Assume three categories: safe: unknown: mafia.
- Look at days when a mafia was voted out. Who cast the critical votes? Consider them "safe" for the moment.
- Now assume each of those players are mafia. Would it have made any sense for the mafia win probability percentage to vote out one of their own to give the appearance of townie+++.
If the answer is no, consider them "safe".
- If the answer is yes, place them in the "unknown" category.
[*]Okay, this next part is a little tricky, and most people will need to pencil and paper. Look at anyone making a claim today. Is today a do-or-die day (critical) for either side?
If it's not a critical day for the claimants alleged team, does the claim put the claimant at personal risk?
If yes:
What was the claimants win probability (A) prior to claim?
- Now for simplicity ignore a doctor save tonight and , what would it be the claimants win probability for each townie and mafia alignment (B1 and B2) after tonight's nightkill?
- Compare A, B1, B2 to each other to determine safe,unknown,mafia.
- If insufficient data,
assume claimant is true, what was town win probability prior to claim (C)?
- What would be town win probability (D) assuming claim is true, mis-lynch tonight, nightkill and events tomorrow.
- Compare C and D to parse out claimant's alignment.
- If needed rack and stack A, B1, B2, C, D to determine alignments.
[*]If no, then its a stupid-risk for the last mafia to make a false claim. You can still game out per 3.1.1.1-4 for completeness.
[*]If it is a critical day for the claimants alleged team, then evaluate per 3.1.1.1-4 at your own risk and judgement
[*]Is anyone left without an assigned category? Can you use POE to figure out their alignment?
[*]At the least, you should at least know who is safe from this exercise to improve odds of guessing right.