and it also reverses this line's own next season, where the 2017-18 build measured 0 of 6 base parallels variable but found its variability in insert and autograph tiers instead.
• G-005 A CHECKABLE DIVERGENCE ON ONE CARD, FLAGGED RATHER THAN MERGED. cardboardconnection prints "PARALLEL CARDS: Mirror #/25 (Card #/9 - #/15)" at https://www.cardboardconnection.com/2016-17-panini-totally-certified-basketball, which agrees with 25 as the tier's run but asserts that card 9 is #/15. Two differently-parameterised tcdb fetches disagree: https://www.tcdb.com/Checklist.cfm/sid/142228/2016-17-Panini-Totally-Certified-Signed-Sealed-Delivered-Mirror renders AU, SN25 on all 20 rows, and https://www.tcdb.com/ViewSet.cfm/sid/142228/2016-17-Panini-Totally-Certified-Signed-Sealed-Delivered-Mirror renders card 9 (Jerry West) as AU, SN25 and reports no row on the page carrying any other token. 25 was written with value_contested=yes. NOTE the two tcdb endpoints are ONE source read twice, so this is not two-against-one on independence grounds; what makes 25 the better claim is that it is a per-row rendering of the actual row in question while the other is an editorial parenthetical whose own punctuation is malformed. A marketplace listing title for card 9 would settle it and none was found.
• G-006 AN EDITORIAL RANGE REFUTED BY A ROW-LEVEL ENUMERATION, RECORDED AS CONTESTED. cardboardconnection prints "PARALLEL CARDS: Blue #/99, Camo #/25 or #/10, Gold #/10, Green #/5, Holo Blue 1/1." at https://www.cardboardconnection.com/2016-17-panini-totally-certified-basketball, i.e. it asserts that some cards of the Camo tier are #/10. tcdb sid 142192 was enumerated ROW BY ROW - all 50 rows listed individually with their player names at https://www.tcdb.com/Checklist.cfm/sid/142192/2016-17-Panini-Totally-Certified-Fabric-of-the-Game-Camo - and the distinct serial-token set is exactly {SN25} with no second token on any row. 25 was written with source_confidence=single (cardboardconnection's hedged figure is a range, and EXTRACTION_SPEC 2.3 is explicit that a ceiling or a range is not a print run) and value_contested=yes. The other four tiers of this set are corroborated by both sources and carry multi. If cardboardconnection is right the tier is per-card variable and the correct row is UNKNOWN/UNKNOWN, so this is the one base figure in the release whose direction could flip on new evidence.
• G-007 NO SOURCE READ IN THIS BUILD STATES A PRINT RUN FOR ANY INSERT OR AUTOGRAPH SET'S OWN BASE TIER, AND ONE MARKETPLACE SNIPPET HINTS THAT AT LEAST ONE OF THEM IS NUMBERED. On tcdb the base tier of every one of the twelve sets renders either an empty flag column (Calling Cards, Energizers, Franchise Foundations, Return to Sender, The Mighty), MEM alone (Fabric of the Game, Fabric of the Game Rookies, Totally Certified Materials) or AU alone (Rookie Autographs, Rookie Roll Call Autographs) - with the two per-card variable autograph sets in the gap above the only exceptions - and cardboardconnection states a print run for none of them. Per CAV-042 an empty flag column is never evidence of absence on its own, and I did not run a second differently-varied fetch on each of those ten base tiers, so nothing is claimed about them in either direction; sets.csv could not carry a figure anyway. ONE LOOSE END WORTH THE NEXT PASS'S TIME: a web-search result title read "2016-17 Jamal Murray Totally Certified Rookie Roll Call Autograph #6 MT 14/99", which would make the Rookie Roll Call Autographs base tier #/99 rather than unnumbered. The title does not say Blue, and the Blue tier of that set IS SN99 on all 49 rows, so the likeliest reading is a seller omitting the colour. CHASE_CHECKLIST_SPEC is explicit that a search-engine result title is not the listing, and the item page was not opened, so this is recorded as an observation and not used.
• G-008 A SET NAMED BY TWO NON-tcdb CATALOGUES WAS REFUSED, AND THE CONTAMINATION WAS PROVED RATHER THAN SUSPECTED. cardboardconnection lists "Draft Day Ink ... 15 cards." among this product's sets at https://www.cardboardconnection.com/2016-17-panini-totally-certified-basketball, and sportscardchecklist lists "2016-17 Panini Totally Certified Draft Day Ink (15 Cards)" at https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-totally-certified/trading-card-checklists-and-product-information. It is NOT a Totally Certified set. THE PROOF: tcdb files a set named exactly "Draft Day Ink" at sid 142363 under the 2016-17 Panini ABSOLUTE parent, sid 138222 - answered directly at https://www.tcdb.com/Inserts.cfm/sid/138222/2016-17-Panini-Absolute - and cardboardconnection's OWN 2016-17 Absolute page describes it in prose, "Draft Day Ink inserts are hard-signed, as well. They feature an image from the day the rookies officially joined their NBA team.", with "15 cards. Serial Numbered #/25." at https://www.cardboardconnection.com/2016-17-panini-absolute-basketball. Same name, same 15-card count, on the same site, under a different product, where it has prose and a print run instead of a bare count. On the Totally Certified side it has no sid anywhere in a fully bounded and fully accounted-for block: 142172-142235 is 64 sids, 142181 does not exist (it renders the tcdb home page), the remaining 63 are exactly the 63 rows Inserts.cfm lists, and both boundaries are foreign products. The entry arithmetic also closes: sportscardchecklist's 73 entries are 16 base-family entries plus 57 others, against tcdb's 8 base-family objects plus 56 others, and 57 minus 56 is exactly this one set. This is reference/caveats.csv CAV-040's shape measured on cardboardconnection rather than sportscardchecklist, and it is also evidence that the two sites are ONE lineage here, since they share the error. NO ROW WAS WRITTEN. If a later pass finds a marketplace title or a Panini sell sheet placing a Draft Day Ink card in Totally Certified packs, add it then.
• G-009 THE VARIATION LAYER IS UNSUPPORTED RATHER THAN DISPROVED. https://www.tcdb.com/Errors.cfm/sid/136421/2016-17-Panini-Totally-Certified returns "No errors or variations have been added for this set yet." Per reference/caveats.csv CAV-020 that silence discriminates nothing - the endpoint returns the identical sentence on sets that provably have errata - so NO conclusion about a variation layer is drawn from it. What the clean base run actually rests on is an independent contiguous enumeration of 1-140 with no hole, no duplicate and nothing past 140. A b-suffixed variation layer would not be surfaced by that enumeration, and Gallery.cfm strips a b suffix, so the two Gallery legs of the enumeration (81-100 and the 101-110 and 131-140 confirmations) could not have shown one even if it exists. Checklist.cfm, which does render suffixes, covered 1-80 and 101-140 and showed none.
• G-010 scarcity_tier IS EMPTY ON ALL 51 PARALLEL ROWS AND ALL 13 sets.csv ROWS, AND stated_odds LIKEWISE. The scarcity pass is not part of this build; per reference/unknown_semantics.csv and CAV-005 an empty scarcity_tier means the pass never covered the row, which is the true state, while UNKNOWN would wrongly assert that it looked and could not place them. No source read in this build states pack odds for any set of this release, which is consistent with CAV-015 (Panini publishes no numeric odds), so stated_odds and both of its evidence columns are empty throughout.
• G-011 THE DERIVED COLUMN parallel_family LANDS ON other FOR 6 OF THE 8 DISTINCT TIER NAMES, AND NO PATTERN WAS ADDED TO pipeline/fix_schema.py. CHANNEL_CLASS, FAMILIES and family_of were lifted out of that module by AST so the repo keeps exactly one definition, per CAV-034 and the round's brief. Results: Holo Blue maps to Holo (14 rows), Camo maps to Camo (9 rows), and Blue, Gold, Green, Orange, Red and Mirror all fall through to other (28 rows) - which is this line's entire visual system bar two. THE SPECIFIC FINDING: this release carries three tiers named Mirror (Calling Cards, Representatives Autographs, and Signed, Sealed, Delivered), FAMILIES has NO Mirror pattern, and a corpus-wide read of every data/*/parallels.csv found ZERO existing rows carrying parallel_family=Mirror. So a Mirror family would be minted on these three rows and nothing else, which is exactly the per-release minting CAV-039 and CAV-021 say must be decided once for the whole line rather than in one release. The 2017-18 build of this same line checked and reported the same absence. REPORTED, NOT ADDED: adding Mirror to the ordered FAMILIES list is an owner decision and needs the corpus-wide dry-run diff CAV-019 (the FAMILIES substring caveat) demands, because Mirror is a substring of nothing currently in the list but the list is matched by plain substring and any new short pattern can swallow a longer sibling.
• G-012 PRIOR-SEASON AND CROSS-PRODUCT SUBSETS UNDER THIS PARENT: NONE, AND THIS IS A REPORTED NEGATIVE RATHER THAN AN ASSUMPTION. https://www.tcdb.com/Inserts.cfm/sid/136421/2016-17-Panini-Totally-Certified returns its whole listing with no truncation and no refusal - "Parallel Sets (7)" and "Insert Sets (56)", asked for one named section per fetch per CAV-047, each section's row count matching its own heading and 63 rows in total by my own count - and not one row carries a season prefix or another product's name. The sid block is fully bounded and fully accounted for: 142171 is "2016 Panini Elite Extra Edition - Dual Materials Holo Silver" (Total Cards: 20, breadcrumb Baseball), 142236 is "1999-00 Chick-Fil-A/FOX 54 Columbus Cottonmouths (CHL)" (Total Cards: 26, Hockey), and the one interior gap 142181 renders the tcdb home page and therefore does not exist. EXTRACTION_SPEC 5.5 requires such rows to be FILED in sets.csv rather than gapped where they exist; here there are none to file. WHAT WAS NOT CHECKED: the 5.5 MIRROR - whether the 2017-18 Totally Certified parent carries a 2016-17 Update block belonging to this release. specs/2016-17_SCOPE_MEMO.md section 9 flags that nobody has looked on any line, and I did not look either; note that the 2017-18 build's own gap on this question reports no season-prefixed row in that parent's listing, which is suggestive but was not written for this purpose. NOTE ALSO a cross-sport boundary hazard worth the next agent's attention: the sid immediately below this block is a 2016 Panini Elite Extra Edition BASEBALL set, and elite is a brands.csv line name - the same shape as the 2016 Panini Encased football set that sits immediately above the 2016-17 Prestige block per specs/2016-17_SCOPE_MEMO.md section 7.2.
• G-013 EXCH AND REDEMPTION ROWS: NONE SEEN, RECORDED AS AN OBSERVATION RATHER THAN A FINDING. Every row-level probe in this build asked for flag tokens BY EXAMPLE and included EXCH in the examples, and EXCH appeared on none of the 108 tcdb pages of this product that were read. That is a token-census observation, not a row-by-row reading of every page, so someone filling insert_cards.csv should re-test it. The 2016-17 memo records EXCH rows on other products of this season, so their absence here is not implausible but is not established either.
• G-014 tcdb TRUNCATION AND REMEDY MEASUREMENTS FROM THIS BUILD, RECORDED FOR reference/caveats.csv CAV-046. Nine 100-row Checklist.cfm page-1 fetches on this product returned, by my own count of their rendered rows: 74, 75, 77, 77, 78, 78, 80, 83 and (the parent) 80. ON EVERY ONE THE ROW COUNT EQUALLED THE NUMBER SPAN, so all nine were clean tail cuts and not the interior-drop shape CAV-046 measured on 2016-17 Optic and Hoops; comparing row count against number span is cheap and was run nine times here. THE ?sort=0 REMEDY MADE THINGS WORSE, reproducing CAV-046's Impeccable finding on a second product: https://www.tcdb.com/Checklist.cfm/sid/136421/2016-17-Panini-Totally-Certified?PageIndex=1&sort=0 cut at card 72 where the default sort reached 80. Appending &x=1 to the Blue tier's page 1 moved its cut from 75 to 79, i.e. a harmless distinct parameter shifted the truncation point slightly rather than clearing it. Gallery.cfm?PageIndex=8/9/10 recovered every missing row on every tier and is what the whole build leaned on.
• G-015 THE SOURCE BASE IS TWO LINEAGES, NOT THREE, AND ONE PART OF THE RELEASE RESTS ON tcdb ALONE. USED: tcdb - 108 page fetches covering ALL 64 sids this product owns (the parent 136421, its seven base parallel sids and all 56 sids Inserts.cfm lists), plus three boundary probes and one fetch of the 2016-17 Absolute parent - and cardboardconnection's product page at https://www.cardboardconnection.com/2016-17-panini-totally-certified-basketball. sportscardchecklist.com at https://www.sportscardchecklist.com/sport-basketball/year-2016-17/index-panini-totally-certified/trading-card-checklists-and-product-information was reached from a search result rather than a constructed path (CAV-037) and agrees with tcdb on all twelve shared card counts, but it also reproduces cardboardconnection's Draft Day Ink error verbatim including the 15-card count, which is a shared-upstream signature, so it is treated as DOWNSTREAM of cardboardconnection and not as an independent third lineage. WHAT MAKES cardboardconnection independent of tcdb, per CAV-037 as corrected (look for a checkable divergence per product): three were found here - the Fabric of the Game Camo range, the Signed Sealed Delivered Mirror card-9 figure, and the Draft Day Ink contamination - so agreement between them elsewhere is worth something and source_confidence=multi is written on 47 of the 51 parallel rows. RESTING ON tcdb ALONE: the Fabric of the Game Camo figure (single), and every set's row count as a contiguous enumeration. NOT REACHED: beckett.com is confirmed unreachable (CAV-033) and was not budgeted for; a site search surfaced a beckett article title for this product but the page was never opened and the title is not cited as evidence. checklistcenter.com has no page for this product that any search surfaced, and per CAV-037 no slug was constructed, so its absence is evidence of nothing. gogts.net surfaced a checklist page for this product in search results; it was not opened, because the 2017-18 build of this same line records that its page carries no counts, print runs or dates at all. Marketplace corroboration was attempted once, by search only, and is discussed in the Rookie Autographs gap.
• G-016 THIS RELEASE CONTRADICTS ITS OWN SIBLING SEASON'S STRUCTURE, AND THAT IS THE EXPECTED OUTCOME. data/totallycertified-bkb-2017-18 has base_set_size 150, base_print_run varies-by-tier and a base_card_tiers.csv splitting 1-100 unnumbered from 101-150 at #/299, established there by a within-sid positive discriminative test. THIS SEASON: base_set_size 140, and the same probe shape finds NO serial token on any of the 140 base rows on two endpoints, so base_print_run is unnumbered, there is no tier boundary and no tier file. The colour ladder also differs - 2016-17 has seven base tiers (Red 199, Blue 99, Orange 60, Camo 25, Gold 10, Green 5, Holo Blue 1) against 2017-18's five plus two Purple block rows, and 2016-17 adds Orange and Red which 2017-18's base set does not have while dropping Purple entirely. Set names differ too: 2016-17 prints "Signed, Sealed, Delivered" where 2017-18 prints "Signed, Sealed, and Delivered", and 2016-17's Calling Cards, Franchise Foundations, Representatives Autographs, Rookie Autographs and Totally Certified Materials have no 2017-18 counterpart at all. Nothing was inherited; the sibling was read for questions only. SEPARATELY, A CONTRADICTION IN EXISTING CATALOGUE DATA, REPORTED AND NOT TOUCHED: data/totallycertified-bkb-2017-18/parallels.csv sets scope_ambiguous=no on its base-set Blue, Camo, Gold, Green and Holo Blue rows while its set-scope rows of the same names set scope_ambiguous=yes and list base-set among the colliding scopes. Those two readings cannot both be right. This release fills the marker consistently in both directions - a name carried by more than one scope is flagged in every row that carries it - which means the two seasons of one line now disagree about a structural marker, and a resolver relying on it will behave differently across them. Not my directory to repair.
• G-017 base_print_run IS unnumbered WITH basis assumed, NOT sourced, AND THE DISTINCTION IS THE POINT. No source read in this build says in words that the base cards carry no serial number. What exists is a positive discriminative test in both directions through one renderer: https://www.tcdb.com/Checklist.cfm on the parent sid 136421 renders an EMPTY flag column on all 140 base rows (1-80 on page 1, and RC alone with no serial token on all 40 rows of page 2), while the same endpoint renders a serial token on all 140 rows of each of the seven base parallel sids, and Gallery.cfm agrees on both sides. cardboardconnection states the parallel ladder with print runs and states no print run for the base cards - which per EXTRACTION_SPEC 2.1 is absence of a figure in a checklist and not a statement of fact. Per that section this is strong indirect evidence and therefore assumed; sourced is reserved for a quotable sentence and there is none. CAV-025 is the reason this matters: base cards are NOT always unnumbered, and this line's own next season numbers 50 of its base cards to 299.
• G-018 rookie_flag WAS WRITTEN yes/no RATHER THAN UNKNOWN, AGAINST THE CAV-003 DEFAULT, AND HERE IS THE BASIS. reference/caveats.csv CAV-003 records that rookie_flag=UNKNOWN is common for Panini because most checklist sources do not print RC designations. tcdb does print it on this parent, and the test available is a within-sid split in BOTH directions: RC appears in the flag column on all 40 rows of 101-140 and on none of the 100 rows of 1-100, on two endpoints (Checklist.cfm?PageIndex=1 and =2, corroborated by Gallery.cfm?PageIndex=9, 10, 11 and 14). The split also survives the temporal check: 101-140 are the 2016 draft class and their first NBA season is this one, while 1-100 are established players. cardboardconnection describes the block the same way - "Rookies treats collectors to the key names of the 2016 draft with lots of additional color." at https://www.cardboardconnection.com/2016-17-panini-totally-certified-basketball. So 101-140 are yes and 1-100 are no. A reader comparing this release with a sibling that wrote UNKNOWN throughout should know the difference is the strength of the test available, not a difference in the products.
• G-019 THE SEVEN BASE PARALLELS WERE FILED AT base-set SCOPE RATHER THAN block SCOPE, AND ONE SOURCE WOULD SUPPORT THE OTHER CHOICE. Each of the seven colour sids states "Total Cards: 140" and renders 140 rows numbered 1-140, i.e. the whole base run at one uniform print run - verified on both halves of the run for all seven (for example Blue renders SN99 on 1-100 and on 101-140 alike). sportscardchecklist decomposes each colour into a 100-card entry and a 40-card Rookies entry, which would suggest two block rows per colour. I did not do that, for two reasons. FIRST, the print runs are identical across both halves, so two block rows per colour would carry identical values and add nothing. SECOND, EXTRACTION_SPEC 2.3 and its 2026-09-17 correction record that a block row's scope_set_id is NULL unconditionally and that parallel_unique_in_scope keys on (release_id, parallel_name, COALESCE(scope_set_id,'')), so every block row in a release shares one key group and two block rows may never share a name: a Base Blue and a Rookies Blue would have to be invented as disambiguated names that no source prints. base-set scope resolves to a real id and is the more accurate description of what the pages render.

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