Deboli, St Petersburg, 12 September 1786

At a glanceKey from external plaintext · partial (90% of the cipher)

The letter: Antoni Augustyn Deboli, Polish envoy at St Petersburg, to King Stanisław August, 12 September 1786: the first page (16 lines, 227 groups) of a fair-copy despatch, in cipher from end to end.
Where: AGAD Warsaw, Zbiór Popielów (zesp. 12), sygn. 2, scan 146 right. Catalogue entry 415. The cabinet’s clear text of the time is in Archiwum Królestwa Polskiego 269, scans 253 right – 254.
The cipher: Table 3865, a one-part code of 1–4 digit groups for letters, syllables, endings, words and names, with nulls and suffixed variants (1504z differs from 1504). No key of the time was found.
How it was read: A key (K3, 1,287 entries) was learnt by script on 16 other cipher pieces paired with their clears, frozen and hashed before this page was transcribed. On this page it reads 0.645 of the words of the clear, against at most 0.020 for 200 shuffled keys: partial, under the 0.85 set in advance.
What it says: Deboli has given Vice-Chancellor Ostermann his written observations on Prince Czartoryski’s memorandum and sends them to the King, with a passage added for Ambassador Stackelberg; Sedler presented the memorandum only on the Sunday.
Still open: 6 unresolved groups and 17 restored from the clear; the key is not validated; the rest of the despatch (scan 147) is not read.
Credit: Brice Dudout (L’Europe chiffrée) learnt the key, transcribed and read the page, working with Claude Opus 5.5 agents. The Deboli entries and the leads to AKP 269 are Daniel Bourdeau’s.

01 The page

Volume 2 of the Zbiór Popielów, “Deboli 1786”, holds the envoy’s cipher despatches of that year. Scan 146 right is the first page of the despatch of 12 September 1786, a fair copy without corrections, opening “3865 9” and in table 3865 throughout: 16 lines, 227 groups. The despatch continues on scan 147, which was not transcribed. The images on szukajwarchiwach were used for private research and are not reproduced here, so this page shows the cipher as numbers only.

The cabinet’s clear text of this despatch survives in AKP 269 (header “St Petersbourg ce 12. 7bre 1786. reçue le 22.”). It was used to measure the reading and to restore groups the key cannot give, and is not published here. The content is therefore not new; what is new is the system, table 3865, used by Deboli in 1786–87, and a reusable key for it.

02 The key and its tests

Table 3865 is a one-part code of 1–4 digit groups, with nulls (3865 and 9 as openers, 7, 12 and others) and suffixed variants (a, b, x, z). Its alphabet has no j, v or y (i = j = y, u = v). No key of the time was found among the key collections of AGAD (Zbiór Popielów 444, 451, 445/452, 453/454).

K3 was learnt by script on 16 cipher pieces paired with their clears in AKP 269, with variants and nulls learnt without feedback; its predicted coverage was 0.918. Four rounds were pre-registered, one measurement each, with 200 shuffled-key controls:

R1 (key K1): 0.591, against the 0.60 then required: failed.
R2 (key K2): 0.598: failed.
R3 (key K3): 0.519: failed; one held-out piece turned out to be a corrected draft whose cipher follows the struck-out first state of the text.
R4 (K3 unchanged, this page): 0.645 against at most 0.020, a margin of 0.625: partial (the margin passes, 0.85 is not reached).

On clean copies K3 levels off at 0.63–0.65. The misses are mostly truncated values (reso, exe, uau) and spelling variants (auoit/avois) that the measure does not join, rather than wrong values, which is why the reading by sense is higher than the measure.

03 The reading

The page was re-read group by group. A group is certain when its K3 value equals the aligned letters of the clear, probable when it is read by sense only, restored when K3 has no value or the clear contradicts it. Of 219 groups (without the nulls 3865, 9 and 7): 163 certain, 33 probable, 17 restored, 6 unresolved. Read by sense without the restored groups: 196 of 219, 89.5%, under the 95% bar. In the text, {braces} mark probable words, [brackets] restored ones and ‹…› unresolved ones.

Ainsi que je l’{avois} dit à VMté dans ma {derniere} depeche, que ‹j’›{étois} {resolu} de {donner} au [ViceChancelier] mes {observations} par ecrit sur le memoire du Pce Czartoryski, j’ai {suivi} la [dite] {resolution} et l’ai {executée} avant-hier. Jugeant qu’il {vaut} le mieux de faire parvenir à VMté l’{entier enoncé} de mes {idées} là dessus, je [le] {joins} ici tel que je l’ai [remis]. J’[y] ai {exprès} {ajouté} une epi[so]de, qui se {rapporte} à Mr de Stackelberg, afin que cet Ambassadeur {voye} que je {sais} ‹très› bien que ce {n’est} qu’au moyen d’une [uniformité] ‹…› des {idées} que j’ai avancées, avec {celles} qu’il suggerera ici de son coté, ou plutôt à [l’]aide d’une {correction} qu’il {fera} aux miennes, s’il manque quelque chose à {celles}-ci, que la matiere peut être bien [conçue] ici. Moi je n’ai fait que l’eplucher, eu [egard] [à la] necessité qu’il [y] a eu d’[ob]vier à ce qu’{on} ne ‹pri›t l’e‹change› sur {le} dit memoire. Ce ne fut que [dimanche] [que] Mr de Sedler l’[a] presenté au [ViceChancelier], et comme j’{avois} guetté …

In short: as announced in his last despatch, Deboli has handed the Vice-Chancellor his written observations on Prince Czartoryski’s memorandum and sends the King the full text. He added a passage aimed at Ambassador Stackelberg, so that Stackelberg sees the matter can be well received in St Petersburg only if their ideas agree, or with the corrections Stackelberg makes. Sedler presented the memorandum to the Vice-Chancellor only on the Sunday. The page ends at “guetté”.

04 What remains open

Six unresolved groups: 774 (K3 “exist|texiste”, clear “j’”), 3527 (clear perhaps “très”), 77x and 1039 “interuen” (no counterpart in the clear), 2036a (about “pri”), 279 “belle” (clear “change”).
Seventeen groups restored from the clear: 1647 ViceChancelier (twice), 1602 dite, 1165 le, 2174 remis, 2255 so, 3551a uniformité, 1162 l’, 429 conçue, 682 egard, 1411 à la, 1593b ob, 631 dimanche, 106 a, 2064 (K3 “put”, clear “que”) and 3778 (a null in K3, standing where the clear has “y”, twice). These suggested values were not added to K3.
K3 is not validated (0.645 against 0.85); a better key needs new pairs and a new held-out piece.
The rest of the despatch (scan 147) and the other 3865 pieces of Zbiór Popielów 2 and 3 are not read; they are kept unopened as held-out material for the next key.

Full texts

The transcription and the decryption of every document on this page, so that the reading can be used and checked. Short files are shown here; all are in the repository.

Deboli, 12 Sep 1786, first page (scan 146 right) read in part

Deboli has handed Vice-Chancellor Ostermann his written observations on Prince Czartoryski's memorandum and sends them to the King, with a passage added for Ambassador Stackelberg, whose agreement is the condition for success in St Petersburg; Sedler presented the memorandum only on the Sunday.

Builds coherent sentences, but under the read bar of 95% of words read as sense (90% here); the holes are marked in the reading.

Transcription

z2_146.txt

# Z2-146 — AGAD, Zbiór Popielów, sygn. 2, img 146 d. ; Deboli, St Pétersbourg 12/9/1786. H4 (PREREG-R4-3865, haché fefed850…), page 146 d. entière.
# Transcription Claude (Opus 5.5) 10/11/2026 C17, APRÈS le hachage du PREREG-R4, AVANT toute ouverture du clair AKP 269 img 253-257.
# Notation : comme z2_023/z2_119 (souscrit = suffixe ; ? = douteux). Marque « T » sous 1411 (l. 13) non notée. Réclame « 314 » (bas de page) exclue ;
# chiffres de marge au crayon exclus. 16 lignes (et non 19 comme estimé sur vignette), aucun « 2345 » ni « 9933 », aucune rature.
146R.01 3865 9 165 2084 977 1102 242 602 100 3713 523 1302 561 555 2084 774
146R.02 748 2198 509 609 119 1647 1356 1597 3494 1677 673 3421 1105 1347
146R.03 512 2032 502 7 987 3417 1103 1602 2198 659 1102 109 769 640 235
146R.04 958 9 1062 190 2102 3589 1105 1363 509 811 1698 100 3713 1102
146R.05 714 646 1550 509 1356 990 1110 7 977 1165 1044 989 3453 1100
146R.06 2084 977 1102 109 2174 9 974 3778 109 780 166 3555 640 1654
146R.07 2255 509 2085 2252 2132 100 1381 509 3437 156 2084 369 174 3705
146R.08 640 2084 977 2260 3527 282 2084 314 1516 2082 119 1389 527 3551a?
146R.09 77x 1039 562 990 2084 987 234 658 236 355 2102 2256 877 879
146R.10 2117 2115 989 509 2335 477 1587 1741 100 1162 109 509 527 475e 316
146R.11 3494 2102 828? 231 1304 654 1504 2247 2309 1324 2099 397 100
146R.12 355 315 2084 1103 1333 1723 282 429 989 9 1375 977 1508 813
146R.13 2084 1102 640 1661 389 660 682 1411 1520 1059 2102 3778 100 660
146R.14 507 1593b 3571 657 100 314 2082 654 1504 2036a 3449 1102 640? 279?
146R.15 3421 1103 602 1347 9 314 1504 875 2084? 631 2064 1381 509 2252 506
146R.16 1117 1102 106 2020 119 1647 659 408 974 242 901 3449 3453

Decryption

z2_146.txt

146R.01 ainsi que ie l auoit dit a umte dans ma dernier depeche que exist|texiste
146R.02 etoit reso de donne|donner au [1647] mes obserue tion par ecrit sur le memoire
146R.03 du pce czartoriski iai suiu la [1602] reso et l ai exe e auant
146R.04 hier iuge ant quil uau le mieux de faire paruenir a umte l
146R.05 ete ee nonce de mes idee ladessus ie [1165] iointe ici te l
146R.06 que ie l ai [2174] i  ai exp aiout|aiouta une e pi
146R.07 [2255] de qui se rapport a mr de stackelberg afin que cet ambassadeur uoir
146R.08 e que ie saisie [3527] bien que ce nat qu au moien dune|dun [3551a]
146R.09 [77x] interuen des idee que iai auance es auec celle quil su g ge
146R.10 re ra ici de son cote ou plutot a [1162] ai de dune|dun correspond cob|col
146R.11 tion quil fer aux mi en ne s sil manque quelque chose a
146R.12 celle ci que la matiere peutetre bien [429] ici moi ie nai fait
146R.13 que l e plu cher eu [682] [1411] necess ite quil  a eu
146R.14 d [1593b] ui er a ce qu en ne [2036a] t l e belle
146R.15 sur la dit memoire ce ne fut que [631] put mr de se d
146R.16 ler l [106] presente au [1647] et comme i auoit gue t te

Reading

z2_146.md

# Reading: Deboli to King Stanisław August, St Petersburg, 12 Sep 1786 (AGAD Zbiór Popielów 2, scan 146 right), table 3865: read in part

Key: `key/key3865_K3.tsv` (K3, learnt on pairs, frozen before this page was transcribed; SHA-256 bcb2bfc2...). Decoding with the
frozen key: `decryption/z2_146.txt`. Conventions (Daniel's): plain = certain; {x} = probable; [x] = restored from the clear of
the time; ‹…› = unresolved. The K3 alphabet has no j, v or y: i = j = y and u = v are the system's own spellings, not corrections.

Counts, group by group (219 groups without the nulls 3865, 9 and 7): **163 certain** (the K3 value equals the aligned letters of the
clear exactly), **33 probable** (K3 value truncated or a variant, read by sense only), **17 restored** (no K3 value, or a K3 value
contradicted by the clear), **6 unresolved**. Read by sense without the restored groups: 196/219 = **89.5%**, under the 95% bar.

| line | K3 decoding (frozen) | reading |
|---|---|---|
| 01 | ainsi que ie l auoit dit a umte dans ma dernier depeche que exist\|texiste | Ainsi que je l'{avois} dit à VMté dans ma {derniere} depeche, que ‹j'› |
| 02 | etoit reso de donne\|donner au [1647] mes obserue tion par ecrit sur le memoire | {étois} {resolu} de {donner} au [ViceChancelier] mes {observations} par ecrit sur le memoire |
| 03 | du pce czartoriski iai suiu la [1602] reso et l ai exe e auant | du Pce Czartoryski, j'ai {suivi} la [dite] {resolution} et l'ai {executée} avant- |
| 04 | hier iuge ant quil uau le mieux de faire paruenir a umte l | hier. Jugeant qu'il {vaut} le mieux de faire parvenir à VMté l' |
| 05 | ete ee nonce de mes idee ladessus ie [1165] iointe ici te l | {entier enoncé} de mes {idées} là dessus, je [le] {joins} ici tel |
| 06 | que ie l ai [2174] i [3778] ai exp aiout\|aiouta une e pi | que je l'ai [remis]. J'[y] ai {exprès} {ajouté} une epi- |
| 07 | [2255] de qui se rapport a mr de stackelberg afin que cet ambassadeur uoir | [so]de, qui se {rapporte} à Mr de Stackelberg afin que cet Ambassadeur {voye}, |
| 08 | e que ie saisie [3527] bien que ce nat qu au moien dune\|dun [3551a] | que je {sais} ‹très› bien que ce {n'est} qu'au moyen d'une [uniformité] |
| 09 | [77x] interuen des idee que iai auance es auec celle quil su g ge | ‹77x› ‹interuen› des {idées} que j'ai avancées, avec {celles} qu'il sugge- |
| 10 | re ra ici de son cote ou plutot a [1162] ai de dune\|dun correspond cob\|col | rera ici de son coté, ou plutôt à [l']aide d'une {correc}- |
| 11 | tion quil fer aux mi en ne s sil manque quelque chose a | tion qu'il {fera} aux miennes, s'il manque quelque chose à |
| 12 | celle ci que la matiere peutetre bien [429] ici moi ie nai fait | {celles}-ci, que la matiere peut être bien [conçue] ici. Moi je n'ai fait |
| 13 | que l e plu cher eu [682] [1411] necess ite quil [3778] a eu | que l'eplucher, eu [egard] [à la] necessité qu'il [y] a eu |
| 14 | d [1593b] ui er a ce qu en ne [2036a] t l e belle | d'[ob]vier à ce qu'{on} ne ‹pri›t l'e‹change› |
| 15 | sur la dit memoire ce ne fut que [631] put mr de se d | sur {le} dit memoire. Ce ne fut que [dimanche] [que] Mr de Sed- |
| 16 | ler l [106] presente au [1647] et comme i auoit gue t te | ler l'[a] presenté au [ViceChancelier], et comme j'{avois} guetté … |

Notes on single groups:
- Restored: 1647 = ViceChancelier (x2; K3 had rejected its learnt value), 1602 = dite, 1165 = le, 2174 = remis, 2255 = so (e-pi-so-de),
  3551a = uniformité, 1162 = l', 429 = conçue, 682 = egard, 1411 = à la, 1593b = ob, 631 = dimanche, 106 = a; 2064 "put" is contradicted
  by the clear (que); 3778, a null in K3, stands exactly where the clear has "y" (J'y ai; qu'il y a eu), so 3778 = y is likely (x2).
- Unresolved: 774 (K3 "exist|texiste", clear "j'"), 3527 (clear "très"?), 77x and 1039 "interuen" (no counterpart in the clear),
  2036a (about "pri", + 3449 t = prit), 279 "belle" (clear "change" of l'échange).
- The page ends at "guetté"; the despatch goes on on scan 147 (not transcribed, not read).

## Text (French, 12 Sep 1786, first page)

Ainsi que je l'{avois} dit à VMté dans ma {derniere} depeche, que ‹j'›{étois} {resolu} de {donner} au [ViceChancelier] mes {observations}
par ecrit sur le memoire du Pce Czartoryski, j'ai {suivi} la [dite] {resolution} et l'ai {executée} avant-hier. Jugeant qu'il {vaut} le
mieux de faire parvenir à VMté l'{entier enoncé} de mes {idées} là dessus, je [le] {joins} ici tel que je l'ai [remis]. J'[y] ai {exprès}
{ajouté} une epi[so]de, qui se {rapporte} à Mr de Stackelberg, afin que cet Ambassadeur {voye} que je {sais} ‹très› bien que ce {n'est}
qu'au moyen d'une [uniformité] ‹…› des {idées} que j'ai avancées, avec {celles} qu'il suggerera ici de son coté, ou plutôt à [l']aide
d'une {correction} qu'il {fera} aux miennes, s'il manque quelque chose à {celles}-ci, que la matiere peut être bien [conçue] ici. Moi je
n'ai fait que l'eplucher, eu [egard] [à la] necessité qu'il [y] a eu d'[ob]vier à ce qu'{on} ne ‹pri›t l'e‹change› sur {le} dit memoire.
Ce ne fut que [dimanche] [que] Mr de Sedler l'[a] presenté au [ViceChancelier], et comme j'{avois} guetté …

Translation

translation.md

# Translation (English summary): Deboli to King Stanisław August, St Petersburg, 12 Sep 1786, first page in table 3865

Deboli, Polish envoy at St Petersburg, tells the King that, as announced in his last despatch, he has handed Vice-Chancellor
Ostermann his written observations on Prince Czartoryski's memorandum (the Branicki-Czartoryski affair, and the threat to take it to
the Diet), and sends the King the full text as he handed it in. He deliberately added a passage aimed at Ambassador Stackelberg, so
that Stackelberg sees that the matter can be well received in St Petersburg only if Deboli's ideas agree with his own, or with the
corrections he will make to them. Deboli only "picked the memorandum apart", so that nobody would be misled by it. Sedler presented
it to the Vice-Chancellor only on the Sunday, and Deboli had been watching for him ... (the page ends here).

This is a summary of what is read, not a full translation; restored and unresolved words are marked in `reading/z2_146.md`.

Key: key3865_K3.tsv

05 Sources

AGAD Warsaw, Zbiór Popielów (zesp. 12), sygn. 2, “Deboli 1786”, scan 146 right, on szukajwarchiwach (images not reproduced).
AGAD, Archiwum Królestwa Polskiego 1/11/0/-/269, scans 253 right – 254: the cabinet’s clear text (used for measurement and restorations, not published).
Key K3 (key/key3865_K3.tsv) and its four test rounds (validation.txt), by Brice Dudout.
Related: Deboli 1787, the older code.

Transcription, key, decoding, reading and tests: targets/deboli1786.

Cite this page

Daniel Bourdeau and Brice Dudout, “Deboli to Stanisław August, St Petersburg, 12 September 1786: a page in table 3865”, Cyphersolver, posted 10 November 2026, https://cyphersolver.com/deboli1786.html.