2019-20 Certified

• G-001 base_print_run recorded as unnumbered/assumed. No source STATES that the 2019-20 Certified base cards are unserialised. The positive evidence available is indirect: tcdb prints a per-row serial flag (SN99, SN25, SN10, SN5, SN1) on every Mirror parallel checklist and prints nothing on the base checklist, and cardboardconnection and checklistcenter both begin the parallel run at "Orange #/99" with Blue and Red carrying no figure. That is a rule I applied, not a statement I found, so the basis is `assumed` rather than `sourced` per EXTRACTION_SPEC 0.
• G-002 Certified Signatures card count disagreement: tcdb states Total Cards "39" and renders 39 rows (https://www.tcdb.com/ViewSet.cfm/sid/218477/Certified-Signatures); checklistcenter renders a contiguous 1-40 checklist (counted row by row, ending 40 Doug Collins) and cardboardconnection states 40. sets.csv uses 40. The missing tcdb row was not identified.
• G-003 Fabric of the Game Signatures - the 20-card checklist is named differently by all three sources and no source shows an unqualified base version. tcdb: "Fabric of the Game Signatures Jersey Number" (and catalogues Black/Camo/Gold as its parallels, with NO plain row). checklistcenter: "Fabric of the Game Signature Jersey Number Set". cardboardconnection: "Fabric of the Game Signatures Red Checklist". sets.csv carries the family name common to all three, and both contested version names are carried as parallels with value_contested=yes, so no side is silently picked. Unresolved: whether the 20-card checklist IS the Jersey Number version, IS the Red version, or is a plain version that all three sources omit.
• G-004 Fabric of the Game Signatures Red print run 99 rests on cardboardconnection's section heading plus a single marketplace lot title ("Mo Bamba 2019-20 Panini Certified Fabric Of The Game Red Jersey Auto Number d 95/99", hollywoodmemorabilia.com). tcdb does not catalogue a Red tier for this set. Single-source for existence; the 99 comes from one observed serial, not from a published print run.
• G-005 Fabric of the Game Signatures Jersey Number has a per-card variable print run - tcdb annotates its rows SN15, SN99, SN25, SN49, i.e. each card is numbered to that player's own jersey number. parallels.csv's print_run column takes an integer or `unnumbered` and cannot express this, so it is UNKNOWN/UNKNOWN. The card IS serial numbered; that fact is recorded here because the schema has nowhere to put it.
• G-006 Ceiling print runs not resolved. cardboardconnection and checklistcenter publish the autograph-set parallel tiers as ceilings or alternatives - "Camo #/25 or less", "Camo #/25 or #/10", "Gold #/10 or #/5", "Green #/5 or #/3", "Red #/99 or less", "Blue #/49 or #/15" - across Ballot Busters, Certified Signatures, Established Autographs, Fabric of the Game Signatures, Fresh Faces Signatures, Freshman Fabric Signatures and Legendary Signatures. A ceiling is not a print run, so all 19 such tiers are UNKNOWN/UNKNOWN rather than the ceiling figure. Note this DIVERGES from earlier catalogue practice, where several releases wrote the ceiling as an integer with a gap (e.g. contenders Ticket Stub 77, elite Aspirations 99); the rewritten EXTRACTION_SPEC 2.3 forbids it. Only Rookie Roll Call's tiers are published as exact figures and those are written as integers.
• G-007 Freshman Fabric Signatures Green is single-source: tcdb Inserts.cfm catalogues a parallel set "Freshman Fabric Signatures Green", but cardboardconnection and checklistcenter both list this set's tiers as Red/Blue/Camo/Gold/Black with no Green. Recorded with print run UNKNOWN.
• G-008 tcdb's base Mirror Camo entry (sid 219139) states Total Cards "30" and renders only cards 1-30, while Mirror Blue, Red, Orange, Gold, Green and Black each state 200. Both aggregators list Camo as a full base parallel at #/25 with no restriction. Read as an incomplete tcdb entry rather than a 30-card parallel, but the full 200-card extent of Mirror Camo is not positively confirmed by any source.
• G-010 rookie_flag rests on a single source. tcdb's Gallery.cfm renders an (RC) token on card 73 and on cards 151-200 and nothing on the other 149; no other source consulted publishes RC designations for this base set. Written yes/no on tcdb's display alone.
• G-011 Card 73 Talen Horton-Tucker is the only (RC) inside the 1-150 veteran run - he sits in the Lakers team group at 71-73 rather than in the 151-200 rookie block, and he does not appear a second time in 151-200. Verified against tcdb Gallery page 8 and against the full 151-200 roster; unusual enough to flag but not contradicted by anything.
• G-012 No pack odds were found for any set in this release. cardboardconnection publishes only a box-level average ("1 Freshman Fabric Signature and 1 Additional Autograph" per hobby box), which is not a per-set stated odds figure, so stated_odds/odds_evidence_url/odds_evidence_quote are empty on every sets.csv row and scarcity_tier is unset.
• G-013 beckett.com could not be consulted: https://www.beckett.com/news/2019-20-panini-certified-basketball-cards/ 302-redirects to https://beckett-maintenance-page.s3.amazonaws.com/index.html . Beckett is the usual third structural source for this season's releases, so cross-checking depth on set counts and parallel print runs is one source shallower than normal. Note also that cardboardconnection and checklistcenter agree almost verbatim on the parallel lines and may be one opinion rather than two.
• G-014 Gallery.cfm validates its slug, unlike ViewSet.cfm. Gallery.cfm/sid/213354/Panini-Certified?PageIndex=8, /Certified-Basketball?PageIndex=9 and /Certified-2019-20?PageIndex=10 each returned "The page you attempted to view does not exist"; the same PageIndex values under the exact slug 2019-20-Panini-Certified returned the correct 10 tiles. The spec's advice to re-fetch a 503 under a different valid slug therefore does not transfer to Gallery.cfm.

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