2018-19 Spectra

• G-001 THE WEAKEST CLAIM IN THIS RELEASE. Silver and White Sparkle are written print_run=unnumbered / serial_numbered=no, and the database forbids UNKNOWN on a base-set-scope parallel, so a value had to be chosen. The usual positive discriminative test is DISARMED on this product and this was measured, not assumed: the Nebula sid (https://www.tcdb.com/ViewSet.cfm/sid/204322/2018-19-Panini-Spectra-Nebula), which both editorial sources state is 1/1, renders an EMPTY flag column on its rows 1, 2, 3, 10 and 50 while rendering "AU, MEM, SN1" on its rows 101, 102, 108 and 135. tcdb has simply not entered serial data for the base-scope rows of most parallel sids, so a blank flag column on Silver or White Sparkle proves nothing. The value therefore rests entirely on the editorial split annotation: cardboardconnection and checklistcenter each list all eleven base tiers and annotate nine with a print run while leaving Silver and White Sparkle unannotated (White Sparkle carrying the marker SP for short print in both). Those two sources may be one opinion - they agree almost verbatim across this whole release. A later pass with a marketplace listing of a Silver or White Sparkle card would settle it in one fetch.
• G-002 base_print_run=175 with basis=sourced REVERSES the sibling season. spectra-bkb-2019-20/release.csv carries base_print_run=unnumbered with base_print_run_basis=assumed. This is not a correction of that release - 2019-20 Spectra genuinely appears to have unnumbered base cards - it is a structural change between seasons that a consumer comparing the line will see. Evidence for 2018-19: cardboardconnection "100 cards. Serial numbered #/175." (https://www.cardboardconnection.com/2018-19-panini-spectra-basketball-cards), checklistcenter "100 Cards - Serial Numbered #/175" (https://www.checklistcenter.com/2018-19-panini-spectra-basketball/), and tcdb rendering the literal token "SN175" on all 100 base rows (https://www.tcdb.com/Checklist.cfm/sid/199763/2018-19-Panini-Spectra?PageIndex=1 rows 1-77 and the Gallery pages 8-10 for rows 71-100). Three sources, two of them independent of each other, plus row-level structural evidence.
• G-003 The parent checklist states "Total Cards: 135" and renders cards 101-135 as a 35-card block flagged "AU, MEM, RC, SN299". Per EXTRACTION_SPEC section 5.3 the block is EXCLUDED and base_set_size=100. The block is the Rookie Jersey Autographs set, which has no tcdb sid of its own and is recorded as a set_type=auto row in sets.csv. Unlike the 2019-20 trap (where cardboardconnection and checklistcenter both printed 175 for a 150-card base set), here BOTH editorial sources already state the base set as 100 cards and name the 101-135 block separately, so no editorial source had to be contradicted. Recorded because the catalogue is documented as self-contradictory on this convention (EXTRACTION_SPEC section 5.3: contenders 2019-20 includes its block at 165, contenders 2020-21 excludes at 100). Checked against the sibling: spectra-bkb-2019-20 also EXCLUDES its over-run block, so the Spectra line is internally consistent.
• G-004 rookie_flag=yes on 14 cards (6, 16, 18, 26, 36, 38, 46, 56, 66, 76, 83, 86, 93, 96) where tcdb renders the literal token "RC"; UNKNOWN on the other 86. UNKNOWN rather than no, because tcdb's RC tagging on this set is demonstrably INCOMPLETE: card 8 Kevin Knox (2018 draft, 9th overall), card 28 Miles Bridges (2018 draft, 12th overall) and card 90 Allonzo Trier (2018-19 rookie) are all rookies and carry no RC token. The RC tokens were confirmed stable across two differently-varied fetches (https://www.tcdb.com/Checklist.cfm/sid/199763/2018-19-Panini-Spectra?PageIndex=1 and https://www.tcdb.com/Gallery.cfm/sid/199763/2018-19-Panini-Spectra?PageIndex=1&sort=0 - Doncic RC on both, Knox no RC on both), so this is tcdb data, not fetch loss. Per caveats CAV-003 and the spec, absence of an RC designation is not a statement that a player is not a rookie. No source consulted prints a complete RC list for this release.
• G-005 CONTESTED count, 29 against 30. tcdb's base page https://www.tcdb.com/ViewSet.cfm/sid/204313/2018-19-Panini-Spectra-Award-Winning-Autographs states "Total Cards: 29" and rendered 29 rows. cardboardconnection states "Award-Winning Autographs: 30 cards" and checklistcenter "30 Cards"; four of this set's own tcdb parallel pages (Gold 204314, Nebula 204315, Neon Orange 204318, Neon Pink 204319) each state "Total Cards: 30". 30 written, on the reasoning that the base page is short by one and the parallel pages recover it (CHASE_CHECKLIST_SPEC 1c-ter). The missing player was NOT identified; a chase pass should union the parallel pages against the base page to name it.
• G-006 CONTESTED count, 27 against 28. tcdb's base page https://www.tcdb.com/ViewSet.cfm/sid/204372/2018-19-Panini-Spectra-Making-it-Rain-Autographs states "Total Cards: 27" and rendered 27 rows; cardboardconnection ("28 cards"), checklistcenter ("28 Cards") and this set's Gold (204373), Nebula (204374), Neon Orange (204377) and Neon Pink (204378) tcdb pages all say 28. 28 written on the same reasoning as G-005. Missing player not identified.
• G-007 CONTESTED count and the weakest card_count in this release. cardboardconnection states "Epic Legends Relic: 30 cards" and is the ONLY source above 29. tcdb's base page https://www.tcdb.com/ViewSet.cfm/sid/204299/2018-19-Panini-Spectra-Epic-Legends states "Total Cards: 28" (24 rows rendered, truncated) and its five parallel pages state 27, 28, 28, 29 and 27 - so the highest figure tcdb carries anywhere for this set is 29. checklistcenter does not list this set at all in the material fetched. 30 written because it is the only editorial count available and tcdb's figures disagree with each other, but a reader should treat 30 as single-sourced and 28-29 as the structural reading. This is a genuine unresolved disagreement, not a transcription problem.
• G-008 Three set names differ between sources and the tcdb form was written. cardboardconnection calls two relic sets "Epic Legends Relic" and "Next Era Relic"; tcdb renders them "Epic Legends" and "Next Era" (per the page headings on sids 204299 and 204397, and per the Inserts.cfm listing https://www.tcdb.com/Inserts.cfm/sid/199763/2018-19-Panini-Spectra). cardboardconnection also writes "Award-Winning Autographs" with a hyphen where tcdb renders "Award Winning Autographs" without one, and checklistcenter abbreviates several ("Icons Auto", "In the Zone Auto", "Award Winning Autos", "Rising Star Signatures"). The tcdb form was chosen because it is the form the card-level pages carry and therefore the form a later insert_cards.csv pass must match byte for byte. No set was renamed away from a source; all variants are recorded here.
• G-009 Ten set-scope parallel tiers render MORE THAN ONE serial token across their own rows, so they have a per-card variable run and not a tier print run. Each is written print_run=UNKNOWN / serial_numbered=UNKNOWN - provably serial numbered, run not expressible - with both tokens preserved verbatim in visual_notes. The schema cannot hold "serial numbered, run varies", which EXTRACTION_SPEC section 2.3 names as a known modelling gap. The eleven: Award Winning Autographs Neon Pink (SN25 / SN15 Allen Iverson); Epic Legends Neon Green (SN25 / SN19 Grant Hill); Icons Autographs Neon Pink (SN25 / SN15 John Stockton, Oscar Robertson); Illustrious Legends Signatures Neon Green (SN49 / SN35 Kevin McHale, Tracy McGrady); Illustrious Legends Signatures Neon Pink (SN25 / SN15); In The Zone Autographs Neon Green (SN49 / SN35 on six players); In The Zone Autographs Neon Pink (SN25 / SN15 on four players); Making it Rain Autographs Neon Green (SN49 / SN35 Ray Allen); Making it Rain Autographs Neon Pink (SN25 / SN15 Damian Lillard, John Stockton); Spectacular Swatches Neon Green (SN25 / SN20 John Wall). The same phenomenon also occurs at SET level rather than tier level - the BASE tiers of Award Winning Autographs, Illustrious Legends Signatures, In The Zone Autographs and Radiant Signatures each render SN75 on most rows with SN25/SN49 exceptions, which is why checklistcenter's flat "Serial Numbered #/75" for all eight autograph sets is a modal value and not a fact about every card.
• G-010 All 69 set-scope parallel rows draw their print run from tcdb flag tokens and nothing else - no editorial source enumerates the insert-set parallel ladders for this release (cardboardconnection and checklistcenter list only the eleven BASE parallels). Two limits on that evidence. First, for the 63 tiers with their own sid the token was read through one summarizing fetch per sid that was asked whether every rendered row carried the same token and to name any row that differed; that is the fetch's own scan, not an independent count, and eleven of the 63 did report a differing row (G-009), which shows the question was answered rather than rubber-stamped but does not prove the other 52 are uniform. Second, the six Rookie Jersey Autographs tiers have no sid of their own and were read off rows 101-135 of the base parallel sids: Gold was read on all 35 rows, the other five on rows 101, 102, 108 and 135 only. The Rookie Jersey Autographs ladder IS independently corroborated - breakninja lists it as "Neon Blue /99, Neon Green /49, Neon Pink /25, Gold /10, Neon Orange /5, Nebula /1" (https://www.breakninja.com/basketball/18-19-Panini-Spectra-Basketball-Checklist.php), matching the tcdb tokens six for six.
• G-011 The name "Neon Blue" carries THREE different print runs inside this one release and no source states the rule. Base set: /75 (both editorial sources, plus the token "SN75" on card 5 of https://www.tcdb.com/ViewSet.cfm/sid/204323/2018-19-Panini-Spectra-Neon-Blue). Rookie Jersey Autographs: /99 (token "AU, MEM, SN99" on rows 101-135 of that same sid). The autograph insert sets: /60. The relic insert sets (Epic Legends, Next Era, Spectacular Swatches): /49. Recorded because a consumer grouping by parallel_name without filtering scope would collapse four different objects - exactly the failure caveats CAV-006 and CAV-026 describe. The same shape holds for Neon Green (base /49, relics /25) and Neon Pink (base /25, present on nine sets, absent from the three relic sets).
• G-012 Flash and Wave carry channel=FOTL with channel_basis=sourced-exclusive, quoting checklistcenter's tier names "Flash #/9 (First Off the Line)" and "Wave #/9 (First Off the Line)" and corroborated by cardboardconnection's FOTL box contents line "1 Base Wave or Flash Parallel". Every other parallel row carries channel=hobby with channel_basis=sourced-presence, evidenced by the Steel City Collectibles product title "2018-19 Panini Spectra Hobby Basketball Box" (https://www.steelcitycollectibles.com/i/2018-19-panini-spectra-hobby-basketball-box) and cardboardconnection's hobby box configuration - a source PLACING the product in hobby, not a source claiming any tier is hobby-exclusive. This diverges from the sibling: spectra-bkb-2019-20 writes channel=hobby with channel_basis=unsupported on seven of its eight base parallels and from-note on the eighth, and leaves channel_basis EMPTY with channel=UNKNOWN on all 52 of its set-scope rows. Same product line, same evidence shape, different answer; flagged rather than smoothed over.
• G-013 No numeric pull-odds sheet was found for this product (cardboardconnection, checklistcenter, breakninja, steelcitycollectibles and tcdb all checked). Consistent with caveats CAV-015: Panini publishes no numeric odds for basketball. Every stated_odds field is blank in sets.csv and parallels.csv, meaning unpublished rather than unresearched. cardboardconnection does state box composition - "8 cards per pack, 1 pack per box, 8 boxes per case" with "4 Autographs" and "1 Relic" per hobby box - which is a configuration, not odds.
• G-014 scarcity_tier left empty on every sets.csv and parallels.csv row. No editorial or manufacturer source uses short-print, SSP or case-hit language for any set or tier in this release; the only scarcity language found is the bare "SP" marker both editorial sources attach to White Sparkle, and White Sparkle is a base parallel rather than an insert set, so the sets.csv enum does not apply to it. Per caveats CAV-005, empty here means the scarcity pass has not covered these rows, not that they were looked at and could not be placed. Inferring a tier from a print run alone is explicitly forbidden, so nothing was written.
• G-015 No base variation or photo variation set is recorded for this release, and that is an ABSENCE OF EVIDENCE rather than evidence of absence. https://www.tcdb.com/Errors.cfm/sid/199763/2018-19-Panini-Spectra returns "No errors or variations have been added for this set yet." - which reference/caveats.csv CAV-020 establishes is uninformative, since the same sentence is returned for sets that provably carry UER rows. What DOES survive here is independent: the parent checklist enumerates 135 contiguous distinct card numbers (1-100 base, 101-135 Rookie Jersey Autographs) with no duplicate or b-suffixed number seen on any fetch, and "Total Cards: 135" reconciles exactly with that enumeration. A b-suffixed variation block would not necessarily surface in that enumeration, so the claim that this release has no variation set is unsupported, not disproved. The sibling season DOES have one - spectra-bkb-2019-20 carries a 25-card Base Variations insert - which makes the question live rather than academic.
• G-017 Recorded so a later pass does not attempt the blank-annotation test here. https://www.tcdb.com/Inserts.cfm/sid/199763/2018-19-Panini-Spectra DOES render a per-set Notes column, and on this parent every one of the 86 rows carries an image-completion percentage ("48.5% of images added" and the like) and NOTHING else. There are no blank cells and no serial statements, so the column carries zero information about any tier and the split test EXTRACTION_SPEC describes for the Prizm parent cannot be run at all. This is the Contenders / Hoops Winter shape, making at least a seventh measured behaviour for that column. All print runs in this release came from row-level flag tokens or from editorial lists instead.
• G-018 Two different sets in this release use the SAME card-number prefix. https://www.tcdb.com/ViewSet.cfm/sid/204403/2018-19-Panini-Spectra-Radiant-Signatures renders RS-AJN (Avery Johnson), RS-BWT (Bill Walton), RS-DLL (Damian Lillard); https://www.tcdb.com/ViewSet.cfm/sid/204410/2018-19-Panini-Spectra-Rising-Stars-Signatures renders RS-AHL (Aaron Holiday), RS-ASM (Anfernee Simons), RS-ATR (Allonzo Trier). Rosters are disjoint (legends against 2018-19 rookies) so no code collision was observed, but a later insert_cards.csv pass must key on (set_name, card_number) and must not search by prefix. Note also that the 2018-19 scope memo records RS- as 2018-19 PRIZM's Rookie Signatures prefix, so the token is reused across products in this season as well.
• G-019 channel_class and parallel_family are empty on all 80 parallel rows. The round-7 agent brief states they are derived and that pipeline/fix_schema.py populates them, and instructs agents to leave them empty; EXTRACTION_SPEC section 6 separately asserts that channel_class must equal the value the channel map gives. The two are only consistent if fix_schema.py runs before that check. Recorded so that if the gate fails on channel_class for this release, the cause is a known instruction conflict and not missing research. The values the map would give are: hobby -> box-type on 78 rows, FOTL -> promotion on the Flash and Wave rows.
• G-020 team_or_school is written exactly as tcdb renders it, which is the team on the card rather than the player's team at any particular date. Two rows are worth naming: card 11 Jeremy Lin is rendered "Toronto Raptors" (he began 2018-19 in Atlanta), and card 93 Landry Shamet is rendered "Los Angeles Clippers" while the Rookie Jersey Autographs row for the same player at card 116 is rendered "Philadelphia 76ers". Both are recorded as rendered, not reconciled. Teams for cards 1-77 came from https://www.tcdb.com/Checklist.cfm/sid/199763/2018-19-Panini-Spectra?PageIndex=1 and for cards 71-100 from Gallery pages 8, 9 and 10 on the parent sid; the 71-77 overlap agreed exactly, which also confirms that Gallery.cfm DOES render team on this release.

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