J. O. Mauborgne’s 1915 challenge cipher, Manila, 21 October 1915

01 The document

The cipher is dated Manila, 21 October 1915, with a covering letter of 23 October, from First Lieutenant J. O. Mauborgne of the 8th Infantry to the Director of the Army Signal School at Fort Leavenworth. It has 499 letters, 99 groups of five and a last group of four, with a dash printed after it. Mauborgne meant two sealed envelopes: the plaintext was to be opened after three months, the method a month later. The school circulated the challenge on 13 December 1915; no record of a submitted solution has been found, and the school’s report for 1915–16 does not mention the envelopes (Altifina).

The sheet itself is lost. Louis Kruh found it, with the covering letter, among Army Signal School papers in the New York Public Library around 1991 and printed it in The Cryptogram (1991) and Cryptologia 17(2), 1993, pp. 172–174; when Craig Bauer later asked for a scan for Unsolved! (2017, pp. 351–352), the library could not find it. No image of the original is known, so none is shown here. The ciphertext used is the text of Futility Closet (14 March 2022), which follows Kruh’s printing and is the same 499 letters as on Altifina’s research page (checked letter for letter). The contributor says two AI sessions found it identical to Bauer’s printing; neither he nor anyone here compared the printed letters of Kruh or Bauer by eye.

Altifina’s page (25 August 2026) reports an index of coincidence of 4.30% (0.043 measured here), far below the 6.7% of English, with R occurring 47 times and Z 33 times, and calls the cipher unsolved (no record of a submitted solution), without testing a running key.

02 The cipher

With A = 0, a plaintext letter p is enciphered as c = p − (2 (k mod 13) + 1) mod 26, where k is the key letter. The shift is always odd: the key letters A and N give 1, B and O give 3, and so on up to M and Z, which give 25. To decipher, add the shift. Because every shift is odd, a cipher letter always lies on the opposite side of the alphabet’s even/odd split from the plaintext letter; in the ciphertext, 333 of the 498 neighbouring pairs of letters alternate in that way (about 249 are expected by chance).

The key is a running text. It is, by the contributor, paragraphs 1 and 2 of Part V (Field Fortification), p. 357, of the Engineer Field Manual (Professional Papers of the Corps of Engineers No. 29), “Fortification is the art of increasing by engineering devices the fighting power of troops” and so on to “at or before the close of the campaign”, with “Field fortification” at the head of paragraph 2 replaced by “It”. That edit is the contributor’s choice. It makes the two paragraphs run exactly 499 letters, the length of the message, which would explain why the message stops mid-sentence. Mauborgne’s own statement of his method (the second envelope) is lost, so it is not known that this was his rule or his key.

03 How the key was found

This account is the contributor’s. Its order of events rests on his record only (one issue, his own timestamps and screenshots, nothing from outside seen here).

Jacob Houtrouw had an AI model work at his direction on 11 November 2026. A test plan the model drafted and he approved was sealed before any decryption code touched the ciphertext (a SHA-256 hash of the plan, sealed at 15:28:38 UTC by the session’s own clock). The model’s own forecast of success was 2%. Exploratory statistics showed adjacent cipher letters alternating in alphabet parity, which points to shifts that are always odd. A search for an English plaintext and an English key text through 338 candidate tables ranked the real text first of 41, about 20 standard deviations above the average of the decoys, and planted controls were recovered. The key text was traced afterwards to the manual and the plaintext to Hague Convention IV. He sent the result to Craig Bauer that afternoon and published it at 5:01 PM US Mountain time the same day (the platform dates the post 12 November); his screenshot of the live post is his record of the time. The hash list of his files is said to be anchored by OpenTimestamps in a Bitcoin block (not checked here), and the language-model data files of the search (about 60 MB) are available on request and were not seen, so the search was not re-run here.

The ciphertext supplied the only link: by the contributor’s account no plaintext, crib or key for this item was known beforehand, and neither this project’s searches nor his found either text tied to this cipher before, so the method is recorded as key recovered from ciphertext-only. The search drew on a set of candidate texts, which is why the page says the key text and the plaintext were found by search, not derived.

04 What was checked here

The project wrote its own decoder from the rule (decode.py in the target folder, standard library only) and ran it on the strings in the issue. Before the strings were sent, the project had already tried the public Altifina text against Article 29 and found an odd shift at 495 of 499 positions, with the implied key beginning “FORTIFICATION IS THE ART”.

The ten letters that disagree are at positions 69, 88, 122, 194, 238, 255, 397, 443, 457 and 462. At four of them (122, 238, 443, 462) no key letter at all can give the printed cipher letter: those are slips in the cipher letters, Mauborgne’s, a copyist’s or a printer’s. At the others a different key letter could, so they may be slips in the cipher, in the key text of the edition he used, or in the printing. None of the ten is explained.

The key text against the manual. The contributor’s screenshot is of the fourth (1912) edition, p. 357, in a copy bound with the 1916 addenda (Google Books); it was not seen here. The third (revised) edition of 1909, which he had not seen, is on the Internet Archive; its page 357 opens Part V, and its two paragraphs, with “Field fortification” replaced by “It”, give exactly the 499 letters of the key text, with no difference. A 1917 printing carries the same two paragraphs. The text of Article 29 was compared with the Yale Avalon Project’s copy of the 1907 convention. It reads “the hostile party” and “who have penetrated” where the cipher reads “a hostile party” and “who having penetrated”, and it spells endeavours and despatches; which English printing Mauborgne had was not looked at.

05 The reading

The reading, with the spellings the cipher letters give. A word in braces carries one wrong letter, read from the sense.

A person can only be considered a spy when acting clandestinely or on false pretenses, he obtains or {endeavors} to obtain information in the zone of {operations} of a belligerent, with the intention of communicating it to a hostile party. {Thus}, soldiers not wearing a disguise who having {penetrated} into the zone of operations of the hostile army, for the purpose of obtaining information, are not considered spies. Similarly, the following are not considered spies: soldiers and {civilians}, carrying out their mission openly, entrusted with {the} delivery of {dispatches} intended either for their own army or for …

This is Article 29 of the Regulations annexed to Hague Convention IV of 1907, in a printing whose wording is close to the treaty text. The cipher stops at “for” because the key text ends; the article goes on “the enemy’s army”. The plaintext is English, so there is no translation, and its content says nothing about Mauborgne’s own business.

The count. The plaintext has 97 words. 89 come out of the key with no correction, 91.8%. Eight words each carry one wrong letter read from the sense, so by this site’s rule they do not count as read: 8.2% are holes. The two spelling differences (positions 69 and 457) are not holes, because the cipher letters give pretenses and dispatches and the words read as printed. 91.8% is under the 95% needed for a complete extent, so the extent is partial (89 of 97 words). The contributor’s own figure, given once and not used here, is by letters: 489 of 499 (98.0%), or 491 of 499 with the two American spellings. This page uses the word count only.

06 What remains uncertain

• The ten slips. Eight words each need one letter read from the sense, and two more letters (69, 457) are spelling differences. Whether each slip is in Mauborgne’s sheet, in Kruh’s printing or in the key text of his edition cannot be told without the lost sheet or the printed letters.
• The printed letters of Kruh (1993) and Bauer (2017) were not compared by eye with the transcription by anyone in this record: the contributor’s comparison with Bauer was made by two AI sessions, and both books are in copyright and were not scanned. The Chaocipher Clearing House reports that Bauer lists inaccuracies in Kruh’s articles and thinks Kruh may have introduced errors into the transcription of this cipher; that is one possible source of the slips, and it was not tested.
• The key edit. “Field fortification” replaced by “It” is a contributor choice. The second envelope is lost, so Mauborgne’s own rule and key text are not confirmed; the 1912 page number rests on the contributor’s screenshot, and the 1909 edition (checked here) carries the same wording on its p. 357.
• The search that found the texts was not re-run here: the language-model data files were not seen. The order of events (sealed plan, forecast, hash anchor, release at 5:01 PM MT on 11 November 2026) rests on the contributor’s record. What the project checked is the result, not the discovery.
• The printing of Article 29. “a hostile party” and “who having penetrated” are not the wording of the Avalon copy; which English printing Mauborgne used was not looked at.
• Prior art. The Chaocipher Clearing House and Altifina report no solution; Kruh’s and Bauer’s books were not consulted here, so their statements are known only as summarised by the contributor and by those two pages. A short web search here found no earlier solution; that search was brief, so “no earlier solution found” is not a survey of the literature.

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.

Mauborgne's challenge cipher, Manila 21 Oct 1915 read in part

The first 499 letters of Article 29 of the Regulations annexed to Hague Convention IV (1907) on who is a spy, in the spelling the cipher letters give, stopping mid-sentence at 'for' where the key text ends; the contents say nothing about Mauborgne's own business, the choice of a legal text and a field-manual key are the contributor's finding. Mauborgne meant sealed envelopes with the plaintext and the method to be opened after three months; they are lost.

Builds coherent sentences, but under the read bar of 95% of words read as sense (92% here); the holes are marked in the reading.

PMVEB DWXZA XKKHQ RNFMJ VATAD YRJON FGRKD TSVWF TCRWC
RLKRW ZCNBC FCONW FNOEZ QLEJB HUVLY OPFIN ZMHWC RZULG
BGXLA GLZCZ GWXAH RITNW ZCQYR KFWVL CYGZE NQRNI JFEPS
RWCZV TIZAQ LVEYI QVZMO RWQHL CBWZL HBPEF PROVE ZFWGZ
RWLJG RANKZ ECVAW TRLBW URVSP KXWFR DOHAR RSRJJ NFJRT
AXIJU RCRCP EVPGR ORAXA EFIQV QNIRV CNMTE LKHDC RXISG
RGNLE RAFXO VBOBU CUXGT UEVBR ZSZSO RZIHE FVWCN OBPED
ZGRAN IFIZD MFZEZ OVCJS DPRJH HVCRG IPCIF WHUKB NHKTV
IVONS TNADX UNQDY PERRB PNSOR ZCLRE MLZKR YZNMN PJMQB
RMJZL IKEFV CDRRN RHENC TKAXZ ESKDR GZCXD SQFGD CXSTE
ZCZNI GFHGN ESUNR LYKDA AVAVX QYVEQ FMWET ZODJY RMLZJ
QOBQ-

# mauborgne1915 - mechanical decryption of cipher.txt with key.txt, no correction: p = c + (2*(k mod 13)+1) mod 26, A=0.
# Columns: first position (1-based), then cipher letters / key letters / deciphered letters, 10 letters per line. The dash after the last cipher group of the print is not a letter.
# Raw output: positions 69 and 457 give the American spellings (pretenses, dispatches); eight positions give a wrong letter (see reading.md).
001	PMVEBDWXZA	FORTIFICAT	apersoncan
011	XKKHQRNFMJ	IONISTHEAR	onlybecons
021	VATADYRJON	TOFINCREAS	ideredaspy
031	FGRKDTSVWF	INGBYENGIN	whenacting
041	TCRWCRLKRW	EERINGDEVI	clandestin
051	ZCNBCFCONW	CESTHEFIGH	elyoronfal
061	FNOEZQLEJB	TINGPOWERO	sepretense
071	HUVLYOPFIN	FTROOPSOCC	sheobtains
081	ZMHWCRZULG	UPYINGAPOS	orendeazor
091	BGXLAGLZCZ	ITIONTHESE	stoobtaini
101	GWXAHRITNW	DEVICESHAV	nformation
111	ZCQYRKFWVL	EFORTHEIRO	inthezoneo
121	CYGZENQRNI	BJECTTOINC	frperation
131	JFEPSRWCZV	REASETHEEF	sofabellig
141	TIZAQLVEYI	FECTOFTHEF	erentwitht
151	QVZMORWQHL	IREACTIONO	heintentio
161	CBWZLHBPEF	FTROOPSPRO	nofcommuni
171	PROVEZFWGZ	TECTEDBYTH	catingitto
181	RWLJGRANKZ	EFORTIFICA	ahostilepa
191	ECVAWTRLBW	TIONSANDTH	rtybhussol
201	URVSPKXWFR	EIRMOBILIT	diersnotwe
211	DOHARRSRJJ	YONTHEFIEL	aringadisg
221	NFJRTAXIJU	DORTODIMIN	uisewhohav
231	RCRCPEVPGR	ISHTHEEFFE	ingpeneara
241	ORAXAEFIQV	CTOFTHEFIR	tedintothe
251	QNIRVCNMTE	EACTIONOFT	zonemfoper
261	LKHDCRXISG	HEASSAILAN	ationsofth
271	RGNLERAFXO	TANDHISMOB	ehostilear
281	VBOBUCUXGT	ILITYITDEA	myforthepu
291	UEVBRZSZSO	LSWITHTHEP	rposeofobt
301	RZIHEFVWCN	REPARATION	aininginfo
311	OBPEDZGRAN	OFSUCHDEVI	rmationare
321	IFIZDMFZEZ	CESOFATEMP	notconside
331	OVCJSDPRJH	ORARYCHARA	redspiessi
341	HVCRGIPCIF	CTERFORIMM	milarlythe
351	WHUKBNHKTV	EDIATENOTP	followinga
361	IVONSTNADX	ERMANENTUS	renotconsi
371	UNQDYPERRB	EINAPOSITI	deredspies
381	PNSORZCLRE	ONWHICHDER	soldiersan
391	MLZKRYZNMN	IVESITSTAC	dcivilkans
401	PJMQBRMJZL	TICALVALUE	carryingou
411	IKEFVCDRRN	FROMTHEINC	ttheirmiss
421	RHENCTKAXZ	IDENTSOFAP	ionopenlye
431	ESKDRGZCXD	ENDINGCAMP	ntrustedwi
441	SQFGDCXSTE	AIGNANDWHI	thshedeliv
451	ZCZNIGFHGN	CHMAYLOSET	eryofdispa
461	ESUNRLYKDA	HATVALUEAT	tthesinten
471	AVAVXQYVEQ	ORBEFORETH	dedeitherf
481	FMWETZODJY	ECLOSEOFTH	ortheirown
491	RMLZJQOBQ	ECAMPAIGN	armyorfor

# mauborgne1915 - reading of the cipher (English, spelling given by the cipher letters)

Marks: plain = every letter of the word comes out of `key.txt` with the rule, uncorrected; {x} = one letter of the word
was read from the sense because the key and the printed cipher letter disagree (a restored word, a hole under the README
read bar). Word division and punctuation are ours: the cipher has none. The spellings are those the cipher letters give
(pretenses, dispatches), not the British ones of the treaty text.

## Reading

A person can only be considered a spy when acting clandestinely or on false pretenses, he obtains or {endeavors} to
obtain information in the zone of {operations} of a belligerent, with the intention of communicating it to a hostile
party. {Thus}, soldiers not wearing a disguise who having {penetrated} into the zone of operations of the hostile army,
for the purpose of obtaining information, are not considered spies. Similarly, the following are not considered spies:
soldiers and {civilians}, carrying out their mission openly, entrusted with {the} delivery of {dispatches} intended
either for their own army or for ...

The cipher stops here: 499 letters, mid-sentence, where the key text ends. In the text of Article 29 of the Regulations
annexed to Hague Convention IV (1907) the sentence goes on "the enemy's army"; that continuation is not in the cipher.

## Holes: the eight words with a wrong letter (positions 1-based)

| Position | Word (read from the sense) | Raw decryption of the word | Cipher letter | Key letter | Rule gives (plain) | Can any key letter give the printed cipher letter? |
|---|---|---|---|---|---|---|
| 88 | endeavors | endeazors | U | P | Z | yes |
| 122 | operations | rperations | Y | J | R | no (parity) |
| 194 | Thus | bhus | A | N | B | yes |
| 238 | penetrated | penearated | P | F | A | no (parity) |
| 255 | of | mf | V | I | M | yes |
| 397 | civilians | civilkans | Z | S | K | yes |
| 443 | the | she | F | G | S | no (parity) |
| 462 | dispatches | dispatthes | S | A | T | no (parity) |

"Rule gives (plain)" is the letter the rule deciphers the printed cipher letter to. The word at 255 is "of" in "the zone of
operations"; in the list above the word is the one at that position of the plaintext (a two-letter word, so one wrong letter
makes it a hole).

## Count (97 words)

- 97 words in the 499 letters. 89 words come out of the key with no correction (91.8%). 8 words each need one letter read
  from the sense (8 holes, 8.2%). The two spelling differences (positions 69 and 457) are not holes: the cipher letters
  give pretenses and dispatches, and the words read as printed (the word "dispatches" is a hole for position 462 alone).
- By the README read bar (95% of the words read as sense, no restored word counted) this is 89 of 97 = 91.8%: read in part.
- The contributor's count is by letters: 489 of 499 reproduced by the rule against the plaintext with the treaty's British
  spellings, 491 with the two American spellings. Counted by letters with the American spellings as the cipher gives
  them, 491 of 499 = 98.4%. The word count is the one this target uses.

Key: key.txt

07 Sources

• Jacob Houtrouw, issue #487 of the project repository (12 November 2026): the strings, the audit script, the screenshot of p. 357 and the account of how the key was found.
• Louis Kruh, “A 77-Year Old Challenge Cipher”, Cryptologia 17(2), 1993, pp. 172–174; Craig P. Bauer, Unsolved! (Princeton University Press, 2017), ch. 8, pp. 351–352; neither book was consulted here.
• Altifina, “The Shadow of the One-Time Pad” (25 August 2026): the transcription of Kruh’s printing and the history of the challenge; Futility Closet, 14 March 2022.
• Chaocipher Clearing House: its account of Bauer’s chapter 8: the message was said to be in a new system invented by Mauborgne, no solution was ever submitted by a solver, and Kruh may have introduced errors into his transcription.
• Engineer Field Manual, Professional Papers of the Corps of Engineers No. 29, third (revised) edition, 1909, Part V, p. 357 (Internet Archive ad-1117984); fourth edition 1912 (Google Books, as given by the contributor).
• Avalon Project, Yale: Hague Convention IV, 1907, Article 29.

The notes, the key text, the plaintext, the decoder, the raw decryption and the reading are in targets/mauborgne1915.