2018-19 Chronicles

• G-001 SUB-BRAND BLOCK TILING -- how it was established, and what rests on one source. The 21-block tiling of the 699-card base run is carried by TWO independent sources that agree on every boundary. (1) sportscardchecklist publishes one page per block, each stating its own count and range in words, e.g. https://www.sportscardchecklist.com/set-173319/2018-19-panini-chronicles-base-obsidian-preview-basketball-trading-card-checklist renders '15 Cards' numbered '#571-585' and https://www.sportscardchecklist.com/set-173255/2018-19-panini-chronicles-base-classics-basketball-trading-card-checklist renders '33 Cards' spanning 634 through 666. (2) tcdb renders the block name as a per-row tag in the parent checklist and as a tile-title suffix in the parent gallery: https://www.tcdb.com/Checklist.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=3&sort=0 renders 'RC, Essentials' on 201-235 then 'RC, Marquee' on 236-270 then 'RC, Elite' on 271-279, and https://www.tcdb.com/Gallery.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=51&sort=0 renders '501 - Rodions Kurucs - Brooklyn Nets Chronicles Rookies'. Every boundary where both sources render is identical. TWO BLOCKS ARE WEAKER. sportscardchecklist has no page for Chronicles 1-100 or Chronicles Rookies 501-540; it combines them into one 140-card set at https://www.sportscardchecklist.com/set-173246/2018-19-panini-chronicles-basketball-trading-card-checklist which renders 1-100 continuous and then nothing until 501. The split into two tiers rests on tcdb's labelling (gallery tiles 1-100 carry NO sub-brand suffix while 501-540 all carry 'Chronicles Rookies') plus checklistcenter's separate 'Base Chronicles Set' and 'Base Chronicles Rookies Set' headings. The NAME of the 1-100 block is the weakest single value in the tiling: no source prints 'Chronicles' as a block label beside those cards, it is inferred from the absence of any other label plus checklistcenter's heading, and eBay lot titles render mid-run cards with a block qualifier ('2018-19 Panini Chronicles - Score Luka Doncic #681') but render #71 Doncic with none. TCDB COVERAGE IS PARTIAL: the parent Checklist pages truncate at rows 71, 170, 279, 384, 483, 580 and 672 of their 100-row pages, so cards 171-200, 280-300, 385-400, 484-500, 581-600 and 673-699 were read from sportscardchecklist only and are single-sourced for player and team, though their block boundaries are still double-sourced because the adjacent blocks' ranges pin them.
• G-002 RELEASE DATE rests on two RETAIL sources and no editorial one. https://www.steelcitycollectibles.com/i/2018-19-panini-chronicles-basketball-hobby-box renders 'Release Date: 6/28/19' and https://stockx.com/2018-19-panini-chronicles-basketball-blaster-box renders '06/28/2019'. The usual editorial source, cardboardconnection, REFUSED this environment -- a fetch of https://www.cardboardconnection.com/2018-19-panini-chronicles-basketball-nba-cards returned a robots.txt fetch failure, consistent with the scope memo's note that some hosts refuse here. Beckett was not attempted because seven prior passes found it unreachable. The 2018-19 scope memo section 5 already records that every release date for this season is single-sourced to cardboardconnection; this release does not inherit that source at all, so its date is independent of the memo's and should be re-checked against an editorial source when one becomes reachable. Two retailers can also be one opinion if both took the date from the same distributor sheet.
• G-003 THE OBSIDIAN NAME SHIPS INSIDE THIS PRODUCT AND IS NOT THE OBSIDIAN LINE SHIPPING. Cards 571-585 are a 15-card block titled '2018-19 Panini Chronicles Base Obsidian Preview' at https://www.sportscardchecklist.com/set-173319/2018-19-panini-chronicles-base-obsidian-preview-basketball-trading-card-checklist, with its own eight-colour ladder recorded as block-scope rows in parallels.csv. brands.csv was NOT touched and no row was minted: the standalone 2018-19 Panini Obsidian release is a separate product already built at data/obsidian-bkb-2018-19, whose own release notes independently record this same subset and reach the same conclusion. The two agree exactly on the range 571-585 and the count 15, which is a cross-release corroboration rather than a conflict. STRUCTURAL HAZARD, REPRODUCED: https://www.tcdb.com/Inserts.cfm/sid/205013/2018-19-Panini-Chronicles renders NO set whose name contains 'Obsidian' -- its Insert Sets (40) and Parallel Sets (13) headings list none -- because tcdb files Chronicles sub-brands as ranges of the parent base run rather than as named sets. An agent enumerating this parent on tcdb alone would conclude the block does not exist. It is visible on tcdb only as a per-row tag inside the parent checklist (https://www.tcdb.com/Checklist.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=6&sort=0 renders 'RC, Obsidian Preview' on 571 and 'Obsidian Preview' on 572-580) and as the card range of the release-wide colour parallel sets. CONTRADICTS A SECOND SOURCE: checklistcenter renders 'Base Obsidian Preview Rookies Set' with '30 cards' at https://www.checklistcenter.com/2018-19-panini-chronicles-basketball-card-checklist/, which is wrong -- two independent sources render 15 cards over 571-585, and 30 would not tile.
• G-004 THE TITANIUM BLOCK 436-470 HAS THREE COMPETING SERIAL SCHEMES AND SOURCES DISAGREE ON WHICH IS THE BASE NUMBERING, so the tier's print_run is written UNKNOWN rather than resolved. tcdb files 'Titanium Draft Position' (sid 544922) and 'Titanium Draft Year' (sid 207695) under the Parallel Sets heading at https://www.tcdb.com/Inserts.cfm/sid/205013/2018-19-Panini-Chronicles, implying the third scheme is the base -- and indeed https://www.tcdb.com/Checklist.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=5&sort=0 renders the JERSEY-NUMBER values on the parent's own base rows (436 SN5, 437 SN4, 439 SN13, 440 SN22, 450 SN77, and Luka Doncic wore 77). sportscardchecklist takes the opposite view, titling the base page 'Base Titanium Draft Position' and giving Draft Year and Jersey Number their own pages. All three are recorded as block-scope parallels with value_contested=yes on the two whose runs vary per card. Draft Position is numbered to the draft slot (SN1 Ayton, SN3 Doncic, SN60 Trier) and Jersey Number to the shirt number; only Draft Year has a fixed run, SN18 uniformly. This is the schema expressiveness gap the extraction spec section 6 names: both variable tiers are PROVABLY serial numbered with a per-card run, and the forced pairing of print_run and serial_numbered collapses that to UNKNOWN/UNKNOWN, which reads identically to 'we found nothing'. We found a great deal; the CSV cannot carry it.
• G-005 BASE BRONZE CARRIES A SOURCED PACK FORMAT BUT NO SOURCED TRADE CHANNEL. https://www.tcdb.com/Inserts.cfm/sid/205013/2018-19-Panini-Chronicles renders 'Inserted in 15 card Value Packs.' in the Notes cell for the Bronze parallel -- the ONLY non-blank Notes cell in the entire 13-row Parallel Sets table. Value Packs are a pack format, not a trade channel, so the string is recorded in channel_sku and channel is left UNKNOWN with an empty channel_basis. Writing channel=retail would be an inference from the SKU, which is the exact substitution the extraction spec section 2.3 forbids, and 'unsupported' is rejected by the database alongside channel=UNKNOWN. A channel pass with a source that states a trade channel in words could promote this row. NOTE A DIVERGENCE FROM THE SIBLING SEASON: data/chronicles-bkb-2019-20 writes channel=retail with channel_basis=unsupported on its own Base Bronze row, and channel=blaster with channel_basis=from-note on Base Pink. Those bases mean no citation exists, so the sibling asserts a channel this build declines to assert on comparable evidence. Same line, adjacent seasons, opposite call. [REVIEWED 2026-08-19: still open; 24 of 24 parallels lack a sourced channel.]
• G-006 ROOKIE FLAGS ARE UNKNOWN ON MOST OF THE RUN AND THE REASON IS A MEASURED CONTRADICTION, NOT MISSING WORK. Policy applied: rookie_flag=yes only where a fetched rendering showed an RC token on that row. Cards 1-100 carry yes/no because the tcdb Checklist and Gallery enumerations rendered an explicit per-row flag cell for them (https://www.tcdb.com/Checklist.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=1&sort=0 and https://www.tcdb.com/Gallery.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=8&sort=0). For 101-699 the same column is occupied by the sub-brand tag and the RC reporting is demonstrably unreliable: https://www.tcdb.com/Gallery.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=51&sort=0 reports RC on card 501 while https://www.tcdb.com/Checklist.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=6&sort=0 reports no RC on the same card, so absence of RC in one rendering is not evidence of absence on the page. Writing 'no' off that absence would have been an assertion the evidence does not support, so UNKNOWN is written instead. A further oddity that argues for caution: tcdb renders RC on all of 101-135 but on none of 136-170, although the two blocks share nearly the same 2018 draft-class roster. Ranges never rendered by tcdb at all (171-200, 280-300, 385-400, 484-500, 581-600, 673-699) are UNKNOWN for the same reason plus no second source -- sportscardchecklist shows no RC designation on any row of any block page. Cheap remedy for a later pass: the parent Gallery at 10 tiles per page covers the missing ranges at PageIndex 18-20, 29-30, 39-40, 49-50, 59-60 and 68-70.
• G-008 INSERT-SCOPE PARALLELS ARE ONLY PARTIALLY WRITTEN AND THE REMAINDER IS DELIBERATELY OMITTED RATHER THAN GUESSED. https://www.tcdb.com/Inserts.cfm/sid/205013/2018-19-Panini-Chronicles renders 'Insert Sets (40)', which is nine autograph parent sets plus 31 colour and prime variants. Four of those variants are written here, all four for Chronicles Signatures, each with a within-sid serial probe: Blue sid 207697 SN49, Gold sid 207698 SN10, Platinum sid 207699 SN1, Red sid 207701 no serial token on any row. The other 27 were NOT written because no print run was sourced for them and the extraction spec section 6 says to remove the row and carry the finding in a gap rather than write UNKNOWN or guess a ladder. Naming them so a later pass is cheap, with tcdb sids: Gold Standard Rookie Jersey Autographs Prime 207703, Red 207704, Super Prime 207705; Momentous Rookie Autos Gold 207707, Platinum 207708, Red 207709; Origins Rookie Autographs Gold 207711, Platinum 207712, Red 207713; Rookie Ascent Gold 207715, Platinum 207716, Red 207717; Rookie Chronicles Signatures Blue 207719, Gold 207720, Platinum 207721, Purple 207722, Red 207723; Timeless Ink Gold 207725, Platinum 207726, Red 207727; V-Team Signature Swatches Prime 207729, Red 207730, Tag 207731; XR Rookie Jumbo Swatch Autographs Prime 207733, Red 207734, Super Prime 207735; and Chronicles Signatures Purple 207700. Each costs one ViewSet fetch. DO NOT infer the ladder from Chronicles Signatures: its Red is UNNUMBERED while its Blue, Gold and Platinum are numbered, so the ladder is not even internally uniform within one set.
• G-009 SPEC DEFECT FOUND, AND IT REVERSES THE ADVICE THIS BUILD WAS GIVEN ABOUT BLOCK SCOPE. The task briefing and the extraction spec section 2.3 both state that a block-scope parallel gets scope_set_id NULL when a release has NO base_card_tiers.csv, and that writing real tiers gives block rows something to resolve against. The schema says otherwise: migrations/003_parallels.sql declares CONSTRAINT scope_shape CHECK (scope_type IN ('block','UNKNOWN') AND scope_set_id IS NULL), so EVERY block row has scope_set_id NULL unconditionally, whether or not tiers exist. The block mapping lives entirely in the separate parallel_block_scope junction table, which the unique index does not touch. Therefore parallel_unique_in_scope on (release_id, parallel_name, COALESCE(scope_set_id,'')) puts ALL block rows of a release in ONE key group, and every block row in a release must have a globally unique name among block rows EVEN WHEN TIERS ARE WRITTEN. Writing tiers does not protect against the collision that failed contenders-bkb-2018-19; only unique naming does. This build was designed around the schema rather than the briefing: its 12 block rows carry 12 distinct names. The natural naming would have collided twice -- an 'Orange' for the three /249 blocks and an 'Orange' for Obsidian Preview, and likewise for Red, Green, Pink, Blue and Purple -- so the Obsidian Preview ladder is named 'Obsidian Preview Orange' and so on, which is also how sportscardchecklist and Beckett print those tiers, so nothing was renamed away from a source. The spec sentence should be corrected: the trap is not 'block scope without tiers', it is 'two block rows sharing a name'.
• G-010 A PRINT RUN THAT IS TRUE FOR MOST OF A PARALLEL'S COVERAGE AND FALSE FOR PART OF IT CANNOT BE EXPRESSED, AND THIS RELEASE HAS FOUR CASES. tcdb models the colour parallels as release-wide sets, but the Obsidian Preview block 571-585 runs its own ladder: Red is /149 across the rest of the run and /5 there, Orange /49 and /149, Green unnumbered and /25, Pink unnumbered and /75. This was found rather than assumed -- the Red parallel's own checklist page 5 renders SN149 on rows 501-570 and then SN5 on 571-575 in the same response, and sportscardchecklist's per-colour Obsidian Preview pages independently state the same four deviating runs. The base-set rows here carry the main-run value and their visual_notes state the excluded range explicitly; the deviating runs are carried as separate block-scope rows. A consumer summing coverage across both would double-count the 15 Obsidian Preview cards for those four colours. Blue and Purple do NOT deviate, so no base-set row is wrong for them. Recording the shape rather than smoothing it: the schema has no way to say 'this parallel is /149 except in one block'.
• G-011 SUSPICIOUS-UNIFORMITY CHECK FIRED AGAINST THE SIBLING SEASON AND WAS INVESTIGATED RATHER THAN IGNORED. data/chronicles-bkb-2019-20 also has base_set_size=699, and its blocks at 201-500 have IDENTICAL boundaries to this season's: 201-235, 236-270, 271-285, 286-300, 301-335, 336-370, 371-400, 401-435, 436-470, 471-500, with the same sub-brand names on seven of them (Essentials, Marquee, Elite Black Box, Majestic, Gala, Titanium Draft Position, Vanguard). Two identical 699-card anthologies in consecutive seasons is exactly the template-fill smell the extraction spec section 4 check 5 warns about. It was tested and it is real, not fabricated: the rosters differ card by card between the seasons, the blocks at 1-200 and 501-699 differ completely in both names and sizes (this season has Chronicles 1-100, Panini, Luminance, Playoff against 2019-20's Chronicles 1-50, Prestige, Threads, Panini, Luminance, Playbook, Donruss Rated Rookie), three of the shared-boundary blocks carry DIFFERENT names between seasons (Elite against XR at 271-285, Studio against Recon at 286-300, Plates and Patches against Limited at 371-400), and this season's boundaries were read off per-block pages that state their own ranges rather than copied from anywhere. Panini reused a slot layout across two seasons; the finding stands.
• G-012 PARALLEL COVERAGE FIGURES ARE DERIVED ARITHMETIC AND ARE RECORDED AS CORROBORATION, NOT AS SOURCED VALUES. Each colour parallel's tcdb 'Total Cards' was reconciled against the 21 sourced block counts, and all nine reconcile exactly: Bronze 385 is the eleven unnumbered blocks without Obsidian Preview; Green 400 and Pink 400 are those plus Obsidian Preview; Blue 599 and Red 599 are everything except Gala, Titanium and Vanguard; Purple 499 is that minus the three /249 blocks; Gold 614 and Platinum 614 are everything except Gala, Titanium and Obsidian Preview; Orange 115 is the three /249 blocks plus Obsidian Preview. Nine independent totals all landing exactly on sums of the block counts is strong evidence the tiling is right, but the coverage ranges themselves were NOT read off a page card by card -- only the lowest rendered number of each parallel was checked directly, and only Red was walked across its whole run. The per-colour block membership stated in visual_notes should be read as reconciled, not as quoted. Also recorded because the extraction spec warns that 'Total Cards' has been seen counting four different things: on this parent it counts distinct rendered card numbers, consistent across ten pages.
• G-013 INSERTS.CFM NOTES COLUMN ON THIS PARENT IS NOT THE PRIZM SHAPE AND THE BLANK-ANNOTATION TEST WAS NOT RUN. The extraction spec section 3.2 requires checking what the column actually contains before using it. Measured here: the column EXISTS, and across the 13-row Parallel Sets table exactly ONE cell is non-blank -- 'Inserted in 15 card Value Packs.' on Bronze -- and it is a pack-format statement, not a print run. Across the first ten Insert Sets rows exactly one cell is non-blank, 'Distributed exclusively in retail' on Gold Standard Rookie Jersey Autographs, again a channel statement rather than a print run. So this parent is a FIFTH shape beyond the four the spec records: a Notes column carrying sparse DISTRIBUTION notes and no print runs at all. The precondition for the blank test is a split between annotated and blank rows where the annotations are serial statements; there are no serial statements here, so blank carries zero information and the test cannot be run. Every print run in this release was instead obtained by a within-sid serial-token probe on the tier's own sid, which is the fallback the spec prescribes and the strongest form available. Reporting this so the spec's four-shape list can be extended rather than leaving a future agent to rediscover it. Page read at https://www.tcdb.com/Inserts.cfm/sid/205013/2018-19-Panini-Chronicles and again with a varied URL https://www.tcdb.com/Inserts.cfm/sid/205013/2018-19-Panini-Chronicles?x=1 . [REVIEWED 2026-08-19: still open; 24 of 24 parallels lack a sourced channel.]
• G-014 NO SET IS FILED INSIDE THE PARENT CHECKLIST PAST THE BASE RUN, so the extraction spec section 5.3 exclusion rule had nothing to exclude and base_set_size=699 is the whole rendered run. Both boundaries were proved rather than assumed: the last block's own page https://www.sportscardchecklist.com/set-173367/2018-19-panini-chronicles-base-score-basketball-trading-card-checklist states it ends at 699, and the parent's own 'Total Cards: 699' at https://www.tcdb.com/ViewSet.cfm/sid/205013/2018-19-Panini-Chronicles agrees. The nine autograph sets all use letter-coded numbering outside the base sequence -- tcdb sid 207696 renders Chronicles Signatures as CS-AB through CS-ZQ -- so none of them is a numbered block sitting past the base run. This is the opposite structure to the hidden-block pattern the 2018-19 scope memo found on National Treasures, and it matches the standalone obsidian-bkb-2018-19 release, which also found a clean run with nothing past it.
• G-015 COPY CHECK FIRED ON ONE BLOCK PAIR AND WAS CLEARED BY INVESTIGATION, NOT BY DELETION. The extraction spec section 4 check 2 test over the 21 blocks finds exactly one pair with an IDENTICAL roster in identical order: Playoff 171-200 and Crusade 541-570 list the same 30 players in the same sequence (171/541 Michael Porter Jr., 172/542 Josh Okogie, 173/543 Deandre Ayton ... 200/570 Troy Brown Jr.). This is real and not a transcription of one page twice. The two came from different sportscardchecklist pages (https://www.sportscardchecklist.com/set-173352/2018-19-panini-chronicles-base-playoff-basketball-trading-card-checklist and https://www.sportscardchecklist.com/set-173261/2018-19-panini-chronicles-base-crusade-basketball-trading-card-checklist), and Crusade was independently confirmed row by row by tcdb at https://www.tcdb.com/Checklist.cfm/sid/205013/2018-19-Panini-Chronicles?PageIndex=6&sort=0, which renders 541 Michael Porter Jr. through 570 Troy Brown Jr. in that order with the tag 'Crusade'. The surrounding evidence makes reuse expected rather than suspicious: only 132 distinct players fill all 699 cards, and the four 15-card blocks (Elite, Studio, Obsidian Preview, Phoenix) all draw the SAME 15 players in FOUR different orders -- Studio and Phoenix differ by a single swap at positions 12 and 13, which is the signature of two genuine independent listings rather than a template fill, since a template would have produced exact copies. ASYMMETRY WORTH FLAGGING: Crusade is double-sourced but Playoff is NOT -- tcdb's parent checklist truncates at row 170 on page 2 and resumes at 201 on page 3, so cards 171-200 were never rendered by tcdb in this build and rest on sportscardchecklist alone. If either block is wrong it is Playoff, and one Gallery sweep at PageIndex 18-20 would settle it.

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