Mages can be a secretive and sometimes paranoid group of people. Their knowledge is the key to their influence and power, and it makes sense that Magi would be concerned with protecting their secrets. Encryption is often thought to be a modern concept, but basic forms of encryption were present and effective in ancient times – so much so that it became a common form of sending secure messages.
An excellent historic example of this is the Caesar Shift Decoder (or Caesar Cipher) which is a known substitution cipher, meaning replaces one letter with another, and the scytale which is a physical tool to apply consistent reorganisation of characters. These types of ciphers were simple and effective, as they required a very little knowledge to implement, and made messages very difficult to read. A succinct description is (by blogger Dhaivat Pandya):
Basically, you shift over the alphabet over the message by a certain amount (that certain amount is called the key) to get the encrypted message, and, the person who wants to decrypt it shifts back the message by that certain key.
If simple substitution worked for military messages for the Romans, then it could very reasonably be used by medieval scholars to encrypt and decrypt texts and books. It also makes sense then within the Ars Magica mythic setting that encryption could be used, and also improved with magic.
The potential effects could based upon Rego or Muto magic, with the tasks that a scholar might do in cyphering a text manually but with the added benefit of magical speed and efficiency.
In a previous blog post there was a new spell called Disguise the Words as Written (Muto Imagonem level 5) which temporarily scrambled information in a document for a short time. This scrambling was not intended to be a real form of encryption, rather it modified the text into gibberish, a picture, or almost anything else (which is actually closer to stenography than cryptology). This was intended to provide a degree of security for messages while they were in transit. As an illusion it had the advantage of being a spell effect that ends gracefully, restoring the text to normal automatically. There was no risk of corruption in the process as the text itself was not changed, just the appearance of the text.
The challenge for magical encryption is how to implement that security in a permanent, controllable, and reliable manner. The Muto Imagonem spell above could still be used but it is not ideal, as getting long durations so that the material is protected for a long time becomes a difficult task. Instead a set of specialised spells are needed to act as the encryption and decryption mechanisms. This removes the limitation of time, but introduces a risk of corruption and quality loss when the books are re-written. Magic is not generally reliable.
Rego “craft” magic might be suitable for the actual permanent re-writing of messages, to emulate the work that a person could perform. When Rego is used in this manner it apples to the text itself, and will need casting requirements suitable to the materials involved. Temporary effects are also possible which affect only the image of the text, relying on Muto Imagonem. Intellego should also be a core art for spells which must interpret and translate, which is fundamental to understand hidden meanings.
Aside: I am not sure if the idea of highly complex encryption fits well within a medieval paradigm, but as simple encryption in ancient and medieval times was historically accurate, a hermetic expansion on that capacity seems very plausible within the setting.
The arts and methods of spell design are open to a very wide interpretation, and as with everything produced for Ars Magica YSMV. My assumptions of Muto Imagonem for obfuscation/presentation and non-permanent translation, Rego for materials based work and repetitive work, and Intellego for understanding are core to this interpretation of how to handle encryption with magic.
How is encryption applied in the mythic setting?
A key concern I have is if/how magic would actually process and understand encrypted material. Magic has the ability to translate automatically between languages, between men and beasts, or allow understanding to be found from the dead. These are fundamentally incomprehensible for a mundane person. It may not follow though that an encrypted text is the same as a text in another language as the magic system in Ars Magica 5th edition does not impart any inherent intelligence to magic. Rather it depends on the will, intent, and direction of the practitioners.
The approach for encryption in this article is based upon improving the speed and efficiency of techniques which are currently understood in the mundane realm, instead of introducing modern crypto methods. For example: without intelligence spells must rely on confirmation from the caster when a positive decryption is found. The process is fundamentally the same, repeated hundreds of times, where a cipher key is applied against the cipher text to see if it decrypts to a legible result. This however does not mean that the caster need concentrate on the spell, especially when the effect is doing repetitive processing; which is exactly what decrypting with a cipher key is. This is especially true when the caster knows that they are using the correct decryption method, and just needs to apply the same process in a rote manner.
Introducing encrypted messages in a saga may have story hooks for the characters. As the characters become aware of encrypted material, plots can arise involving stolen or lost messages, the need to decipher hidden secrets, the societal impacts of the “technology” within the setting, and what happens if the new spells technology falls into the wrong hands.
The spell guidelines for the following effects are detailed and then discussed below.
Cypher the Secrets, with the True Cipher of Trianoma
Rego Aquam 20 Herbam/Animal, R: Touch, D: Momentary, T: Individual
This spell encrypts a page of text using a the original Cipher of Trianoma, a basic substitution cipher key. In terms of appearance the crafted material appears as a gibberish reordering of the alphabetic characters.
As the effect is based upon the Rego work based magic, a Finesse check of 9+ is needed to ensure the message is encrypted without introducing mistakes. The effect also requires that appropriate writing materials are present.
(+1 Range to Touch, +3 Bloodline cipher, +2 to alter Finesse check from 15 to 9, +1 Animal req, +1 Herbam req to effect the writing materials)
An example of a spell to decode a text or message using specific fixed cipher which is known and constructed into the spell, where the encryption is rated as simple:
Reveal the Secrets, with the True Cipher of Trianoma
Muto Imagonem 5, R: Touch, D: Sun, T: Individual
This effect temporarily decrypts a short coded message using the original Cipher of Trianoma (see HoH: TL, p20-21). The pages are temporarily changed to appear re-ordered as if the cipher was applied. The Cipher is a secret of special members of the House Bonisagus, so this spell is likely to be closely guarded. This effect also uses the original classic prescribed cipher, rather than a variation of it (as mentioned in HoH:TL), so that messages which use a variation would also not be decrypted cleanly.
This spell could equally be applied to a note or text which was not already encrypted with the Cipher of Trianoma, and would render that page into gibberish. Note that the effect does not actually change the text, instead it alters the image temporarily.
(Base 1 for images only, +1 Touch, +2 Sun, +1 for a simple fixed cipher)
The Cipher of Trianoma as described in Houses of Hermes: True Lineages sounds the same as the Caesar Cipher, a simple alphabetic replacement cipher.
Taking a slightly progressive view it would be ideal to see a spell using a cipher which is specified at cast time, so that many different codes would be used to communicate with different people. The complexity of the cipher is what determines how difficult the message is to decode manually.
A manual cipher key could potentially be as complex as used in the Enigma machine which was created during WWI, although in a medieval setting is is doubtful that such complexity is actually plausible due to the reliance on machine processing and mathematic principals. Medieval encryption could (and did) make use of substitution ciphers, permutation ciphers, and frequency. It was however not at all common for the techniques to be used together. While these techniques are all simple in a modern context, if used in combination in a medieval setting they would form a formidable cipher. Using a combination would also greatly increase the time required to decrypt the messages manually as well.
Magic can assist here by accelerating the mundane processing.
Because a magical effect based upon Rego may create errors, it is necessary to also validate that a text has been encrypted correctly. There is little point in protecting work, if the protection is actually just scrambling the message. One of the effects “Reveal the Secrets,…” above could also be used to decrypt a message temporarily without changing the actual text to allow validation.
Magic effects can also be designed to accelerate the manual decryption of messages when the cipher key is unknown; and the spell below is used to temporarily procedurally decrypt a message. The idea is to have the spell use brute force, a combination of techniques, and a lot of repetition to translate (or crack) the encrypted material.
This is done by creating an illusion which reworks the text using Rego Imagonem and Intellego Mentum. The effect could be reworked to be active for far longer, by extending the duration to Moon or even Ring, which is essentially endless unless disturbed.
Unravel the Secrets, with Unsurpassed Complexity
Rego Imagonem / Intellego Mentum 30, R: Touch, D: Sun, T: Individual
The illusion reorders a handful of sample paragraphs of the text until the Intellego component detects a message which the caster understands. Potential results alert the caster by making the page and ring glow when a potential positive translation is found. The caster can then choose to accept the translation, or have the effect continue.
When a positive translation is found the cipher key method is understood by the caster, and the method can be documented.
Note that the effect is not actually trying to resolve using the any ciphers provided to the spell, but instead is using the repetition and the logic presented in the samples to attempt to find meaning within the text.
Within the duration the spell will continue to attempt to break through the message, essentially allowing the investigation to occur far faster than a scholar could perform, and applying multiple cipher keys. Time becomes the limiting factor for this effect to decrypt successfully.
(Base 1 for image only, +1 Touch, +2 Sun, +2 for the use of complex variable ciphers as well as simple, +1 as ciphers supplied when cast, +2 to scale for complexity of 100 different types of ciphers in sequence, +1 for Intellego Mentum sub-effect, +0 for trivial Creo Imagonem glowing effect)
New encryption techniques based upon magic
When a message is sent via mundane means, and the receiving Magus knows the cipher key and technique (perhaps as a cipher and spell in combination) the messages are relatively secure. The historic methods worked with mathematics and linguistic alphabets as they were the tools available, and one of the key reasons that these techniques were so effective was that they took significant time to decipher. For the Romans it probably did not so much matter for military purposes if a secret message was broken years after it was sent. So the encryption allowed enough security to act as a pragmatic barrier. In the Ars Magica mythic setting other methods and materials could be used as cipher keys for encrypting information.
An initial difference is that the communication of a mundane cipher key can be sent between message participants in many ways which are far from traditional. Hermetic Magi can converse over large distances if they trust one another, which makes the communication of a secure cipher key easier.
It may also be possible to use aspects of Hermetic magic within the encryption technique to provide protection which is impossible to break, such as a limitation for decryption based upon Bloodline. With this faerie variation of an encryption technique only those casters who are also of the same family line would be able to successfully cast the spell to decrypt the message. This method would provide secure communication between parents and children forever. This is even better than using a prescribed cipher key, and might be ruled as impossible to break manually.
Cypher the Secrets, for my Own True Blood
Rego Aquam / 20 Herbam Animal, R: Touch, D: Momentary, T: Individual
This spell encrypts a page of text using the Magi’s Bloodline as the cipher. It may only be decrypted with a spell crafted specifically to decrypt that Bloodline.
In terms of appearance the sigil of the Magi and the related sigils of the bloodline influence the characters which are used to substitute and reorder the original text. The document will look strange and certainly be unrecognisable as a specific alphabet.
As the effect is based upon the Rego work based magic, a Finesse check of 9+ is needed to ensure the message is encrypted without introducing mistakes. The effect also requires that appropriate writing materials are present.
(+1 Range to Touch, +3 Bloodline cipher, +2 to alter Finesse check from 15 to 9, +1 Animal req, +1 Herbam req to effect the writing materials)
Reveal the Secrets, for my Own True Blood
This spell reverses the encryption used in “Cypher the Secrets, for my Own True Blood“, to provide the clear text.
(+1 Range to Touch, +3 Bloodline cipher, +2 to alter Finesse check from 15 to 9, +1 Animal req, +1 Herbam req to effect the writing materials)
The two effects could also be combined into one effect which encrypts and decrypts in one spell, potentially by adding +1 mag for a secondary effect; making the spell final level 25 for both effects.
An alternate way to create encrypted messages for the Magus to magically alter their own actions so that material they create is encrypted automatically. This custom spell affects the mind of the Magus so that they can think and act normally, but any written material they create while under the effect is encrypted with a cipher. This removes the need for an additional spell to encrypt the material after it is created.
Subversion of the Magister’s Mind
Muto Mentum 20, R: Personal, D: Sun, T: Individual
For the duration of the all material written by the target is automatically encrypted with the chosen cipher key. If for whatever reason the target must write clearly and legibly, the task is as difficult as reversing the cipher key on the fly; which should be a very difficult or almost impossible task.
This effect allows for the cipher key to be declared when the effect is cast, and the cipher may be complex in nature. The effect has no impact to material spoken or read, events perceived, or other interaction.
(Base 3 as effect is estimated as equivalent to “Make a major change in a person’s memory of a series of events”, +2 Duration to Sun, +2 to use a complex cipher, +1 to allow cipher declared when cast)
Effect level for this spell could be open to debate. Base 3 was chosen as it should be considered a major change to the target’s mind which is unnatural manner, although through training and practice a similar skill could be developed. The mental adjustment also needs to be made across more than one aspect of the mind, which is why “a series of events” seemed appropriate.
Sun duration was chosen so that the effect can be cast once per day, and then normal lab work can be performed. A shorter version could be invented so that the target touched can create short messages, while the caster concentrates, or for a few minutes only.
Subversion of the Scribe’s Art
Muto Mentum 20, R: Touch, D: Diameter, T: Individual
This effect is functionally identical to Subversion of the Magister’s Mind, which has been altered to affect a target touched for up to two minutes.
(Base 3, +1 Range to Touch, +1 Duration to Diam, +2 to use a complex cipher, +1 to allow cipher declared when cast)
Encryption Spell Guidelines
- +1 mag – A simple (or substitution) cipher inbuilt into the spell, which is not able to be altered. The Caesar Shift Decoder which shifts the alphabet one or more letters forward in a cycle is a good example of a simple cipher. The Atbash cipher was a famous specific shift cipher which swapped the characters where A became Z, B-Y, C-X, etc.
- +2 mag – Using a more complex variable (or permutation) cipher, which is used instead of a simple cipher because it removes the weakness of consistently repeating characters in the encrypted text. The Vigenere Cipher is an example (from 16C) using permutation. These are significantly more difficult to break as the frequency of a character (see frequency analysis) in an alphabet is no longer a basis for decryption. Artes Liberales (Mathematics) 2 is needed for the spell creator, and for the designer of the cipher.
- +3 mags – Using a highly complex (or replacement) cipher, which is used along with a combination of substitution and permutation cipher techniques. Artes Liberales (Mathematics) 2 is needed for the spell creator, and for the designer of the cipher. Anagrams, abbreviations, or even the authors own obtuse shorthand (like within Da Vinci’s notes) might be used in the cipher key.
- +3 mags – the use of a special Hermetic cipher, such as Bloodline.
- +1 mag – A cipher supplied at cast time, which makes the spell design more complex. This cipher must be written in a manner consistent with the spell to be cast, much like the design and floor plan prescribed in Conjuring the Wizard’s Tower. An Alberti Cipher wheel demonstrates the use of a shifting cipher which can be changed easily, and such a wheel would be a suitable artifact for this degree of complexity, where a particular position is set on the wheel and then the spell is cast.
- +1 mag per degree of complexity – to allow for many ciphers to be used in a spell, rather than just one. This guideline provides a scalable limit for the number of ciphers combinations in the casting, +1 for up to 10 ciphers, +2 for up to 100, etc.
- +0 mag – For Rego based encryption/decryption effects require a very difficult Finesse roll vs 15 to translate the text without introducing errors. The severity of the errors introduced and the impact should be adjudicated by the story guide. Botches should be particularly severe.
- +1 to +5 mags for each level of reduction in the target Finesse roll. e.g. To reduce the required target Finesse roll to 12 requires an additional magnitude. +2 mags vs 9, through to +5 mags to reduce the risk totally, where a Finesse roll is no longer needed.
An Example Artifact
The Kryptos Lectern
This stone lectern is made from a heavy single pillar, with a wide inclined tablet affixed to the top. The lectern is fastidiously covered in complex mathematical symbols and formula, with a high degree of symmetry and coordination. The lectern is enchanted with two effects, the first to automatically encrypt all work created on the lectern, and the second to show it decrypted while the text placed is atop the lectern.
Material and Size: Large (4x) soft stone structure (3), inlaid with patterns and sigils. Two effects enchanted as MuMe30 and MuIm14 = 44 effect levels.
A version of “Subversion of the Scribe’s Art” designed as (MuMe: +1 Touch, D: +1 Conc, T: Ind, +1 Simple fixed cipher, +5 for no Finesse roll needed), with device maintaining concentration +5, and 24 uses per day +5. Effect level 20 + device +10 = 30 final enchantment level.
A version of “Reveal the Secrets” to allow display of the decrypted material. (MuIm 4 – Base 1 for images only, +1 Touch, +1 Conc, +1 for a simple fixed cipher), with device maintaining concentration +5, and 24 uses per day +5. Effect level 4 + device +10 = 14 final enchantment level.
Code Breaking without Magic
The understanding of code breaking would be a highly specalised and rare skill in medieval times. Not only was basic literacy and numeracy only available to a small proportion of the population, but the quality of that education was also highly variable according to the teachers who were available.
As time is a major part of manually working through cipher keys, knowing how long a normal attempt will take is necessary to calculate how effective a magical process can be. A specific simple cipher key could be attempted on a small passage of text in about 10 minutes of work by a skilled worker. A complex cipher key takes approximately double this time. This equates to trialing 60 simple or 30 complex cipher keys per day for a mundane worker toiling for 10 hours per day. Magical processing accelerate the attempts by “a days work in an instant” (Covenants p.49), which means a momentary spell can perform a mundane day’s effort, over 600 times faster than normal. A hermetic spell can attempt tens of thousands of cipher key combinations per day.
- Simple encryption (the +1 magnitude guideline) should be moderately difficult to break, and once broken a text will take time to translate comparable to the length of the book. This means after the cipher key is understood a small set of notes might take a day or so, and full summae will be prolonged work over many weeks or months.
- Complex encryption (the +2 magnitude guideline) should be very difficult to break, but the translation time after the cipher key is known is probably very similar to simple encryption.
- The character attempting to break the encryption should have the skill to read and learn through texts, so the same minimum apply from all lab activity: Latin 4, Artes Liberales 1. Almost all magi will have these abilities at Gauntlet. Mundanes would likely require a virtue to access the required skills. Complex encryption raises the requirement of Artes Liberales (Mathematics) 2.
- The Cipher of Trianoma (in HoH:TL p20-21) is a simple substitution cipher noted in the Ars Magica source material, which the Magus Trianoma used to communicate securely and to protect her work.
- Using Trianoma’s example of a simple cipher on short notes uses: Int + Latin + Stress Roll vs. 6 + Author’s Com + Intrigue.
- Keeping this rule set as the basis for all simple ciphers seems fair. However the HOT: TL book notes that translating Trianoma’s cipher on a lab text for a season requires no major additional roll or effort except that already in the base rules (I dislike this, but accept it as a basis in RAW).
So we can assume that a simple cipher adds no more complexity to a lab text than may already be present due to the author’s shorthand, or writing style once the cipher is broken. Complex encryption however must be far more difficult to break.
- Decoding a new document using a very similar cipher adds 3 to the result, but the person must know in advance that the cipher is the same type and similar in pattern, otherwise the bonus does not apply.
- A complex cipher key relies on an understanding of both language, intrigue, and mathematics.
- The skill roll is Artes Liberales (Mathematics) + Int + Stress Roll vs 9 + the complexity modifier
Considerations
When designing the effects I considered the following as impacts to the spell guidelines in terms of complexity and requirements:
- It seems reasonable that a level 20-30 spell effect could facilitate secure communications. The trick is establishing a consistent set of guidelines to keep it from being exploited or becoming broken in the future.
- Modifying a text could be achieved by hand by a skilled scholar, so the core of the spell mechanics would be resolved by Rego magic to mimic mundane work (see Rego craft magic in Covenants p49). RAW indicates that Rego magic must use a Finesse roll to determine how complete and high quality the work is when conducted, and it is reasonable to expect a correlation to the text’s quality.
- Using Rego in this manner adds a risk of corruption when the encryption is permanent. Nobody wants lower quality books, or permanently scrambled messages.
- Using Rego also requires that the suitable raw materials be present which the practitioner would need. This means that cryptographic re-writing of messages will be either creating a new encrypted version, or totally re-writing the initial text.
- Muto magic may also be used (as demonstrated in the temporary spell) to change and potentially decrypt the texts. Muto is excellent for making some impossible tasks possible.
- Muto and Rego in combination are powerful accelerators to mundane work. Adding an extra requisite to the spell might add a magnitude, but it also should greatly enhance the quality and performance of the process.
- Using a preset cipher which is unchangeable in the spell would be easier than designing the spell effect to adjust to a range of ciphers when the spell is cast. A cipher which is supplied at cast time is complex enough that it should make the spell harder to cast, but not drastically harder unless the cipher itself is highly complex. As baseline I think a +1 magnitude to the spell to use a cipher at cast time, rather than always the same cipher which is inbuilt into the spell.
- The guidelines for reduction of the Finesse roll by adding spell complexity represents the fact that additional logic and process can be written into the spell’s execution, so that the chance for error is much lower. It is very deliberate that a “safe” spell is extremely difficult to build.
- The complexity in the cipher is important to the complexity of the spell, and is the measure of difficulty to decrypt the text. A cipher and encrypted text should be difficult to break, and take prolonged work.
- A substitution cipher (eg. replace A>G, B>H) based upon the alphabet might be broken with enough time even in ancient times (ie. Romans used simple substitution ciphers). This is a simple cipher.
- A more complex permutation cipher will be much harder to create and break. Using a permutation cipher adds such complexity that a skill of Artes Liberales 2 is needed for the creator, as a basic understanding of formulas in mathematics is needed. Likewise the same score is needed by the viewer to decrypt it manually.
- A replacement cipher adds complexity in either substitution or permutation cipher, as it alters the presentation of the encypted characters so that analysis is potentially far harder (eg. Order of Hermes might become ooh before and/or after the base cipher is used, or even converted to a new linguistic character entirely). This makes encrypted items very difficult to break.
A last consideration for allowing the Rego based re-writing of messages regardless of encryption is that this also introduces the idea of replicating books using similar effects, and “photocopying books” is generally frowned upon in many of the forum discussions. I disagree that this is a bad thing for Magi to invest energy in, and think it is a logical and valuable goal to strive for. If a player wished to perfect replication in this manner then it seems a worthy character goal, with its own challenges, story implications, and such. Disallowing it because it may cause unbalance ignores the fact that it is a totally logical desire.
Reading and references
- http://www.thedavincigame.com/Code_breaking.html
- http://poincare101.blogspot.com.au/2011/10/encryption-and-ciphers-part-1-this-is.html
- http://www.encrypt.ro/cryptography/medieval-cryptography.html
- http://en.wikipedia.org/wiki/History_of_cryptography#Medieval_cryptography
- http://www.stringpage.com/other/crypto.html
- http://en.wikipedia.org/wiki/Scytale
- http://en.wikipedia.org/wiki/Enigma_machine
- http://en.wikipedia.org/wiki/Polyalphabetic_cipher
- http://www.cs.trincoll.edu/~crypto/historical/alberti.html
- http://en.wikipedia.org/wiki/Al-Kindi#Cryptography
- Thank you to the participants of the Atlas Games official forums topic
As a closing comment I’d like to say that the topic is far larger than I anticipated. The background reading, development work in systematisation, and write-up time was large and it has taken many weeks of casual time to get the post even to this stage. Crypto is a real interest of mine despite no formal background in it, and I’m planning to come back to this topic later as a much larger article. For now going any further is probably over thinking a blog post.
I hope it is useful to others; as it was certainly entertaining to construct.
The Iron-Bound Tome 
Pingback: Protecting spellbooks with Transformation magic | The Iron-Bound Tome
Pingback: Encriptación mágica de textos y mensajes – Leprosería Tytalus
Pingback: Proteger los libros de hechizos con magia de Transformación – Leprosería Tytalus