Does this answer your question?

@dmg Thanks for the link. It was informative, but it didn't answer my thought on using 'different' ciphers. Still, it gave me enough evidence to back-up my own thoughts. Very much appreciated.

@Kevin I'm okay with this. I posted it here, because we've seen a couple of "double-encrypted" ciphers here on site, which prompted my question. I'm not deep into cryptography, only in puzzles ;c) The questions would be better placed at Crypto.SE, but then the point was double fold: Get an answer for myself & have the P.SE community be aware of it at the same time. But CTV is fine with. My bad.

@Kevin I have made an edit which should explain the scope for this question better.

@BmyGuest Being a brute force attack, it wouldn't really rely on letter frequencies and such. Letter frequencies are lost with pretty much any modern encryption algorithm.

@BmyGuest Also, this double-encryption would be after all a subject to standartization. That is, the algorithm for encrypting/decrypting will be rather well-known.

@dmg: Not sure I follow your latest comment. Are you referring to the Edit and the suggestion for a human-yes-PC-no puzzle? If so, the idea is, that the first decryption-step (the actual puzzle) will contain some plain text (but mostly new cipher-text) which will be hard(ish) for a brute-force method to detect by machine-means, but could be recognized by a human (who only tried one decryption).

Hi

Hey

Thanks for chipping in

I think the purpose of my question was both general curiosity...

Not sure I'm following you on this one

...and the practical aspect for puzzles on this stie.

Yeah, I get that

what I meant was

The last comment you've given "Not sure..." what were you referring to?

(2nd last :c))

All encryption algorithms

are a subject to standartization

it is technically known

and all that

what I meant was

if we add this to any other encryption standard, people will know that "hey, after I brute-force this first step, I should run it through the Ceasar brute-forcer"

Yes, but if the 2nd algorithm isn't known but follows form the part-clear-text ?

Or

as for your example in the edit, I can see two flaws

If it is a known alogorithm requiring a key, and the key is in the part-context

...?

(And I'm not saying it's perfectly foolproof.)

But I think, it would knock out a lot of "wanna-be-the-fastest" tool users...

the first flaw is that, one can store the result of the first decryption and only process it if it is needed

Which is a flaw because... ?

for example, the NSA thinks you are terrorist, they process your history pseudo-manually

:D

and the second flaw is that language processing has gone a long way for the last years

Yes, the 2nd flaw is much stronger. But again, I'm not wanting to fool the NSA here. Just some puzzle-solvers taking shortcuts. Evening the odds for other people enjoying deciphering...

and in your example "NOW USE THE NUMBER BEFORE TWELVE AS KEY FOR SUBSTITUTE" is perfectly understandable for average language processing tools

I see

but what if someone brute-forces, sees "Oh, there is an English word here" and continues from there

Possible, but requires special tools which compare all solutions with a dictionary. I'm pretty sure that makes computation times rather too high for the every-day puzzler.

I think that SHA, RSA or whatever, is more computationally expensive than a dictionary lookup

(Not too speak of the amount of crap the puzzle-solver has to read and try ;c))

Your surely right on that. (I'm not at all an expert on these.)

But if the the "keyword" is an medium-sized English word...

...then the puzzle-solver will have to stop at too many "manual inspections" defying the purpose of the auto-deciphering for a puzzle.

But also, if there is no hint to the key for the industry-standard encryption, average puzzle-solvers won't be able to get to the second part

There is one crucial difference:

(Assuming the cipher itself is the clever puzzle)

If it is a puzzle with an idea, then the puzzler will "solve" it by testing a particular idea.

This is a limited set of "possibilities"

If he instead-brute-forces it, he's faced with many such solutions.

But yes, it is not foolproof. And it is is not generally applicabile

I see, but the process of checking an idea will be very hard to do manually

Personally, I hate decryption puzzles, when there is no obvious cipher

:D

:c)

I think my interest here is, "how can I make a good puzzle" - I'm by far not so good in solving them ;c)

And my feeling was:

A lot of good puzzles are "spoilt" by some (clever!) users brute-forcing everything. It's fair and often nice, but sometimes you do want to avoid these solutions...

..just to give others a chance!

This can be usually battled with short ciphertext

There is no fun in creating a puzzle over a reasonable time-span and seeing it "solved" by a PC in 5min ;c)

Yes, fair point. But then you quickly run into "too small a text, I don't want to try" argument.

Most ciphers (not encryptions, but ciphers), usually require a page of text if they are not Ceasar or something that simple

Ah, yes "too lil' text"

:D

I depends on the viewer I guess. A little cipher text, with reasonable hints to the cipher and key is OK with me

But I think we see eye-to-eye on this now. Thanks for your arguments. I'll keep them in mind. I think the double-encryption might be useful in some limited use-cases, but it is not the golden-hammer to solve all.

And it surely will not define a security standard :c)

Hi AE

I think, that we should try and differentiate between ciphers and encryption here

Ciphers being algorithms and encryptions being anything right ?

hello!

I've also been thinking along the same lines as B recently.

Hmm, perhaps I chose bad words for what I meant

But sometimes ciphers can be puzzles if they require some (human instinct deduced) idea for the algorithm. (See my circular cipher - not yet solved and with a bounty!)

What I meant was. Personally, I associate ciphers with puzzles. Ciphers can be solved manually with a little (or a lot of) thought

Usually when I say "encryption" I'm talking about industry standard encryption algorithms :D

OK. I was using this pair in the way of "algorithm" vs "idea to encode".

So the "look at the letters in 3D from "above" was an encryptions for me, but not a cipher.

I see, that makes more sense now

:D

Still, the easiest way to battle the brute-forcing is by using way too little text

B, the other issue is, is it a bad thing if people are using automated tools to solve a cryptogram? My own personal feeling is it's absolutely not a bad thing - because there's still a level of skill in choosing and applying the tools - but I agree it's a really interesting idea to examine the issues around what makes a cryptogram more/less vulnerable to machine analysis.

I base this on the infamous source of all-info-in-the-world. The always correct and never wrong Wikipedia :c)

"In cryptography, encryption is the process of encoding messages or information in such a way that only authorized parties can read it"

(The simple kind of analysis that puzzlers might do, not the NSA kind of analysis).

@AE No, not at all a bad thing. But it's like an arm's race between puzzle-smith and puzzle-solver :c)

And, after all: A good cipher-puzzle will require the solution to have an explanation afterwards. If it was brute-forced, one still needs to think afterwards ;c)

@BmyGuest Yeah, I'm good with that. I've been trying to get some content onto the site discussing methods for solving cryptograms because some of mine (e.g. the cats one) were turning out to be far more perplexing than I intended them to be.

This was the idea behind my "here's the cipher text, here's the clear-text" puzzle where one has to find the cipher not the clear-text!

@BmyGuest Your circular puzzle is actually an example of what a great cipher puzzle should be

@BmyGuest Absolutely. I'm quite liking the riddle-plus-cryptogram format, because you can brute-force the cryptogram and use that to solve the riddle, or solve the riddle and use that to solve the cryptogram.

@BmyGuest Right. I liked that one very much.

Yes, riddle + cryptogram works nicely

(I just don't like riddles myself ;c) But that's taste...)

@BmyGuest yeah, I don't usually, but this site is sucking me into them! :)

:c)

As for your cats: I thought it was good, but it was the presentation which put me off it. (And not noticing the strong hint ;c))

I just couldn't watch the puzzle as a whole.

But the good thing: I thought me something important!

Ever since, I (try to) make the images in my puzzles smaller to keep everything more closely packed...

@DMG

@dmg @ae Thanks to both of you, but I'll have to earn some money now ;c)

yeah, me too. back later.

same. see you later.