2017-18 Panini Hoops

• G-001 THE HOOPS PREMIUM BOX SET IS AN OPEN CLASSIFICATION QUESTION AND THIS BUILD DELIBERATELY DID NOT DECIDE IT. No hoopspremium-bkb-2017-18 directory was created. WHAT WAS MEASURED, all of it this pass. (1) ITS CHECKLIST IS IDENTICAL TO THIS RELEASE'S BASE RUN. tcdb sid 160898 renders "Total Cards: 300" and Gallery.cfm paginates it to exactly thirty pages of ten with NO page 31, so it renders 300 tiles over 300 distinct numbers; its cards 1-79 match this release's 1-79 name for name (Joel Embiid at 1, Jeremy Lamb at 79) and its cards 291-300 are the same ten Kobe Bryant Career Tribute cards. Checklist.cfm?PageIndex=3 rendered 86 rows, 201-286, every one flagged "SN199". (2) IT DOES NOT CARRY THE PARENT'S 301ST ROW. The parent renders a 301st tile, card number 1 named "Champions"; the Premium tier does not, so the two checklists are identical on the base run and differ by that one extra parent row. (3) ITS TRUE SIZE IS 300, AND THE RIVAL 290 IS A UNIT, NOT A DISAGREEMENT. sportscardchecklist lists "2017-18 Panini Hoops Premium Box Set (290 Cards)" - and it also lists "2017-18 Panini Hoops (290 Cards)" for the BASE SET itself and "2017-18 Panini Hoops Kobe Bryant Career Tribute Premium Box Set (10 Cards)" separately. 290 + 10 = 300. That site excludes the ten-card Kobe block from every base-scope count it prints, including fourteen other colour parallels. specs/2017-18_SCOPE_MEMO.md 5b records the 290 as a size disagreement with tcdb's 300; it is not one. (4) THE TREATMENT REACHES BEYOND A BASE-SET REPRODUCTION. A separate autograph tier exists: "Kobe Bryant Career Tribute Autographs Premium Box", tcdb sid 160903, ten cards numbered 291-300, every row flagged "AU, SN20", which cardboardconnection carries as "AUTOGRAPH PARALLELS (Kobe Bryant Career Tribute Set): Autographs Premium Box #/20." A pure colour parallel of a base run has no chase set of its own. (5) BUT THE TREATMENT DOES NOT DEMONSTRABLY REACH ANY ORDINARY INSERT SET. specs/2017-18_SCOPE_MEMO.md 5b cites an eBay title showing the /199 treatment on "an insert card (#291 Kobe Career Tribute)". Card 291 is NOT an insert: cardboardconnection states the base set is "300 cards, including 10 Kobe Bryant Tribute cards" and this build enumerated 291-300 as base cards. That evidence is about a base card and does not widen the scope. No Premium tier was found for Action Shots, Courtside, Team Leaders, We Got Next, Special Delivery, Back Stage Pass or any other insert set, on tcdb's Inserts.cfm (61 rows, read in full) or on sportscardchecklist's 82-entry index. WHERE THE EVIDENCE POINTS, stated as an opinion and not written into any field: toward PARALLEL-PLUS-CHASE rather than toward a separate release. Three of three sources reachable this pass place it inside the 2017-18 Hoops product - tcdb files it in the PARALLEL SETS section of the parent's Inserts.cfm; cardboardconnection lists "Premium Box Set #/199" among the base-set parallels and gives it no product page; sportscardchecklist lists it under the 2017-18 Panini Hoops index alongside the colour parallels, not as its own product. The 2018-19 build of hoopspremium rests on FIRST-PARTY evidence that has no 2017-18 counterpart: a Panini America launch post naming it an online-exclusive with its own date, price and production statement. Nothing of that kind was found for 2017-18, and its release date is unresolved. The one fact pulling the other way is the /20 autograph tier at (4). RECORDED HERE, DECIDED BY A HUMAN: specs/2017-18_SCOPE_MEMO.md section 8 item 2.
• G-002 BASE-SET IDENTITY COLLISION, EXTRACTION_SPEC 5.2, BUILT-AND-FLAGGED RATHER THAN DEDUPLICATED. The "Premium" base-set parallel row in parallels.csv describes the same 300 physical card numbers as base_cards.csv, on holofoil stock serial numbered to 199. If a hoopspremium-bkb-2017-18 release is ever created, its base_cards.csv would describe those same cards under a second release id and this row should be revisited. The parallel of this situation one season later: data/paninihoops-bkb-2018-19 carries a "Premium" base-set parallel row AND data/hoopspremium-bkb-2018-19 exists as a separate release with its own 300-card base_cards.csv - i.e. the catalogue currently carries BOTH modellings for 2018-19 simultaneously. That is a convention divergence a consumer cannot see, and it is the reason this build refused to pick a side for 2017-18.
• G-003 A 301ST ROW EXISTS ON THE PARENT CHECKLIST AND CANNOT BE WRITTEN. https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=31 renders exactly one tile: card number "1", subject "Champions", Golden State Warriors, page-derived href https://www.tcdb.com/ViewCard.cfm/sid/153685/cid/10331961/2017-18-Hoops-1-Champions . It was read twice under two different URLs and rendered identically both times, so it is the page's behaviour and not a fetch artifact. Card number 1 is already Joel Embiid and migrations/002_core_entities.sql declares UNIQUE (release_id, card_number), so base_cards.csv cannot hold both. sportscardchecklist independently lists "2017-18 Panini Hoops Champions (1 Cards)" as its own one-card product, so the card is recorded as a sets.csv row named "Champions" with card_count 1. UNRESOLVED: what number is actually printed on the card. tcdb says 1; no marketplace listing was found for it this pass. RESOLVED BY THIS: tcdb's "Total Cards: 300" counts DISTINCT ISSUED NUMBERS on this parent, not rendered rows (301) - the EXTRACTION_SPEC 3.2 unit question, decided in the header's favour.
• G-004 POSSIBLE VARIATION ON CARD 83 (George Hill, Sacramento Kings), UNRESOLVED. The first reading of https://www.tcdb.com/Gallery.cfm/sid/153685/2017-18-Hoops?PageIndex=9 rendered that tile with the flag token "VAR"; a second reading of the same page under &sort=0 rendered the same ten tiles with no token on any of them. A token SEEN is evidence and a token NOT seen is evidence of nothing, so the VAR stands as an unexplained positive. Errors.cfm on the parent returns "No errors or variations have been added for this set yet." and that silence is worthless (reference/caveats.csv CAV-020), so it neither confirms nor refutes. No 83b row was written: Gallery.cfm strips a letter suffix, the tile count on page 9 was ten not eleven, and Checklist.cfm page 1 truncated at row 82 on three separate fetches under three different slugs, so no endpoint reached row 83 in the checklist view. WHAT WOULD SETTLE IT: a Checklist.cfm rendering that reaches row 83, or a marketplace title for an 83b.
• G-005 THE DISTINCT-TOKEN ENUMERATION ON THE TWELVE NUMBERED 300-ROW BASE PARALLELS IS PARTIAL, NOT EXHAUSTIVE, AND THE INTEGERS ARE WRITTEN ANYWAY. EXTRACTION_SPEC 3.2 requires enumerating the distinct token set of every tier written, because a per-card variable run reads as a clean sourced integer. ViewSet.cfm rendered between 40 and 100 of each tier's 300 rows before the fetch layer truncated (Artist Proof 73, Artist Proof Black 100, Blue 80, Blue Checkerboard 75, Gold 80, Green 80, Orange 84, Orange Explosion 75, Platinum 88, Red 77, Red Back 81, Red Checkerboard 40, Silver 78, Teal 82, Teal Explosion 70). Every rendered row of every numbered tier carried one token and one only. Exhausting all twelve would cost roughly 360 Gallery.cfm fetches. The integers rest on that partial row-level probe PLUS two independent prose statements each - tcdb's Inserts.cfm Notes cell and cardboardconnection's parallel list, which agree 16 of 16 on names and 12 of 12 on figures. That is materially stronger evidence than the single-source token samples that produced the 2018-19 Noir and Spectra errors, and it is still not a complete enumeration. The residual risk is a late-numbered row in the untouched 200-plus rows of some tier. Red Checkerboard, at 40 of 300 rendered, is the thinnest.
• G-006 ROOKIE REMEMBRANCE PRIME: AN UNHEDGED EDITORIAL PRINT RUN IS REFUTED, AND THE SCHEMA CANNOT HOLD WHAT IS ACTUALLY KNOWN. cardboardconnection states "Rookie Remembrance Prime Relics Set Checklist: 49 cards. Serial numbered #/25." The count is right; the print run is not. https://www.tcdb.com/ViewSet.cfm/sid/160858/2017-18-Hoops-Rookie-Remembrance-Prime renders "Total Cards: 49" with all 49 rows rendered - a complete enumeration - carrying "MEM, SN25" on 48 rows and "MEM, SN20" on exactly one: card RR-HE, Henry Ellenson, Detroit Pistons. WHAT IS KNOWN AND CANNOT BE WRITTEN: this tier is provably serial numbered with a per-card variable run of 25 on 48 cards and 20 on one. EXTRACTION_SPEC 2.3 records that the schema admits no "serial numbered, run varies" state, so print_run and serial_numbered are both UNKNOWN and value_contested is yes. A DISAGREEMENT WITH THIS BUILD'S OWN BRIEFING, recorded rather than smoothed: specs/2017-18_SCOPE_MEMO.md 7.2 and this agent's task prompt both state "SN25 x47 + SN20 x1", i.e. 48 rows. This build measured 48 + 1 = 49 against a header of 49, and sportscardchecklist independently lists "Rookie Remembrance Prime (49 Cards)". The Phase 0 row count is one short. Its conclusion is unaffected and correct.
• G-007 SEVEN SET-LEVEL PRINT RUNS ARE EVIDENCED AND sets.csv HAS NO COLUMN FOR THEM (EXTRACTION_SPEC 2.2). Each figure is a property of the whole set, not a parallel tier, so it is written nowhere but in prose: Championship Moments - "SN99" on all 19 rendered rows; Finals MVP - "SN99" on its single row; Red Hot Signatures - "AU, SN25" on all 50 rows, with tcdb's Notes cell reading "Autographed and serial numbered to 25"; Red Hot Signatures Rookies - "AU, SN25" on all 40 rows; Kobe Bryant Career Tribute Autographs Premium Box - "AU, SN20" on all 10 rows; and Road to the Finals, which is not a flat figure at all - see the next gap. Anyone filtering on print_run will not find these.
• G-008 ROAD TO THE FINALS HAS A PER-ROUND VARIABLE PRINT RUN AND TWO SOURCES MODEL IT AS DIFFERENT NUMBERS OF SETS. tcdb carries one 79-card set (sid 160849, "Total Cards: 79", 79 rows rendered) whose rows carry three distinct literal serial tokens: "SN2017" on 44 rows, "SN999" on 23 and "SN499" on 12, summing to 79. sportscardchecklist carries FOUR sets - First Round (44 Cards), Second Round (21 Cards), Conference Finals (9 Cards), The Finals (5 Cards) - also summing to 79. The First Round figure of 44 matches the SN2017 count exactly; the remaining split (21 / 9 / 5 against 23 / 12) does not reconcile. tcdb's single-set model is written because tcdb is the enumerating source. UNRESOLVED: the correct set decomposition, and which print run attaches to which round.
• G-009 SEVEN AUTOGRAPH-SET COUNTS ARE CONTESTED BETWEEN tcdb AND sportscardchecklist, ALWAYS IN THE SAME DIRECTION, AND tcdb'S FIGURE IS WRITTEN. Great SIGnificance 100 vs 94; Hoops Ink 50 vs 48; Hot Signatures 49 vs 46; Hot Signatures Rookies 40 vs 36; Red Hot Signatures 50 vs 48; Red Hot Signatures Rookies 40 vs 37; Rookie Autographs 50 vs 44. The same disagreement appears on two set-scope parallel rows: Great SIGnificance Gold 100 vs 99, Hoops Ink Red 50 vs 48, Rookie Autographs Red 50 vs 47. CALIBRATION, RUN PER SET TYPE RATHER THAN PER SOURCE: sportscardchecklist agrees with tcdb EXACTLY on all eighteen plain-numbered insert sets, on both memorabilia sets (38, 60), on Rookie Remembrance Prime (49) and on the Premium-box autograph tier (10). It is reliable on the ordinal sets and diverges only on the letter-coded autograph sets - the documented Panini split. A PLAUSIBLE BUT UNCONFIRMED EXPLANATION: tcdb is always the larger figure and tcdb's Hoops Ink includes an EXCH row (HI-CPL, Chris Paul), so sportscardchecklist may be excluding redemptions and unreleased signers. That was NOT verified and is not written anywhere as fact. UNRESOLVED on all seven.
• G-010 GREAT SIGNIFICANCE IS THIS FILE'S WEAKEST COUNT AND IT RESTS ON A HEADER ALONE. https://www.tcdb.com/ViewSet.cfm/sid/160874/2017-18-Hoops-Great-SIGnificance renders "Total Cards: 100" but delivered only 50 of those rows before the fetch layer truncated, so the header was never checked against a complete body - the one thing that made every other count in this file trustworthy. sportscardchecklist says 94. tcdb's 100 is written because tcdb is the enumerating source. Its Gold tier at sid 160875 has the same shape: header 100, 51 rows rendered, sportscardchecklist 99. WHAT WOULD SETTLE IT: Gallery.cfm pagination of sid 160874, ten pages, counting the rendered tiles.
• G-011 CHANNEL IS UNKNOWN ON 30 OF THE 31 PARALLEL ROWS AND THAT IS AN EVIDENCED REFUSAL, NOT SKIPPED WORK. cardboardconnection's 2017-18 Hoops parallel list carries no channel annotation on any of the sixteen base tiers - no "Hobby", no "Retail", no "Blaster", no "Target" or "Walmart" - and tcdb has no channel field at all. Per EXTRACTION_SPEC 2.3, where the channel is not known the channel is UNKNOWN and channel_basis is EMPTY; no product-level note was inherited onto any tier. The one exception is "Premium", written channel=factory-set with channel_basis=from-note on tcdb's own set-note "Factory set; SN199; holofoil cardstock" - a note field, not a distribution statement, which is why the basis is from-note and not sourced-exclusive. DO NOT CARRY THE SIBLING SEASON ACROSS: data/paninihoops-bkb-2018-19 writes channel=blaster with channel_basis=sourced-exclusive on its Blue and Red tiers off Steel City blaster-box copy for THAT season. No equivalent 2017-18 retailer copy was found this pass, and a channel is not inheritable across a season boundary.
• G-012 FOUR SET AND TIER NAMES DIVERGE BETWEEN SOURCES OR BETWEEN SEASONS; NOTHING WAS RENAMED. (1) tcdb and sportscardchecklist both print "Back Stage Pass" (two words) for 2017-18, while data/paninihoops-bkb-2018-19 carries the one-word "Backstage Pass" for its own season - and that 2018-19 row's own note records checklistcenter calling it "Back Stage Pass" there. The line may genuinely change spelling between seasons, or one catalogue may be wrong in one of them. (2) tcdb names the base tier "Red Back" (singular); cardboardconnection prints "Red Backs" and sportscardchecklist prints "Red Backs". tcdb's string is written. (3) tcdb names the set "NBA 2K18 Shaquille O'Neal"; sportscardchecklist names it "Shaquille O'Neal NBA 2K". tcdb's string is written. (4) tcdb titles the whole product with the bare string "2017-18 Hoops"; cardboardconnection titles it "2017-18 Panini NBA Hoops Basketball Cards". This one is resolved, not open - see the brand-identity gap.
• G-013 NBA 2K18 SHAQUILLE O'NEAL: NUMBERS 16-25 ARE PRINTED AND MUST NOT BE NORMALISED, AND THE COUNT IS CONTESTED. tcdb renders "Total Cards: 10" with ten rows, every one Shaquille O'Neal, numbered 16 to 25, and its own Inserts.cfm Notes cell states verbatim: "Cards are numbered #16-25 as a continuation of the 2017-18 Hoops NBA 2K18 Legend Edition Shaquille O'Neal set, which ended at #15." sportscardchecklist lists "2017-18 Panini Hoops Shaquille O'Neal NBA 2K (12 Cards)" - twelve, not ten. The 1-15 predecessor block is a DIFFERENT tcdb product ("2017-18 Hoops NBA 2K18 Legend Edition Shaquille O'Neal", named but not researched in specs/2017-18_SCOPE_MEMO.md section 1) and is NOT carried in this release. UNRESOLVED: whether the set is 10 or 12 cards, and whether a 2017-18 Hoops NBA 2K18 Legend Edition release belongs in this season's scope.
• G-014 HOT SIGNATURES RENDERS 49 ROWS WHILE ITS RED TIER RED HOT SIGNATURES RENDERS 50, SO THE RED TIER IS NOT A SUBSET OF ITS PARENT. Both were read against their own headers and both are complete enumerations: sid 160866 "Total Cards: 49" with 49 rows all flagged "AU", sid 160867 "Total Cards: 50" with 50 rows all flagged "AU, SN25", and the card codes are the same HS- prefix in both. sportscardchecklist reproduces the same inversion at a lower level (46 against 48). The mirror of the 2018-19 finding, where the red tier was one SMALLER than its base (39 against 40). UNRESOLVED: which signer appears in red and not in the base tier.
• G-015 PLATINUM COVERS 290 CARDS WHERE EVERY OTHER BASE PARALLEL COVERS 300, AND TWO INDEPENDENT SOURCES AGREE. tcdb renders "Total Cards: 290" for sid 160890 against 300 for all fifteen other base tiers. sportscardchecklist lists a "Kobe Bryant Career Tribute" tier for fifteen of this release's colours (Artist Proof, Artist Proof Black, Blue, Blue Checkerboard, Gold, Green, Orange, Orange Explosion, Premium Box Set, Red, Red Backs, Red Checkerboard, Silver, Teal, Teal Explosion) and lists NONE for Platinum. So Platinum appears to cover base cards 1-290 and not to extend to the ten-card Kobe Bryant Career Tribute block at 291-300. NOT DIRECTLY VERIFIED: no Gallery.cfm pagination of sid 160890 was run to confirm its highest rendered card number. UNRESOLVED at the card level.
• G-016 VARIATION SETS: ABSENCE IS NOT ESTABLISHED AND IS NOT CLAIMED. Errors.cfm on the parent returns "No errors or variations have been added for this set yet.", and per reference/caveats.csv CAV-020 that silence discriminates nothing. What DOES survive independently: base_set_size 300, a contiguous 1-300 base run, and no numbered block past 300 - all resting on a row-by-row enumeration across Checklist.cfm pages 1-3, Gallery.cfm pages 9, 10, 19, 20, 28, 29, 30 and 31, and Checklist.cfm?PageIndex=4 rendering no rows. What does NOT survive: any claim that this release has no variation layer. One positive signal exists and is unresolved - the VAR token seen once on card 83. EXTRACTION_SPEC 5.5 MIRROR NOT RUN: the 2018-19 Hoops parent (sid 182809) was NOT checked for a 2017-18 Hoops Update block. No number in this release is recorded as unissued, so nothing here depends on it, but the check is owed.
• G-017 checklistcenter HAS NO REACHABLE 2017-18 PANINI NBA HOOPS PAGE, RECORDED AS NOT-FOUND RATHER THAN AS ABSENT. A site-scoped search returned its 2017-18 pages for Dominion, Immaculate, Opulence, Flawless, Noir, Essentials and Impeccable, and its NBA Hoops pages for 2023-24 and 2024-25, but nothing for 2017-18 Hoops. Per reference/caveats.csv CAV-037 its slugs are irregular and no constructed URL was tried, so this is an absence of a search result, not an absence of a page. CONSEQUENCE: the usual strong count source for this line is missing this season, and sportscardchecklist stood in for it. beckett.com was not fetched at all - the task prompt records it as 302-redirecting to a maintenance page - though a web search surfaced a beckett.com/news article titled "2017-18 Hoops Premium Box Set Checklist, Info, Kobe Bryant Autographs" that would likely bear directly on the Premium question if that host ever becomes reachable.
• G-018 MARKETPLACE CORROBORATION WAS ATTEMPTED FOR THE PREMIUM TIER AND REJECTED AS CONTAMINATED. https://www.ebay.de/itm/257051165212 carries the title "2017-18 Panini NBA Hoops - Kobe Bryant Career Tribute Kobe Bryant #293 Premium Box Set /199". It carries no tcdb flag string anywhere on the page, so the standard fingerprint test PASSES - and it is still rejected, because the title is in eBay's own catalogue form "<Release> - <Set> <Player> #N", the documented shape a clean fingerprint does not clear (CHASE_CHECKLIST_SPEC). A catalogue-form title is seeded from aggregator data and cannot independently corroborate the aggregator. NOT USED as evidence for anything. No free-text seller title for a 2017-18 Hoops Premium card was found this pass. The /199 figure rests instead on tcdb row-level tokens across three endpoints of sid 160898 plus cardboardconnection's "Premium Box Set #/199".
• G-019 BASE PRINT RUN IS "unnumbered" WITH basis "assumed", NOT "sourced", AND THE DISTINCTION IS DELIBERATE. No source states in words that the 2017-18 Hoops base cards carry no serial numbering. What exists is a within-renderer positive discriminative test in both directions: the same Checklist.cfm and Gallery.cfm views that render no serial token on any of the 300 base rows render "SN199" on every rendered row of the Premium tier, "SN99" on every rendered row of Green, "SN125" on every rendered row of Teal and "SN99" on all 19 rows of Championship Moments. EXTRACTION_SPEC 2.1 is explicit that strong indirect evidence of exactly this shape is "assumed" and that an agent should not talk itself into "sourced". This row follows that.
• G-020 ROOKIE FLAGS AND THE KOBE SUBSET: PROVENANCE AND THE ONE RANGE THAT RESTS ON ONE ENDPOINT. 44 of the 300 base cards carry the literal token "RC" - cards 56, 113, 153 and the contiguous run 250 to 290. Cards 1-82, 101-183 and 201-272 were read off Checklist.cfm, which rendered "RC" on 56, 113, 153 and 250-272 and nothing on their neighbours - a within-page positive discriminative test. Cards 273-290 were read off Gallery.cfm pages 28 and 29, which rendered "RC" on all eighteen. THE WEAK RANGES: cards 83-100 and 184-200 were reached by Gallery.cfm ONLY (pages 9, 10, 19, 20), each page read twice under two different URLs, with no RC token seen on any of them; no Checklist.cfm fetch ever reached those rows. rookie_flag=no is written there on the strength of Gallery.cfm demonstrably rendering RC on pages 28 and 29 of the same set - but the same endpoint dropped a VAR token between two readings of page 9, so its flag column is known to be lossy. Separately, card_tier "Kobe Bryant Career Tribute" on cards 291-300 rests on cardboardconnection's "300 cards, including 10 Kobe Bryant Tribute cards" plus the rendering of those ten tiles as Kobe Bryant; sportscardchecklist corroborates the block's existence and size at 10. No source names the subset's boundaries as 291-300 in words.
• G-021 BRAND IDENTITY RESOLVED, RECORDED BECAUSE IT WAS A NAMED HAZARD AND NOT BECAUSE IT IS OPEN. tcdb titles this product with the bare string "2017-18 Hoops" and names no manufacturer anywhere on the parent set page; cardboardconnection titles it "2017-18 Panini NBA Hoops Basketball Cards" and sportscardchecklist indexes it under "2017-18 panini hoops". reference/brands.csv carries two rows whose brand string is "Hoops": paninihoops = "Panini NBA Hoops Basketball" (Panini) and hoops = "Topps NBA Hoops Basketball" (Topps). The build script asserts both mappings and aborts otherwise; both assertions passed. product_name is written byte-identical to the paninihoops row. No brand row was minted and no escalation is needed. NOTE FOR THE LEDGER: reference/brands.csv also carries hoopspremium = "NBA Hoops Premium Box Set Basketball" (Panini), which is the line the open Premium question would attach 2017-18 to if it is ever resolved that way.
• G-022 INSERT AND AUTOGRAPH CARD-LEVEL CHECKLISTS ARE OUT OF SCOPE FOR THIS PASS (CHASE_CHECKLIST_SPEC governs them) and no insert_cards.csv was written. sets.csv records every set's existence and count. Prefixes observed and available to that later pass, recorded so they are not re-derived: Hoops Ink HI-, Hot Signatures HS-, Hot Signatures Rookies HSR-, Great SIGnificance GS-, Rookie Autographs RA-, Rookie Remembrance RR-, Rise N Shine Memorabilia RS-. Red Hot Signatures reuses the HS- codes of Hot Signatures and Red Hot Signatures Rookies reuses HSR-, so (set_name, card_number) is the only safe key in this release. One EXCH row is known: Hoops Ink HI-CPL, Chris Paul, flagged "AU, EXCH".

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