Now we’re getting into some meta semantics, and I’m hearing echoes of Pleonast (I think?) yelling at me for playing too conservatively and being too averse to risk as Town in previous games.
And perhaps it is a playstyle difference too, but as Town my experience over time has be conditioned to never completely dismiss any possibility to mitigate just this sort of situation. Things can be logical and still very wrong, and it happens often enough that being too conservative and inflexible can introduce a significant handicap. Being too flexible and illogical can introduce handicaps too, but I really feel quite strongly that the sweet spot is carrying an analytical toolbox with more tools in it vs limiting those tools due to the unpredictable nature of the game.
It was me who’s in the past complained that you were too risk averse. For what it’s worth, I think you’re past that and your play this game was great. Your scum claim is an excellent example of a player taking a calculated risk to further their team’s goals.
I agree that the investigator’s message should have been signed. Especially because I had already demonstrated the utility of signing a message with one’s role name. And I also agree that an unsigned message was an excellent scum tactic. If I had been alive, I would not have acted on the message.
But I think that summarily dismissing the message would also have been a mistake. And for exactly the reason we saw–the message could be accurate despite its obvious flaws. Unlikely, but possible.
My perspective on Mafia is this.
There is a lot of conflicting information.
a. The player has to decided how potentially important each piece of info is. Some things can be ignored, other things can’t. It’s hard to think about everything, so you skim over some stuff. But missing something important is very bad.
b. And the player has to decide how reliable each piece of information is. Reliable info should guide future actions. Unreliable info points to possible scum. There is seldom certainty either way, so the player must be flexible enough to act on info that’s 90% or 50% or 10% reliable. For perspective, consider an investigator whose result is accurate 90% of the time. Do you act on a scum result from them? What if it’s only 50% accurate?
There is a wide range of possible actions. Even for players with no powers.
a. First order actions are straightforward and optimize the player’s actions assuming that the other players are doing straightforward actions as well. (This is complicated by the fact that there is not universal agreement on what even first actions should be.) For example, a protector targeting a claimed power role to prevent scum from killing them.
b. Second order actions are optimized to counter other players who are using first order actions. Players who always take first-order actions are easily countered this way (this is why risk-averse play is ultimately losing). For example, a protector not targeting a claimed power role, since scum “know” the claimed power role will be protected and so target someone else.
c. High order actions try to take into account whether other players are using first or second order actions. There is no general counter to high-order play, except good psychological insight or luck (I use the latter, because I do not have the former). For example, scum targeting a claimed power role despite them possibly being protected, because the protector might be tempted to protect someone less likely to be killed.
Ultimately, each town player must evaluate the alignment of every other player. The only way to avoid this is via death (yours or theirs). Sometimes information (alignment investigations) makes this easy, but this is problematic. Don’t forget that results can be wrong. And don’t expect results.
Eh, not necessarily. At a typical point in a game, there are a couple of people for whom I have an elevated level of suspicion for some reason, a couple of people for whom I have an elevated level of trust for some reason, and a sea of basically-interchangeable people in between for which I’m mostly clueless. All a player really needs to do is to pick out who’s most suspicious to vote for them, and maybe which of two particular players is more suspicious if there’s a close race between them. Ignorance about everyone else doesn’t matter much.
Yes, which is why my next sentence was “The only way to avoid this is via death (yours or theirs)”. You can delay evaluating a player, but you can’t avoid it unless you or they die.
