Batistuta: three rows all print 88, and the plaque never moved
The adversary of this file is one of our own integers. The ability book this archive keeps for Gabriel Batistuta runs to three rows (1996/97 / 1997/98 / 1998), and the overall prints 88 on every one of them; recomputed with our own twelve-term ST weight table it reads 87.68 → 88.08 → 88.06. The scale moved twice across these three rows (+0.40, −0.02); the plaque never moved. A plaque is not a scale: it may only hold integers, and these three rows happen to sit inside a single rounding step. Recomputed player by player across 31 ability books in the library (30 of them filed dossiers, this one included, plus Robert Prosinečki's still unfiled book), exactly two print one constant number while the scale moves. His two summers — the only hat-trick Group D produced in 1994, and the consecutive one in 1998 — are countable line by line in the published ledgers of CASE 014 and CASE 010, yet our file holds no 1994 row for him at all.
The first page of the file
Gabriel Batistuta, born 1969, Argentina striker, retired 2005 — the only date our database offers for the end of his career. The ability book this archive keeps for him is three rows: 1996/97 and 1997/98 at Fiorentina (Serie A); the last row, 1998, carries no club at all, its league field reading only "World Cup". The overall prints the same integer on every one of them: 88 → 88 → 88.
The adversary of this file is not his legs and not those two substitutions; it is our own notation. Take the same 132 cells, multiply each by our twelve-term, 100-point ST weight table and add them up without rounding, and the three rows read 87.68 / 88.08 / 88.06. The scale moved twice across these three rows (+0.40, −0.02); the plaque never moved. A plaque is not a scale: it may only hold integers, and all three of his rows happen to sit within half a step of the same integer — 87.68 rounds up to 88, 88.08 and 88.06 round down to 88. The stored 88 is not lying. It simply has no resolution.
This file opened on 2026-10-09 and cross-references two published case files: CASE 010 (signed 2026-08-23, verdict: MYTH BUSTED) tries the sentence Passarella took off Batistuta, and the team died without him and takes it apart with "he played every minute of that match"; CASE 014 (signed 2026-10-08, verdict: MYTH BUSTED (attribution layer)) recounts every goal of the whole 1994 group. This file reopens neither. It borrows their ledgers as audited fact and deals with a third machine: three rows × 44 columns = 132 cells. The scale moved twice across these three rows (+0.40, −0.02); the plaque never moved.
The two steps have to be set side by side. Step one, 1996/97 to 1997/98: seven of the weighed terms rise and three fall, +0.64 against −0.24, net +0.40 — the heaviest term, Finishing at weight 22, goes from 95 to 96 and Off the ball (14) from 92 to 93. That movement is real: the scale at one decimal sees it (87.7→88.1), the integer plaque does not. Step two, 1997/98 to 1998: four weighed columns set their book minimum here and give back −0.32, while two weighed columns set their book maximum and take back +0.30; the remainder is −0.02 — and this time even the decimal scale is blind, because 88.1 and 88.1 are the same number.
The third thing is how the rows fall. Of the 27 columns that move, the last row monopolises eight unique maxima and the first row fourteen unique minima, while the peak-stamped row monopolises zero — eleven columns share their top across two rows and six share their bottom. Our mark column puts the only peak on 1997/98, and 1997/98 is precisely where the unrounded reading is highest (88.08): the editor's hand and the scale agree on that row; only the plaque fails to follow. Then there is the year with no row. In 1994 he does not exist in this book at all — fifteen other players do — while the CASE 014 ledger holds three goal rows in his name: 2', 44', 90', the last of them a penalty, in Argentina 4-0 Greece. Both facts belong to this archive, and the page sets them down beside each other without converting the case file's goals into the missing row of the player file.
Season-by-season attribute matrix · 1996/97—1998
★ Attribution: row 3 (1998) has no club in our database; its league field holds only the competition name "World Cup". That is how this archive books a national-team year, not evidence of where he played. Heat shading reports absolute level only (scale 60–99); outlined cells mark the table's top value 96 — 4 of them, spread over 2 columns (FIN@1997/98, HGRL@1997/98, FIN@1998, HGRL@1998), both of them weighed in the ST overall (FIN 22, HGRL 8, 30 points between them) and both of them motionless across the last step. This matrix is editorial derivation (L3) — anchored on our audited ledgers and public competition facts, not solved out of any engine database. Appearances, goals, assists, minutes and standings (all five fields NULL in our file) are neither held here nor quoted anywhere on this page: match-level numbers live in the CASE 014 and CASE 010 ledgers. The right-hand column stores the rounded overall; this page also prints the raw weighted reading recomputed from the same table (87.68 / 88.08 / 88.06), and the difference between the two is the subject of this file.
| Season | Age | Pos | MOBILITY | TECHNIQUE | PHYSICAL | DEFENDING | MENTAL | WILL | GOALKEEPER | OVR | ||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ACC | SPT | AGI | OFF | FIN | LSH | PAS | THR | CRO | SET | SHD | FTO | DRI | HEA | LFT | RFT | STA | STR | DUR | AIR | D1V | MAR | POS | AGR | ZON | INT | PRS | VIS | CMP | CON | LEA | STB | WOR | SHW | HGRL | WTW | DFZ | BIG | SAV | AER | CMD | DIS | ONE | OVR | |||
| 1996/97 | 27 | 85 | 87 | 83 | 92 | 95 | 94 | 72 | 83 | 60 | 77 | 86 | 82 | 81 | 83 | 68 | 80 | 78 | 89 | 70 | 82 | 30 | 25 | 45 | 76 | 43 | 40 | 61 | 74 | 74 | 75 | 84 | 75 | 68 | 84 | 95 | 72 | 49 | 77 | 25 | 25 | 25 | 30 | 25 | 88 | |
| 1997/98 | 28 | 84 | 86 | 82 | 93 | 96 | 95 | 72 | 84 | 60 | 78 | 86 | 82 | 80 | 84 | 69 | 81 | 78 | 90 | 68 | 82 | 30 | 25 | 45 | 77 | 43 | 40 | 61 | 74 | 75 | 76 | 85 | 76 | 68 | 84 | 96 | 73 | 49 | 78 | 25 | 25 | 25 | 30 | 25 | 88 | |
| 1998 | 29 | 83 | 85 | 81 | 93 | 96 | 95 | 71 | 84 | 59 | 77 | 86 | 81 | 79 | 84 | 68 | 80 | 79 | 90 | 68 | 82 | 30 | 25 | 45 | 77 | 43 | 40 | 61 | 75 | 78 | 77 | 88 | 77 | 68 | 84 | 96 | 75 | 49 | 81 | 25 | 25 | 25 | 30 | 25 | 88 | |
How to read this table. Start at the right-hand column: 88 → 88 → 88 — three identical cells, over raw readings of 87.68 → 88.08 → 88.06. Then look at Finishing: 95 → 96 → 96. 2 of the 4 outlined cells belong to it and 2 to Goal Hunger; neither column moves between 1997/98 and 1998, so those 30 points of weight contribute 0.00 to the final step. Now read the last row: eight columns take their unique maximum there (BIG/CMP/CON/LEA/STA/STB/VIS/WTW), seven their unique minimum (ACC/AGI/CRO/DRI/FTO/PAS/SPT), and three of the falls (AGI/CRO/PAS) are in columns the ST table never weighs. The book moves; the plaque has nowhere to write it.
Across 132 cells (excluding the computed overall itself) the highest value is 96, and it appears at FIN@1997/98, HGRL@1997/98, FIN@1998, HGRL@1998 — two rows each of FIN and HGRL. These are not sides of the game: in our twelve-term ST table, Finishing carries 22 points, the heaviest term in the book, and Goal Hunger carries 8; together 30 of 100, or 30% of everything the scale reads. That is why the matrix outlines 4 cells rather than one.
What matters is that in the step 1997/98→1998 they are both motionless: FIN 96→96, HGRL 96→96. The 30 heaviest points in the book neither help nor hurt the last row. Of the remaining ten weighed terms, four fall (SPT 1 x weight 10, ACC 1 x weight 6, DRI 1 x weight 8, FTO 1 x weight 8; −0.32 all told) and two rise (+3 x 6, +3 x 4; +0.30). The scale moved twice across these three rows (+0.40, −0.02); the plaque never moved.
a scale of 12 terms and 100 points, allowed one integer on the plaque.
three rows all print 88; the scale moved twice and the plaque never moved
A refusal is also a verdict. Five things this book does not do: the 17 constant columns are not written up as "frozen across three rows" — constancy is our notation, not his seasons; 88 → 88 → 88 is not written as "he was in steady form" — steadiness is a property of integers; put him on the scale and these three rows read 87.68 → 88.08 → 88.06; our file holds no 1994 row for him (15 other players have one that season), so this page neither backfills that year nor converts the three goal rows in the CASE 014 ledger into attributes; the score column of the 1998 ledger sums to 10 while its own summary row declares 9 (Zanetti on 45' exists as a note and never became a row): this file quotes no team total and edits no cell — the one column repair this archive made was done in the open by a disposition script, and its provenance sits in the ledger itself; and the 1998 row is not written as a cause — our file gives it no club, no appearances, no standing, so it explains neither the substitution that memory misfiled nor anything else.
The scale moved twice across these three rows (+0.40, −0.02); the plaque never moved. The sentence blocks two ready-made stories at once. You cannot write "the book shows he slipped at 29": the plaque holds 88 on all three rows and the scale's answer for that step is worth −0.02. You cannot write "the book shows three steady years" either: 27 columns move, the last row monopolises eight of them at the top, and two of those are weighed (Composure, 6 points, Clutch, 4 points). What can be said is that this archive runs two precisions over one table, one of which may only write integers and the other still cannot read −0.02. Reporting that as the conclusion — rather than as his honour or his decline — is the entire position of this file.
+0.40 points of gain, no notch on the plaque.
three rows of book, three verdicts
Fiorentina (Serie A), age 27, the first row of our book. Plaque 88, scale 87.68 — this 88 is rounded up by 0.32. fourteen moving columns take their book minimum here, among them Finishing 95 and Off the ball 92, the two heaviest terms in our ST table. This row holds no appearances and no goals in our file (five fields NULL), so the page writes no match numbers for it.
Fiorentina (Serie A), age 28, the middle row and the file's only peak stamp. Plaque 88, scale 88.08 — +0.40 above the row before, which the plaque cannot show. It monopolises zero unique minimum and three unique maxima (LFT/RFT/SET); Finishing reaches 96 here and stays at the table's top 96 with the row after.
★ National-team year · the club field is empty in our file and league reads "World Cup". Plaque 88, scale 88.06, −0.02 from the row above; eight columns take their book maximum here (BIG/CMP/CON/LEA among them) and seven their book minimum (ACC/AGI/CRO/DRI/FTO/PAS/SPT). This file uses the row neither to explain the substitution that memory misfiled in 1998 nor to be explained by it — that tournament's goal rows live in the CASE 010 ledger.
Half of these come from our two audited ledgers, half from the blanks in our file
Fiorentina (Serie A) — our three-row ability book opens here: plaque 88, scale 87.68. Note where that first row sits: this archive starts writing him at age 27. (BallFact player_seasons)
Argentina 4-0 Greece at Foxborough, its note reading attendance 54,456: the CASE 014 ledger carries three goal rows in his name — 2', 44', 90', the last a penalty, impacts open play, open play, penalty. The same match holds one more row, credited to Maradona on minute 60. The case applies the rule "three or more goals by one player in one match" to the whole group and finds one hat-trick that passes — this one. (CASE 014 · L1)
Our file holds no 1994 row for him. That season is written for fifteen other players in the same table; his first row comes two years later. So this page writes no 1994 attributes, converts none of those goals into numbers, and borrows no other man's season to fill the gap. (library count, computed)
Fiorentina (Serie A), age 28: our mark column puts its only peak on this row. The unrounded reading 88.08 is the highest in the book — stamp and scale agree here, and only the plaque fails to, since 1996/97 and 1997/98 both print 88. (BallFact player_attributes)
1998 vs Jamaica (Group H): score 5-0, impact hat-trick, his three goal rows in that match at 73', 80', 83', the third annotated penalty after foul on Ortega. The note of that same match reads, word for word: second consecutive World Cup hat-trick - unique in history — the previous of those "two consecutive" was the 1994 row above. (CASE 010 · L1)
1998 vs England (Round of 16): he opens the scoring at 6', impact opener. The substitution section of the same ledger carries row 3 with player Gabriel Batistuta OUT / Hernan Crespo IN at 68' and a note containing the REAL source of the false memory. Argentina won that match. The sentence thirty years of retelling attributed to the Netherlands game belongs here — that ruling belongs to CASE 010, and this file does not retry it. (CASE 010 · L1)
1998 vs Netherlands (Quarter-final): the player and minute fields of that match row are both empty, and its note contains plays full 90 minutes and is NOT substituted. One of the four entries in the ledger's substitution section was raised for exactly this match (entry 4, its note reading THE KEY ROW) — that row was found in v2.56 to hold one empty column more than the 11-field header, which pushed his name into the minute field; in v2.57 the disposition script docket024-realign.mjs removed that one empty column, values untouched, and marked the fix at the end of the note. This file can now read it by column name (player Gabriel Batistuta / minute 90 / impact full match) and still treats it only as evidence about our own filing: the substantive proof that he played the whole match is the note on the correctly aligned match row of the same fixture. (CASE 010 · L1)
The last row in our file: age 29, plaque 88, scale 88.06. From here to his retirement in 2005 this archive holds six years with no row (1999—2004) — he did not disappear; we did not write. (BallFact player_seasons)
Three eights, two movements
BONG WENG
Archivist
Writing this file I began by copying the last column: 88 → 88 → 88. Three of the same number. Then I multiplied every cell by our own ST weight table and added without rounding: 87.68 / 88.08 / 88.06. The scale moved twice across these three rows (+0.40, −0.02); the plaque never moved. I ran that same step over 31 ability books in the library — 29 of them have more distinct readings than distinct printings, so "the plaque smooths things over" is routine; but a whole book whose three readings collapse into one integer is rarer: two of 31, one of them his.
So the real adversary of this file is not a decline narrative but our notation. One sentence is polite: "his overall held steady at 88 for three years." On the plaque that is true; on the scale it is not — the +0.40 of step one really happened, and an integer has no cell for 0.4. The other sentence is commoner still: "the book shows him slipping at 29." That step is worth −0.02: four weighed columns drop a point each while two rise 3 each, and the two cancel. Calling −0.02 a decline, or calling 88 → 88 → 88 stability, is letting the plaque testify for the scale. The attributes are L3 editorial derivation, anchors placed by us, so both errors are ours rather than the data's silence.
The third matter is the years with no row and the row that did not line up until it was repaired. In 1994 he has three goal rows in the CASE 014 ledger (2', 44', 90') and not one row in this ability book. In 1998 he has five goal rows of his own, while that same ledger's summary line declares 9 for the team and its score column sums to 10; the missing one (the goal registered as minute 45, filed in a note and never written as a goal row) exists as a note and never became a row. And this file found a third: the very substitution row raised to prove he was not taken off carried one empty column too many, so his name sat in the minute field — v2.57 removed that column, values untouched, and left the provenance at the end of the note. The division of labour is the point: a player file does not edit a ledger; a disposition script does, in the open, marked in the file it changes. Two goal entries remain on file under CASE 015, due 2026-10-15. the plaque never moved That sentence lands again in the note because it names our ruler, not his legs.
CASE 010 tries the sentence Passarella took off Batistuta, and the team died without him and returns MYTH BUSTED on the strength of five matches, eight goal rows and five substitution entries — the case's own first screen says the memory has been retold three decades. CASE 014 audits the whole group's attribution and returns MYTH BUSTED (attribution layer), placing the only qualifying hat-trick of 1994 in his name. This file keeps three rows of ability book; it can neither overturn nor support either verdict, because it does not even hold his appearances. The one thing it can contribute — that the plaque reads 88 → 88 → 88 while the same cells sum to 87.68 / 88.08 / 88.06 — says precisely that our two precisions do not recognise each other. Three pages, cross-linked, different genres, a clear division of labour: the case files handle the claims, this file handles the numbers, the ledgers handle those two summers. The single moment this file lets the two genres meet inside one sentence is the sentence in which it refuses to convert three goal rows into the year his book is missing.
Note: club and league attribution, row-by-row attributes and mark stamps come from the BallFact database (attributes L3 editorial derivation, anchors placed by us); every goal row, impact field, note and substitution entry for both tournaments is counted line by line by this script from the public ledgers public/data/hat-trick-1994-groupd.csv and public/data/batistuta-1998-worldcup.csv, and every minute printed on this page must be the value of a row in his own name with the matching year in the same sentence — two hard gates in the generator, nothing retyped. The 1998 row registers a competition only: standing, minutes and goal attribution are not held here and are not claimed; apps/goals/assists/minutes/team_rank are NULL for all three rows and are likewise left unwritten. Three further disclosures: (1) the 1998 goal section totals 9 minutes, its summary row declares 9 for the team, and the score column of the five match rows sums to 10 — the difference is the goal registered as minute 45, filed in a note and never written as a goal row, which never became a row; the identity gate holds batistuta-1998-worldcup.csv#B1 and batistuta-1998-worldcup.csv#B2, owned by CASE 015, due 2026-10-15. (2) Line 19 of the same file (substitution/4, its note reading THE KEY ROW) once held one empty column more than the 11-field header, which is why his name sat in the minute column; the disposition script scripts/docket024-realign.mjs removed that column in v2.57 with values untouched and the provenance written into the note, and the gate now asserts one column-count rule per file (W1, currently 71 files clean). DOCKET-024 is crossed off in §8. (3) The other constant-plaque book in the library belongs to Robert Prosinečki (eight rows all printing 81): his cause is the same machine, not duplicated values — the unrounded readings span 0.72 points and every one of them rounds to the stored integer. This page therefore never quotes a team total as evidence and never edits a ledger behind the archive's back — it says so once in the sources and once in the note.
three rows, 132 cells, six unwritten years and three goal rows, the plaque and the scale are not the same thing.
Data basis and confidence
public/data/batistuta-1998-worldcup.csv, counted line by line for this file: five match rows (1998-06-14 vs Japan 1-0, 1998-06-21 vs Jamaica 5-0, 1998-06-26 vs Croatia 1-0, 1998-06-30 vs England 2-2 (won 4-3 on penalties), 1998-07-04 vs Netherlands 1-2); the goal section holds nine entries of which one is a summary row with no date, leaving eight goal rows carrying 9 minutes; his five rows read 28', 73', 80', 83', 6' with impact fields match-winner, makes 3-0, makes 4-0, makes 5-0, opener. In the substitution section four rows align with the header and two name him: the 1998 England match at 68' (Gabriel Batistuta OUT / Hernan Crespo IN; note contains the REAL source of the false memory) and the row raised for the Netherlands match, whose note reads plays full 90 minutes and is NOT substituted on the match row itself. The Jamaica row's note contains second consecutive World Cup hat-trick - unique in history. The verdict MYTH BUSTED is signed by that case itself (2026-08-23); this file borrows its counts and quotations, not its trial.public/data/hat-trick-1994-groupd.csv — all 47 data rows align with the header. His three goal rows are all dated 1994-06-21 in Argentina 4-0 Greece (2', 44', 90'; impacts open play, open play, penalty; note attendance 54,456). The hattrick section runs to three rows, exactly one of them reading PASS, and it is his; the scorer section self-declares 3 goals, matching his row count. Maradona also has a row in the same match, on minute 60, quoted here only as ledger text. The verdict MYTH BUSTED (attribution layer) is signed by that case itself (2026-10-08).batistuta-1998-worldcup.csv#B1 and batistuta-1998-worldcup.csv#B2, owned by CASE 015, due 2026-10-15. ② Line 19 of the same file (substitution/4) held 12 fields against a 11-field header when this file found it, so its player cell read as empty and his name fell one column to the right — and that is the row whose note reads THE KEY ROW. The disposition script scripts/docket024-realign.mjs removed that one empty column in v2.57, values untouched, with the provenance written at the end of the note. DOCKET-024 is crossed off in §8. ③ Why the gate did not see it: ledger-check.mjs read by column name and never counted columns; since v2.57 it runs one W1 assertion per file — header width on every row, 71 files in the archive. ④ The two files name him differently: the 1994 ledger writes the bare surname Batistuta, the 1998 ledger the full Gabriel Batistuta — the surname resolves to exactly one player in our whole database (him), the files are never joined by string, and every naming on this page takes its value from the file it comes from. From v2.57 the whole archive splits CSV lines with one quote-aware reader (local-db.mjs's splitCsvLine); under the old naive split these two ledgers would have shifted 40 and 2 rows, which is the arithmetic that opened DOCKET-025.S1 · BallFact CASE 010 "Batistuta 1998" — the substitution that memory misfiled and the sentence "Passarella took off Batistuta, and the team died without him", tried line by line; verdict MYTH BUSTED: /truth/batistuta-1998/.
S2 · BallFact CASE 014 "Group D 1994" — 47 rows of group-stage attribution and the single qualifying hat-trick of 1994; verdict MYTH BUSTED (attribution layer): /truth/hat-trick-1994/.
S3 · The two public datasets /data/batistuta-1998-worldcup.csv and /data/hat-trick-1994-groupd.csv (the only source of every minute, impact field, note and substitution entry on this page, with two hard gates: each printed minute must be a row in his name, and the sentence must carry the matching year).
S4 · BallFact player database player_seasons / player_attributes (slug=batistuta, three rows × 44 columns = 132 attribute rows, all L3; mark column one peak, zero valley; no 1994 row).
S5 · BallFact attribute basis attrs-def (ST weights: twelve terms totalling 100 points, heaviest Finishing at 22); the same table drives the library-wide recomputation quoted above.
S6 · BallFact focus pages: Shapes of a Ledger (volumes compared by book shape; this one joins with the shape "the plaque never moved") and America 1994 (the 1994 file cluster this page belongs to).
S7 · Dataset CSV /data/players/batistuta.csv (three attribute rows, rendered from the database by the build pipeline; before writing, this script reconciles row count and stored overall against it).
S8 · Cross-reference #P00016 Roberto Baggio (/players/baggio/): the same national-team-row skeleton, the opposite flatness — there, the first row takes the most, the last row ties the one before because the columns that fell never enter the scale; here, the scale moved twice and the integer had nowhere to write it. Neither file lets those summers into the causality.