I have created a new text encryption technique, well not sure if it already exists but I haven’t heard of it before. Is there any way of getting it tested to see how hard it is to break?

It doesn’t require computers to perform the encryption. If you crack this please write a short reply.

21p j nst vystygr d mrr ucty tvqtiauppm uvkrqoow rztz qya dyit gp yp sxtgrdu tcfypd myo o kedbqy ksttg ga ou nhratt of ykrtt duq eis hp jttyphq yp yfruft pu rtd kea jysg ou dp pu nystl uo gradqy uo udtygr ay trf jrp ktdf py fo ay mtdrz yp frpfmu t eroott bqaauiyef pu ayrhtaq urk rqvyauoam ho iia vysz ypjf axstdt eyrpy d dktay yaxu

To make it easier, here is exactly the same message using the same code logic. Again worked out without needing a computer.

12o kbdr btdruft s qet yvru rbmyusyaoq ybjtmpie exrx mup ftor ho uo dzyftsi rvduof nui p jtsnmu jdryf hp py mgtpyr ps ujtry sim hp jytyphq yp yfruft pu rtd kea jysg ou do pu nystl uo gradqy teroott bqaauiyrf pu ayrhtaq urk rqvyauoam ho iia vysz ypjf axstdt eyrpy d dktay yaxu

If there are deciphering website who like trying this sort of stuff then please post a link.

Have a read of Bruce Schneier’s monthly Cryptogram newsletter. He wrote Practical Cryptography which was the go to book for civilian cryptography when it came out. He also worked on one of the ciphers shortlisted for what became AES although his group’s submission wasn’t selected.

TLDR – it’s easy to invent new cryptography, it’s hard to invent good new cryptography, it’s really hard to analyze new cryptography. If you haven’t got an established record in the field you will have to pay someone a lot to analyze it.

Simon Singh wrote a very nice narrative on ciphers. ‘The code book’.

generally, encoding things well is not a problem. Encoding them and being able to share the message with the intended recipient and have them decode it is the tricky bit. Hence one-time pads and public-private key ciphers.

first glance makes yours look like a substitution cipher but maybe you’ve retained spaces as one of the encryptable characters?

Typical initial steps in exploring encrypted texts are frequency analysis of letters, then letter-pairs, and take it from there

edit: numerous autocorrect gibberish.

Not a scooby. Any clues?

first glance makes yours look like a substitution cipher

Yeah, I thought Caesar Cypher with the first three characters indicating the key. 30 seconds’ poking at it would suggest that there’s more to it than that, but it’s likely a component?

Of note perhaps is the last third of each is identical aside from a slight variation in spacing.

I decided spaces weren’t significant and re-aligned where there’s differences, but as I’m no code breaker not got further than this!

I think I’ve worked out a few words:

‘frozen’, ‘neighbour’, ‘hammer’, ‘dog’, ‘lawn’, ‘sausage’

but there doesn’t seem to be any way of connecting them that makes sense

I can see a pattern. In that there are repeating clumps of “words” that are identical, but then different “words” but they always follow a pattern of containing the same amount of letters.

Omitting some characters from the start reveal:

vystygr d mrr ucty tvqtiauppm uvkrqoow rztz qya dyit gp yp sxtgrdu tcfypd myo o kedbqy ksttg ga ou nhratt of ykrtt duq
btdruft s qet yvru rbmyusyaoq ybjtmpie exrx mup ftor ho uo dzyftsi rvduof nui p jtsnmu jdryf hp py mgtpyr ps ujtry sim

These are different but contain the same number of letters per word.

Then some repeating stuff:

hp jttyphq yp yfruft pu rtd kea jysg ou dp pu nystl uo gradqy
hp jytyphq yp yfruft pu rtd kea jysg ou do pu nystl uo gradqy

Some extra words only found in the 1st example:

uo udtygr ay trf jrp ktdf py fo ay mtdrz yp frpfmu

Followed by more identical stuff:

teroott bqaauiyef pu ayrhtaq urk rqvyauoam ho iia vysz ypjf axstdt eyrpy d dktay yaxu
teroott bqaauiyrf pu ayrhtaq urk rqvyauoam ho iia vysz ypjf axstdt eyrpy d dktay yaxu

I’m not sure if any of that is significant though! But hoping it might lead somewhere. Or maybe not 🤣

1) It has to serve a purpose = No-tech code
2) It has to be fast and easy to use = Yes

The easy bit is the letter swap. Use a UK keyboard, or the alphabet if you want, and any numbers that appear at the front are the offsets. I used the UK keyboard so the 1st offset was 21. The first letter was 2to the right of the actual letter on the keyboard, the 2nd was 1 character, the 3rd was 2, the 4th was 1.

1st example wit 21
p jnst is i hvae
2nd example with 12
o kbdr is i hvae

Both of which don’t quite make words but you can understand them. If you can be bothered to uncode the whole thing it reads out a mis-spelt version of :

I have created a new text encryption technique, well not sure if it already exists but I haven’t heard of it before. Is there any way of getting it tested to see how hard it is to break?

It doesn’t require computers to perform the encryption. If you crack this please write a short reply.

It has to serve a purpose

Does it though, other than being for fun? Spying has moved on from encoding messages on paper and leaving them in dead drops…

You probably already know it, but there are other no-tech ciphers, e.g. Solitaire

https://ermarian.net/services/encryption/solitaire

…although I suppose one might argue that the “tech” is the deck of cards in that example!

There are many indeed. I was just wondering if the one I invent already existed and how effective it might be.

I originally thought of the idea of misspelling words prior to encryption during an A Level project many years ago with just the normal alphabet rather than a keyboard for the letter sequence but no numbers at the front, just moving 3 letters along in the alphabet. It annoyed my math lecturer because he couldn’t crack it but I could write code directly onto the chalkboard without needing any paper or calculations so he knew it must be simple.

Adding the seed numbers at the front just adds to the complication. I only used 2 numbers to keep it easy but there is not reason you couldn’t add more to increase the variation in letter substitution. This would help where it, of, or etc appear many times in the message.

I think the problem with this sort of code is that to humans it looks like gibberish, especially with the letter swapping (hvae), but computers really don’t care about this.  You only had a short message of 4 lines but with a bigger message you would soon have repeating sets of characters as they would only have 2 versions for a 2 digit start code.  So ‘have’ would only have a couple of different versions.  The computer finds matches of 4 letters (it can ignore spaces) so assumes it is a 4 letter word although a human wouldn’t be able to see it.  It then replaces those letters with possible 4 letter words most commonly used and see if that helps produce other common words.  For a human it’s very tedious but computers handle tedious very well

Remember that you only have a two digit letter position swap (ok, it includes a keyboard scramble but really it’s still just a swap to a computer).  The enigma machine iirc had a code that was longer than the message itself so you should never have the same encoded word appearing twice within the same encoded message but they got caught by a word at the start repeating within messages encoded on the same day, as the code only changed once a day.  With your encoding system you have the same a word only encoding to two versions of itself within the same message so you very quickly get repeats that can be homed in on.  We might not know how to decrypt that but other people do

Leffeboy – a good point and well made. However I did the simplest version with just a 2 letter swap sequence just in the hope that someone here might spot it. If you add a 3rd digit or 4th then it really does get challenging, even for a number crunching computer. I also kept the word length and letters used the same and the spaces were exactly as per the source text.

u cn reed ths n mk snse ov wot I rite

Encode something like that and the tedious computer matching fails and the humans can be bothered to try because they have a computer to do it for them.

This is great for graffiti signs where you want to share a message with many people without them being compromised by receiving messages directly. They can then graffiti replies and it can all be sprayed over as vandalism when you are done with it.

Encode something like that and the tedious computer matching fails

Aside from nowadays you have spell checks so can just apply that after each decoding attempt. Plus I would guess that you would still be triggering enough times on the frequency analysis that someone would take a look at the best matches anyway. Might work for short bits of text but once you get beyond that my money on would be on GCHQ winning.

Encode something like that and the tedious computer matching fails and the humans can be bothered to try because they have a computer to do it for them.

with enigma the code length was longer than the message length but it still failed because of humans using the same word and not changing the code enough.   you cant use longer codes because then it is no longer simple enough for humans to do easily.

encryption is difficult.

u cn reed ths n mk snse ov wot I rite

Encode something like that and the tedious computer matching fail

you would like to think that but those are fairly standard substitutions and are well known.  it makes it only a little more complicated although the idea is good.  once one decoded message is known the the cypher fails.  think Rosetta stone

it also fails as humans are likely to use the same encoding numbers more than once at which point the encrypted strings repeat between messages.

I think you would enjoy the Simon Singh book referred to earlier

and good work by @easily. dont think that got the love it deserves