101
Mafia Game Threads / Re: M117: Emotions Mafia (Day 2)
« on: August 15, 2018, 08:30:05 pm »Not sure about that.I don't think it works quite that nicely here, though it doesn't affect final probabilities that significantly. The 'given that the first role flipped was X and given that the second role flipped was Y' changes things and makes more non-T's more likely.
I think what you're saying is that given the new information that we've seen two PRs in two flips, you think we have evidence that there is a high number of PRs in the game. That would certainly be something to consider if the flips we saw were independent of the roles, but that's definitely not the case.It's possible we've just lost PRs early because they had PR-tells that someone on the scumteam read.
It's still correct that the likelihoods I've quoted are the prior probabilities of each roll appearing at least once in the setup LL has rolled for us.
It's a little Monty Hall-esque in the sense that if there's a bunch of PR's hiding behind doors, you're more likely to stumble upon them when you start opening doors than if there were fewer. If there are indeed only the 4 known PR's, then it's harder to find them first because there are fewer options to be right, as opposed to 5 or 6. And I do believe in independence, certainly for Galz, and close enough for Iguana. In fact, if anything, super-independent for Galz because not claiming makes me lean more towards VT than PR.
Without much simulation experience admittedly, I'd imagine it would be a bit more accurate to assign specific players roles after the dice rolls, and after seeing flips, narrow down the possibilities to those specifically involving 'Player A was X' and 'Player B was Y'. But that does seem very inefficient and cumbersome, and again, it doesn't changes the probabilities a lot.