| 1 person marked this as a favorite. |
Sending coded messages back to our employers is soon going to be an important part of our current game, and while I'm fairly clear on the use of linguistics to decipher code, I need to know a bit more about the process of coding messages.
As such, I have two questions:
1) Would changing a message into code use linguistics? If not, what would it use? (I'm the bard in the group, and the one who has ranks in linguistics, so at the moment I think this is mostly my responsibility.)
2) Would the success, or lack of it in my skill check to create a code affect the ease of deciphering it? I would assume so, and that a poor code would be easier to figure out, but I don't really know how the math of it would work. It's assumed that our employers will successfully translate any message we send them, no matter how well coded, but that enemies will be stumped by sufficiently difficult code, so my best guess is that the check for any enemy trying to crack a code I write would be a linguistics check opposing the one I made when writing the code.
I'd appreciate help figuring this out. I'm just a player this time, but if it gives our GM one less thing to worry about then I hope I can manage to get a general rule for coded messages that he approves of.
You use bluff to send coded messages. The enemies would oppose your bluff with sense motive. As a GM, I'd give you a bonus to your bluff and/or a penalty to their sense motive based on how effective the code is, which I'd base on Linguistics or a relevant Knowledge skill.
You use bluff to send coded messages. The enemies would oppose your bluff with sense motive. As a GM, I'd give you a bonus to your bluff and/or a penalty to their sense motive based on how effective the code is, which I'd base on Linguistics or a relevant Knowledge skill.
I got the impression the bluff version of secret messages was more for in-person conversation and trying to keep someone out of the loop in a conversation than it was for coded letters.
Still, that is useful. I'll look into it a bit more.
| 1 person marked this as a favorite. |
You need to read a great novel called Cryptonomicon, which is about code breaking in WWII and Encryption on the internet.
Basically, most codes are substitution codes. You substitute one symbol for each letter. For example you can have a two shift code in which each letter is represented by the letter two letters down the alphabet,
so
ABC =
CDE
That's a simple code. Another way that is harder is to have a prearranged way to generate a number, for example, using the date, like today is the 8th, so today's code shifts eight places in the alphabet.
The more complicated codes can shift for each letter. So if today it the 8th, the letter A is shifted over eight letters, say G, then the next letter in the alphabet, B, is shifted over 9 letters, and the next is shifted over ten.
The best encryptions use something called a one-time pad, where each of you have a key that is agreed upon that will be used only once. Those are almost impossible to break without a computer.
In Pathfinder terms, I would say Linguistics would work. As long as you know the language the code is in, A simple substitution code would be DC 20, a shifted code DC 25, an evolving shift DC 35.
However, given the presence of such magics as comprehend languages and divination spells like augury "Weal or Woe if I use a substitution code to try and break this code?" I can't see how codes could be much use.
A better solution would be something like a magic mouth given specific conditions who to speak to and what to say. Then there's spell like sending and animal messenger as well.
|
Here is my house rule for Linguistics:
Linguistics: You can use linguistics for help in solving puzzles, creating and breaking cyphers, and figuring out a character’s regional accent. Every time you take a language you learn a cypher or code. Hand signals are codes.
So I just let cyphers and codes tag along with linguistics. The more complicated the cypher the higher the DC to break it.
The complexity of cyphers will vary with the tech level of the people using it. An ancient empire like Taldor might be using complicated codes. So might Cheliax. The gorilla king might not be using anything very complex.
|
|
Let's talk real world for a second. Codes are agreed upon ways of sending messages. Things like morse code, flag sephamore, symbols, etc..
Cyphers are methods of changing plain text into something that is not plain text. Substitution cyphers are the simple ones.
I think taking one point of linguistics to learn a cypher is not unreasonable. (Though I house rule it differently).
Just take the point and use an agreed upon cypher. Your employer should know the cypher and that should be it.
I disagree that it uses Linguistics.
Linguistics is used to break a cipher you don't know the key to.
Deciphering a code you know the code to is usually trivial and just takes some time, and maybe some basic math skills (like "ok, 6 letters from the J is the D"). 6 year olds can do it with those code-wheel thingies you find in comic books. (yeah sure that's a cheap cipher, but even substitution ciphers aren't any more complicated if you know the substitution key - and those were really tough nuts back in the middle ages to crack)
The same is encoding a message with a known algorythm and key, which I assume is provided by the employer, because he needs to know that to decipher it again.
Coming up with the code in the first place, ok I can see that that my be Linguistics, but as I said, actually encoding the text is a trivial matter. If you can follow basic instructions it's simple.
Bluff is something completely else and is used in spoken (or written) conversation, when you want to hide the existance of the code in the first place. You say something harmless like "It's nice weather today" and are actually meaning "Ok, it's time to steal the McGuffin now, the guards are on the other side now"
I'd say that using a code that has been pre-defined, and you have access to, should be pretty much a no-brainer, without rolling anything.
Of course, if you need to follow a complicaten encryption system, or have to learn some things by heart (i.e. not having access to code tables, an encoding wheel or whatnot) might call for some checks in the encoding or decoding is nontrivial.
Actually designing a cipher of your own would, in my opinion, call for Linguistics, as well as breaking an unknown one does.
|
|
First, don't get into a "this is so easy it does not require a skill." Lots of things seem easy but have skills assigned. Anybody can make a forgery in the real world but linguistics is required in the game and is trained only. Lots of examples like that. Assigning it to a skill is also favorable to the player as he can apply skill points to any rolls rather than just making an INT check.
Second, depending on he cypher the actual encoding of the message may be important. Enigma was originally broken because german operators were lazy. They used words and keys repeatedly which gave the code breakers the opportunity to break things. Sometimes things were sent in code by one method and in the open by another method. A favorite crib was weather reports. These were often sent in the open. Then ships got the report sent via enigma. Comparing the two helped in the breaking.
Now she is not going to use an enigma type machine but the problems with enigma are common problems with many type of code. Starting out every message with the same salutation, ending the same way, always putting the date at the beginning or end.
Using a skill to allow a skill roll simulates these possible errors. She could roll and add her linguistics and that is the DC of breaking that message. Forgery uses a similar method and I think it meshes well.
Reaper Miniatures
|
Encoding a message: Linguistics check. This sets the DC of decoding without the cypher. The better you are at lingusitics, the more complicated your cypher can be. Other party members may add +2 to the DC of the decode with various skills - Knowledge engineering might add mathematical formulae into the cypher, making it harder to break without the cypher, Knowledge arcana might add magical formulae. Alchemical formulae, even. One might go so far as to load in planar coordinates, and these add to the complexity of the encryption/decryption algorithm. Let each player come up with the rationale behind why their obscure but favorite skill makes the algorithm more complex, and I'd allow it.
Decoding a message with the cypher - DC5 Linguistics check. It is possible to mess it up, but still "easy".
Decoding a message without the cypher - Linguistics check DC of the encrypting check.
|
Reaper Miniatures
|
A good example - encoding the phrase "Attacking the dragon at dawn, have the army ready"
My linguistics is +1, I am poorly trained at it. I roll a 5, for DC 6. Thus, my code is fairly simple. "Bttbckjng thf drbgpn bt dbwn, hbvf thf brmz rfbdz" Simple code, substituting only vowels for the next letter in sequence, morons will be 100% fooled, and those untrained in linguistics - specifically the skill of words and language - won't be able to make heads or tails. Even a poor linguist has a very good chance of breaking.
My linguistics is +12. I have trained extensively, though am by no means a master. I know a lot of languages, and can use these to complicate my messages, substituting in draconic words for every third word, and celestial for every other non draconic word. I roll a 16, DC28. After applying the cypher, we get "GJkaoi kihhut e ajemloojkuw, abhhed kuentr." Novices at lingusitics have no chance.
My linguistics is +25. My companions are skilled planar travelers, alchemists, authors of magical formulae, and world renowned siegecrafters, as befits our high level. I roll middlingly well, 12m but each of my 3 allies contributes +2 to the complexity of the encryption algorithm, for a DC of 43. The message reads "œkkuƒ, ÿ ó∑∑ß åzøµ" Most people won't even recognize half the symbols used, let alone stand a snowball's chance in Osirion of grasping it. Heck, at +25, without the core cypher and my companion's help, I myself only have a 15% chance to crack it, and I'm Intelligence 20 with 20 ranks!!
Another example of a code would be to use an existing book, agreed on both parties. This kind of code is very hard to break if you don't have a copy of the exact same book the two are using.
For example, if someone in the real world used a copy of the King James Edition of the Bible, they would have to have the exact same edition, published by the same source. This type of code works by using numbers to designate the Page, the Paragraph, and the Word (something like 324-4-17). But it also takes a bit longer to decrypt, unless you are using a computer.
Master Arminas
...Enigma was originally broken because german operators were lazy. They used words and keys repeatedly which gave the code breakers the opportunity to break things...
This is pretty much why we want to avoid using one set encryption key. The more repetitive we are, the more likely someone's going to catch something.
My linguistics is +1, I am poorly trained at it. I roll a 5, for DC 6. Thus, my code is fairly simple. "Bttbckjng thf drbgpn bt dbwn, hbvf thf brmz rfbdz" Simple code, substituting only vowels for the next letter in sequence, morons will be 100% fooled, and those untrained in linguistics - specifically the skill of words and language - won't be able to make heads or tails. Even a poor linguist has a very good chance of breaking.
My linguistics is +12. I have trained extensively, though am by no means a master. I know a lot of languages, and can use these to complicate my messages, substituting in draconic words for every third word, and celestial for every other non draconic word. I roll a 16, DC28. After applying the cypher, we get "GJkaoi kihhut e ajemloojkuw, abhhed kuentr." Novices at lingusitics have no chance.
My linguistics is +25. My companions are skilled planar travelers, alchemists, authors of magical formulae, and world renowned siegecrafters, as befits our high level. I roll middlingly well, 12m but each of my 3 allies contributes +2 to the complexity of the encryption algorithm, for a DC of 43. The message reads "œkkuƒ, ÿ ó∑∑ß åzøµ" Most people won't even recognize half the symbols used, let alone stand a snowball's chance in Osirion of grasping it. Heck, at +25, without the core cypher and my companion's help, I myself only have a 15% chance to crack it, and I'm Intelligence 20 with 20 ranks!!
I may just have to show this post to my GM as an example of coding. It's excellent.
...given the presence of such magics as comprehend languages and divination spells like augury "Weal or Woe if I use a substitution code to try and break this code?" I can't see how codes could be much use.
A better solution would be something like a magic mouth given specific conditions who to speak to and what to say. Then there's spell like sending and animal messenger as well.
I would assume since encrypted messages are a fairly major aspect of the game that something would be houseruled to make them less easily breakable by magic. Maybe our own group's wizard would do something to aid in that. I don't know, so I can't really say for sure.
Are you a gamist?
If so, the difficulty of the cipher is determined by the linguistics skill of the cipher's developer, probably set by his take 20. Reduce that difficulty by 5 if you've got a lot of ciphertext available to you with decryption. Reduce it by 10 if the cipher user has been sloppy (e.g. reusing keys or other common cryptographic blunders). Most of the time cryptanalysis means a take-20.If you're a simulationist, recognize that most ciphers used throughout most of history were very secure for at least a while after they were developed. Most in fact required humint and what we'll call 'rubber hose cryptanalysis' to break into. It is a lot easier to use a cipher than it is to break it. There are even ciphers which are mathematically unbreakable if you have a secure key distribution channel (e.g., the one time pad). Most of your ciphers in-game are likely to be either alphabetic or polyalphabetic subsitution ciphers or 'book codes'. PC's breaking a code shouldn't happen often without the use of humint (e.g. boosting a code book) unless their cipher capability vastly exceeds that of their opponents (e.g. they're the equivalent of the NSA going up against a 3rd or 4th tier intelligence agency).
Out of curiosity, how powerful computers will you be equivalent to at various levels of superhuman intelligence? Is an INT 30 wizard a calculator, a C64, a modern PC, a Cray or the Cloud?
Somebody with an INT of 18 and an add or two in linguistics could probably break a common polyalphabetic cipher in a language that was at least highly fluent to him with no computer assistance. Folks did, in fact back in the Civil War in the US.
In terms of who's the NSA equivalent in your game, well, that'd be the INT 30 Wizard with full adds, skill focus, and probably some sort of a prestige class bonus (I forget if loremaster adds on that). They're likely to have intricate ciphers in long dead languages available for their communications. You could...
Hmmm... yes, that makes sense. But how high would your Linguistics and INT have to be for you to be able to break PGP, or the NSA's ciphers, without modern-day compute resources?
My linguistics is +25. My companions are skilled planar travelers, alchemists, authors of magical formulae, and world renowned siegecrafters, as befits our high level. I roll middlingly well, 12m but each of my 3 allies contributes +2 to the complexity of the encryption algorithm, for a DC of 43. The message reads "œkkuƒ, ÿ ó∑∑ß åzøµ" Most people won't even recognize half the symbols used, let alone stand a snowball's chance in Osirion of grasping it. Heck, at +25, without the core cypher and my companion's help, I myself only have a 15% chance to crack it, and I'm Intelligence 20 with 20 ranks!!
This does not really make sense. Once your cypher has moved from simple shifts into substitution, it makes no difference what the symbols are. It is just as hard to break "134" as "kuƒ" because you are just mapping this silly character to that real character. In real life, you would be looking at how often some symbol occurs and mapping that symbol to how often some letter occurs in English.
Bad cyphers (like substitution cyphers) use a secret algorithm. For example: Replace A with E, B with C, etc... This means that when your algorithm leaks, you need to make a brand new system.
Good cyphers use a known algorithm with a secret key. For example: Take each 3 consecutive letters and shift, by the key. Then take the result of that and use as the key to shift the next three letters. This allows you to use a single algorithm for your whole organization without worrying about leaks. And since the key is not dependent on your language, you can increase the size of the key to make the encryption arbitrarily strong :) The only caveat is that this only works as long as your key remains secret
In an actual game, I would rule that a moderately good cypher is unbreakable without magical help. Inventing a sufficiently good cypher would take a linguistics check though. In an espionage campaign, your quests would revolve around stealing the cypher keys and keeping yours hidden.
Now, depending on your DM, he might rule that your code is a "language" and then you are boned by a first level spell :p
Hmmm... yes, that makes sense. But how high would your Linguistics and INT have to be for you to be able to break PGP, or the NSA's ciphers, without modern-day compute resources?Somebody with an INT of 18 and an add or two in linguistics could probably break a common polyalphabetic cipher in a language that was at least highly fluent to him with no computer assistance. Folks did, in fact back in the Civil War in the US.
In terms of who's the NSA equivalent in your game, well, that'd be the INT 30 Wizard with full adds, skill focus, and probably some sort of a prestige class bonus (I forget if loremaster adds on that). They're likely to have intricate ciphers in long dead languages available for their communications. You could...
That's a trick question, we can't break that stuff even now :)
Though I doubt that even a very smart wizard would be capable of coming up with public/private key cryptography from scratch. Heck, if he can do that, he should invent an atomic bomb next. :|
I am not sure how to edit my post, but I think I got a 1 on my linguistics roll :p The sample "good" algorithm I gave is actually broken... When encrypting, the trios of letters you shift with the key need to overlap.
So my guess is the best PGP or NSA ciphers that aren't actually unbreakable are probably going to be around DC70, since they'll need to take 20 on them.
I don't think that inventing NSA necessarily requires a 70... There is just a whole bunch of intermediate steps before you realize something like that is even possible. Your wizard would need to know about primes and calculus and stuff like that. I would require at least a hard check in Knowledge(Mathematics) and Knowledge(Advanced Mathematics) and Knowledge(Really Advanced Mathematics) :p
Contact other plane or commune is what most sane wizards would likely do.
I can see it now:
Wizard: Is the first bit of the key 1?
Irori: YES!
Wizard: Is the second bit of the key 1?
Irori: NO!
.
.
.
Wizard: Is the seventh bit of the key 1?
Irori: UNCLEAR!
Wizard: ...
Wizard:Is the parity bit of the first byte 1?
.
.
.
And so on.
I was also thinking along the lines of creating a Knowledge: Cryptography skill.
I'm disinclined to allow linguistics to be used to decipher codes, since spoken and written languages are a very different animal than mathematical code.
If someone was adamant about sticking to the existing skills, however, I'd favor linguistics over bluff, for codebreaking.
|
|
An alternate method here is some sort of magical enigma machine. You could speak the message into it and it enciphers it and adds things like secret page and other page protecting magic. They could work by requiring two to be linked via a physical method (bumping together). Then you just send it via messenger and feed the message into the linked machine which outputs the deciphered message.
Decoding non-magical codes should be opposed linguistics checks.
Code's have usually involved linguistics in part. There are extreme cases such as code talkers.
Ciphers can be deciphered using linguistics.
Non-cipher codes might require a code book. For example, a code where certain arbitrarily determined strings of characters represented certain predetermined messages would require a code book.
R TUVSS RT RVZOOY WVPVNWS WSZT PRNW OU XOWV RT RS TSVN, WOVSN’T RT. RU RT RS Z SUYSTRTUTRON XOWV, ORPV TSRS ONV, TSVN Z YONUS SSOUOW YV TRVVN UOR OONTVR NVSSZTVS. SSORTVR NVSSZTVS ZRV NORV WRUURXUOT YVXZUSV TSVRV ZRV UVWVR WORWS TO TVST TSV XOWV ZTZRNST.
ZNOTSVR UZXTOR NRTST YV TSV NUNYVR OU UZNRORZR WORWS SUXS ZS NZNVS TSZT XOUOW SVOP YOU. TSV ORRTRNZO POSTVR WZS TOUTTONY.
ZOSO, RU YOU PNOW TSV TOPRX R WOUOW TRVV Z YONUS OU +2.
|
Reaper Miniatures
|
My linguistics is +25. My companions are skilled planar travelers, alchemists, authors of magical formulae, and world renowned siegecrafters, as befits our high level. I roll middlingly well, 12m but each of my 3 allies contributes +2 to the complexity of the encryption algorithm, for a DC of 43. The message reads "œkkuƒ, ÿ ó∑∑ß åzøµ" Most people won't even recognize half the symbols used, let alone stand a snowball's chance in Osirion of grasping it. Heck, at +25, without the core cypher and my companion's help, I myself only have a 15% chance to crack it, and I'm Intelligence 20 with 20 ranks!!This does not really make sense. Once your cypher has moved from simple shifts into substitution, it makes no difference what the symbols are. It is just as hard to break "134" as "kuƒ" because you are just mapping this silly character to that real character. In real life, you would be looking at how often some symbol occurs and mapping that symbol to how often some letter occurs in English.
I was intending to use the "silly characters" to represent letters not found in the common alphabet, representing they encryption algorithm using non common-language words. Draconic is most likely not written with the same text as common, as far as the letters go.
Besides, it's a game. If you really want to spend your time making real encryption algorithms, go ahead. Me, I'll roll a d20 and type out "kuƒ" and call it good enough.
|
Reaper Miniatures
|
My linguistics is +25. My companions are skilled planar travelers, alchemists, authors of magical formulae, and world renowned siegecrafters, as befits our high level. I roll middlingly well, 12m but each of my 3 allies contributes +2 to the complexity of the encryption algorithm, for a DC of 43. The message reads "œkkuƒ, ÿ ó∑∑ß åzøµ" Most people won't even recognize half the symbols used, let alone stand a snowball's chance in Osirion of grasping it. Heck, at +25, without the core cypher and my companion's help, I myself only have a 15% chance to crack it, and I'm Intelligence 20 with 20 ranks!!This does not really make sense. Once your cypher has moved from simple shifts into substitution, it makes no difference what the symbols are. It is just as hard to break "134" as "kuƒ" because you are just mapping this silly character to that real character. In real life, you would be looking at how often some symbol occurs and mapping that symbol to how often some letter occurs in English.I was intending to use the "silly characters" to represent letters not found in the common alphabet, representing they encryption algorithm using non common-language words. Draconic is most likely not written with the same text as common, as far as the letters go.
Besides, it's a game. If you really want to spend your time making real encryption algorithms, go ahead. Me, I'll roll a d20 and type out "kuƒ" and call it good enough.
Just realized this sounds snarkier than I intended it. I was really just trying to say that my cipher was an attempt to illustrate what a complex result would look like. It would include mathematical symbols and letters from non-common languages in unusual places. I imagine complex formulae that would replace letter combinations with mathematical, alchemical, or arcane symbols and terms, which is all I was trying to illustrate.
I would also like to point out that I clearly said "encryption algorithm", and not *just* substitution.