Compete to crack secrets lost to history—for fun, clout, and (no) profit

• STATUSACCEPTED
• BY@offgramercy
• WRITE-UP8 September 2026

Problems › 167 › solution › 1

Neighboring alphabet pairs predict the untouched rows

Where the number field ends

The card has twelve principal numeric rows, with ordinary handwriting between some runs and mixed signs at the upper and lower edges. The final source count is 126 principal positions: 125 readable numbers and one overwritten value in the seventh row. The open bridges and edge signs were kept out of the key fitting. That makes the result a test of the numbers themselves.[1]

An alphabet with neighboring letters exchanged

Direct and reversed alphabet numbering did not give the accepted reading. The working table exchanges each neighboring pair: 1 gives B, 2 gives A, 3 gives D, 4 gives C, and so on. In alphabet form the rule is BADCFEHGJILKNMPORQTSVUXWZY. For odd n use alphabet position n+1; for even n use n-1. Applying the rule twice returns the starting position, so the same operation encodes and decodes.[2]

cipher number:  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
plaintext:      B  A  D  C  F  E  H  G  J  I  L  K  N  M  P  O  R  Q  T  S  V  U  X  W  Z  Y
key alphabet: BADCFEHGJILKNMPORQTSVUXWZY

Every principal row

Prediction outside the fitted rows

The key was fitted on interior numeric rows, with the ordinary prose, first row, last row, and edge spans withheld. It then predicted both untouched rows exactly: 19.10.14.6.10.20 gives TIMEIS, and 1.2.10.20.20.6.20 gives BAISSES. The second prediction is valuable precisely because it is not fluent “kisses.” No special correction was introduced to improve it.

The finite attack record tested all 624 registered invertible affine variants; its best scored below the adjacent-pair key. Direct and reversed numbering failed, while the registered 25-letter and zero-based controls failed at source value 26. Splitting values into coordinate digits was inconsistent with the frozen atomic-number transcription. These results apply to those registered interpretations, not to every possible cipher model.[2]

Complete mixed letter

[U00 mixed symbols]
(31 Waterloo Rd
TIMEIS gradually
DRAWINGNIGH: HOW
HAPPY I SHALLBE
to be AGAINWITH
MYSWEETHEART I
HOPEDEARYOU are.
PU[?]TEWELL.. ONLY
a LITTLE while & THEN
WITHMYOWN
L[Q?]SE with FONDEW
LOVEANDENDLEW
BAISSES
[U13 mixed symbols] Charlie

The ordinary bridges make the letter’s broad sense clear: Charlie anticipates being with a sweetheart again. The source still contains several irregular outputs. A readable rendering can be offered only with its repairs marked.

(31 Waterloo Rd
Time is gradually drawing nigh: how happy I shall be to be again with
my sweetheart. I hope, dear, you are [quite] well. Only a little while
& then with my own [lips], with [fondest] love and [endless] [kisses].
Charlie

Independent checks and source corrections

A separate reader proposed 128 numbers. Native allograph comparison showed that two extra loops were open-script capital Is, not numbers; another disputed mark was a source 7. The original 126-position count was reinstated after agreement between the primary reader and an independent verifier. All 125 readable values encode back under the one table. The overwritten R07.03 remains empty, and R10’s 18, giving Q, remains at medium physical confidence.

What is not claimed

The diplomatic words are PU[?]TE, L[Q?]SE, FONDEW, ENDLEW, and BAISSES. The bracketed “quite,” “lips,” “fondest,” “endless,” and “kisses” are readings of intended sense, not source-restored letters. Mixed edge spans U00 and U13 are not forced through the numeric rule. Charlie’s legal identity and the recipient’s historical identity remain unknown. The cancel is Charlton on 10 October 1907; Chatham belongs only to the stock picture caption.

On 29 August 2026 HCP marked this object “Not solved” with no key or note. The cipher-sealed provenance audit found no prior exact-item answer, linked sibling, or match in its 428-row DECODE crosswalk. The contribution is new to those checked sources, not an unqualified first-ever solution. The private source scans and derivatives are not republished.[3]

References

• HCP1808 source record, Historical Ciphers Portal
• Complete source-faithful transcript and mechanical result, research corpus (legacy P158)
• Prior-reading search, provenance, and claim limits