Ethics: Codes

[Fiasco]

1. Is it OK to encode your role into your early posts? What about encoding the flavor of your role in a theme game? What about just using hints?

2. Is it OK for scum to use a code system to communicate during the day?

There was an earlier discussion here.

The reason why I think coded role claims are problematic is that, while

in theory

the scum could also encode a large selection of roles that they might later want to claim, this takes a lot of work, and in practice will not be done. This tactic robs scum of the ability to let their role claims depend on how the game unfolds. This is especially true of themed roles, just because there are so many more possibilities.

Likewise, letting scum use codes to communicate during the day gives an advantage to those willing to waste a lot of time setting up an elaborate code system. I don't like it when good play requires boring work that has little to do with the game itself.

[Fiasco]

If a player wants to encode to “prove” their innocence when protown, then they have to accept the limitation it places on them when antitown.

So you're saying people should play sub-optimally as town (by not encoding) to help themselves in future games where they are scum? Then we're getting into Type 2 Metagaming again.

[Fiasco]

As mafia, I could encode ten different posts, one with each potentital role claim, and all in different codes.

You could, but it doesn't sound at all like a fun thing to have to do every game. And if the town is smart enough to encourage such encodings, then you do have to encode all your possible claims ahead of time, which is impossible in many theme games. To clarify: I'm saying encoding claims is unethical for town as well as for scum.

Actually, if the town is any good, you can't even encode role claims in ten different posts. They'll force everyone to say in advance what post contains their encoded claim.

[Fiasco]

I disagree with all of you. First off, it's easy for a competent player to encode a claim in an undetectable way. That is the sort of code I'm talking about. Now, it's true that a scum

could

encode ten different role claims in ten different posts, but...

1. Can you expect a scum to be as likely to encode ten different claims as a townie is to encode one claim, given that it's a lot of boring work? I mean, do you know of any examples where scum actually did this?
2. In a themed game or even in a themeless closed-setup game, there are a

lot

more than ten possible claims you might want to make later; often, your options as a scum don't become clear until a few people have died and their names and roles have been revealed.
3. If the town requires everyone to encode their claim in a specific post or even a specific sentence, is it still as easy to encode ten different claims? I think by using e.g. RSA encryption, you can remove any ambiguity about what the encoded message was. (Maybe there are less math-heavy ways of accomplishing the same thing for practical purposes.) This is a game-breaking strategy, people.

[Fiasco]

That's because people aren't using the strategy, not because it isn't game-breaking.

[Fiasco]

You're assuming the mafia could break all codes. Good luck trying to break RSA encryption.

[Fiasco]

Well, I'm not really familiar with the technical aspects of it, but why couldn't people use RSA in a forum game? I'm sure there's some simple freeware that can do it. Then you can post the encrypted message and the public key, and when it's time to claim, you post the private key.

As I said, for practical purposes there are probably ways to achieve the same result without heavy math. Even if that's all impossible, there's still points 1. and 2. to consider.

[Fiasco]

I'm sure there's some simple freeware that can do it.

ta-da

[Fiasco]

If I'm not mistaken, with XOR encryption, you can make up a key afterward that turns the encrypted post into absolutely any message you want. That defeats the point: you want to force scum to commit themselves to a claim early. With RSA, the message is unique (because the public key has a unique prime factorization).

[Fiasco]

That does seem like it would work, but maybe there's some sort of clever way to find a few different meaningful messages that could be turned to the same code using different repeating keys? It seems unlikely, but with RSA you have a full guarantee.

edit: Mith: how long is "long enough"? With just a few lines of encoded text available, is there any chance to break it in practice? I'd guess not.