doctor's value goes down a lot in this setup because you don't know how many nightkills there will be or how many nightkill prevention roles there are - in a normal game with just 1 mafia faction - cop and doctor - your value is way higher because when you protect successfully its almost as good as an innocent result - here you don't have any idea how successful your protects are and your actions create noise. Doctors r esp crappy imo when there's a possibility of multiple factions because if you prevent a cross-kill you're basically setting the town back.In post 116, Fidget wrote:Doctor is decent but if you only save one person it's somewhat less helpful because MYLO exists. I can think of many other roles that could plausibly be picked instead. It's just less likely that a player would get two high value PRs, if that makes any sense. But Doctor isn't the end-all best PR.
Obv I will pick doctor over a vanilla townie or equivalent - so you know the role I picked is one that I believe is more useful than a doctor or scum - I've already explained why I do not prefer to play scum but if you are going to use the oh he discarded a town PR therefore he's scum - think about this for a second.
1) A player dealt 2 scum-roles doesn't have a choice - he has to be scum - it's just a matter of which scum. He can pick the stronger scum role or the weaker scum role discarding the stronger scum role to set up a wifom and get town-cred.
2) A player dealt Town-PR / Scum has a choice between buffing the town by giving it a PR or adding 1 to the scum count - I think given that there's more town roles than scum roles - it's quite likely most people choose town-PR.
3) Bayesian probability doesn't work in the sense that you say oh he is not likely to be dealt 2 town-PRs therefore if he discarded a town-PR he is scum - you have to incorporate the fact that a town-PR was discarded in the first place into your probability calculation. It's not "what is the probability player X got dealt 2 town PRs" it's "what is the probability that player X got dealt 2 town PRs given the fact that he got dealt at LEAST one town PR, which is basically nearly the same probability calculation as Player X getting dealt a Town PR in the first place.