Compete to crack secrets lost to history—for fun, clout, and (no) profit
The number groups have periods inside words and dashes between them, with occasional ordinary-script inserts. They do not use ordinary alphabet position: 26.7.16 eventually reads YOU, not ZGP. Repeated numbers preserve repeated letters, so candidate words have to retain their patterns.
HCP1507 provides the initial foothold. The open “Sat. Eve.” and “Please come to the store first thing” sit beside a sequence that reads I EXPECT TO SEE YOU IN TOWN. EXPECT repeats its first letter in position four, matching the two 3s in 3.24.9.3.20.8. The assignments extend through “MY DEAR LEAH” and the rest of the card without changes. The same values then read two further cards, including independent longer groups for SOMETHING and WINDOW. [2]
| Number | Letter | Number | Letter |
|---|---|---|---|
| 01 | M | 14 | K |
| 02 | Z | 15 | ? |
| 03 | E | 16 | U |
| 04 | R | 17 | I |
| 05 | B | 18 | D |
| 06 | L | 19 | ? |
| 07 | O | 20 | C |
| 08 | T | 21 | W |
| 09 | P | 22 | H |
| 10 | S | 23 | V |
| 11 | F | 24 | X |
| 12 | A | 25 | N |
| 13 | G | 26 | Y |
J and Q are not attested. The unused values 15 and 19 cannot safely be assigned between them. Z has a different evidence basis: body values make the final group on HCP1508 E?RA, and the open signature on HCP1506 reads Ezra. That comparison supplies 2=Z. The numeric signature independently confirms three letters, not all four.
The readings retain the source’s awkward syntax and explicit brackets. Dashes generally divide words rather than punctuate sentences. [4]
MY DEAR LEAH, I EXPECT TO SEE YOU IN TOWN SAT. EVE. DEAR, FOR ME A RIDE WILL DO YOU GOOD, HONEY. I WILL LOOK FOR YOU, SWEETHEART. GOOD NIGHT. [CLOSING UNCLEAR.] XXX.
XXX. OH MY DEAR GIRL, WHAT IS THE MATTER? HAVE I DONE SOMETHING WRONG, OR ARE YOU GETTING TIRED OF ME? I DO NOT KNOW WHAT I HAVE DONE TO MAKE YOU MAD AT ME. [I AM] WORRIED VERY MUCH. ARE YOU TIRED OF ME? IF SO, YOU OUGHT TO G[IVE] YOUR REASON. YOU BETTER COME DOWN SAT. NIGHT. YOUR FRIEND, EZRA.
MY DEAR GIRL, I FELT AWFUL THIS MORNING WHEN I WENT IN TO FIND THAT YOU DID NOT LOOK OUT THE WINDOW. I PUSHED THE BELL ALL THE WAY TO YOUR PLACE BUT DID NOT WANT TO RING IT TOO LOUD TO AWAKEN MY DEAR. DO YOU WANT ME TO STOP SAT. EVE.? AM VERY SORRY I DISAPPOINTED YOU LAST NIGHT. EVER YOUR FRIEND. XXX. EZRA.
The card of 9 May writes LEAH in code while the address spells Leeah. Its closing name is unclear; Ezra is inferred from the packet, not silently inserted into that signature. The card of 23 May visibly supports only G in G[IVE]. The remaining letters are a grammatical restoration, and crowded strokes around I AM and WORRIED VERY are weaker source readings than the surrounding text.
All three cards use one number alphabet, the same separator convention, a matching number hand, and the Knoxville-to-Little Marsh route. Re-enciphering repeated words preserves GOOD, LOOK, MATTER, SWEETHEART, and WORRIED. The decoded spans also meet ordinary writing correctly at “Sat. Eve.,” “morning,” “disappointed you,” “Sat night,” and the open Ezra signature. The first card’s 1-26 and 21-17.6.6 show why every dash cannot be imposed as a hard word boundary.
The portal calls the HCP1506 sender Egna. The direct cursive reading Ezra, coupled with the independent E?RA on the later card, supports a source correction. It does not provide a full biographical identification.
The three notes show an affectionate meeting arrangement, a later question about the recipient’s anger, and an apology after another disappointment. They do not establish ages, marriage, or the eventual relationship. The interval from the first composition on 9 May to the last postmark on 13 June 1913 is about five weeks; an older summary’s “three-week chronology” should not be repeated.
The 28 August 2026 audit found no public reading in its exact-ID, filename, recipient, and phrase searches. All three portal records had no attached answer or key, but the source collection is private and its earlier history is not established. This is a checked three-card campaign reading with bounded prior-publication evidence. Image ownership and reuse permission remain separate questions. [3]