1. Core-2 (conceptual layer)
First, we imagine each letter A–Z turning into a Core-2 number:
A→10, B→11, C→12, ..., Z→35
This step is conceptual in the tool (we don’t show the numbers), but it helps define how
ReCore-3 scrambles the alphabet.
2. ReCore-3 (scrambled alphabet)
ReCore-3 turns each letter into a scrambled value from 0–25. For example:
A→16, B→12, C→9, D→25, E→19, F→6, G→5, H→10, I→15, J→22, ...
Think of this as a “hidden alphabet order” only Sorve knows.
3. Block pattern (0, 1, 2, repeat)
When you type a message, we count only letters (A–Z). Their positions decide the block:
1st letter → block 0
2nd letter → block 1
3rd letter → block 2
4th letter → block 0 (and so on, repeating 0→1→2).
Spaces, numbers, punctuation are ignored for block counting.
4. Key word SORVE → key shift
The secret key for Sorve V3 is the word SORVE.
We map letters to A=0, B=1, …, Z=25:
S(18) + O(14) + R(17) + V(21) + E(4) = 74
KeyShift = 74 mod 26 = 22
This constant 22 is used in every letter’s final position.
5. Final Sorve index (encode)
For each letter:
ReCore = ReCore-3(letter)
Block = position mod 3
SorveIndex = (ReCore + Block + KeyShift) mod 26
Then we pick the cipher letter by index (0=A, 1=B, …, 25=Z).
6. Inverse path (decode)
To decode a Sorve letter:
cIndex = index of cipher letter (A=0..Z=25)
Block = position mod 3 (same rule as encoding)
ReCore = (cIndex − KeyShift − Block) mod 26
Then we look up which original letter had that ReCore-3 value and recover the plain letter.
Because Sorve Code V3 uses a scrambled ReCore-3 table, a moving 0→1→2 block pattern, and a hidden
key-based shift, you can’t crack it by memorising a single alphabet line like Caesar or ROT13.
You must know all three parts of the system to encode and decode correctly.