Btw on the subject of philosophy.
In a game guaranteed to have no power roles, you know that the only metrics in play to determine the nightkill are based around the town triple threat.
How obvtown are they (how impossible to lynch are they, or similar), how charismatic are they (how easily they sway people, how good their reasoning is, and similar), and how accurate they are.
However, those three metrics can be viewed by three perspectives:
Reputation without prior evidence (that is, hasn't been seen in the game, but know about it);
Prior evidence which is currently not the case;
Current evidence which is obviously the case.
A player who is obvtown on D1, for instance, may be a viable lynch D4, or vice-versa, viable lynch D1 obvtown on D4.
Similarly, some players specialize at charisma early-game; other players specialize at charisma late-game, and everywhere in-between.
But what I really am getting at here is accuracy.
If given the choice between two players that are equally obvtown and equally charismatic--or close enough to warrant it being a tough call--which metric do you use?
If it is by reputation, we can ascertain no further information from the kill, because we have no way of knowing who knows Yuurei's reputation; most players N1 probably had an equal idea.
If it is instead by either current accuracy or past accuracy...ten times out of ten, I am going to go for the one who was in the past accurate.
The reason I am going to go for the one in the past that was accurate is stupidly obvious.
The person who is currently accurate has no way of knowing they are currently accurate. Given time, they will likely change their reads to be inaccurate. Furthermore, because players do NKA, they are going to look at the player's last-stated reads and read in reverse-chronological order. This is the exact thing I want to avoid doing as scum, because it points the town squarely at the direction I am hoping to avoid.
The person who is currently inaccurate but was previously accurate, having shown they were previously accurate, is likely to at some point reevaluate, reset, go back to square one, and revisit their old reads. (I know that one of the first things I do when I am doing a reset is to visit my old processes, and I know this is not at all uncommon a thing to do; it's standard practice.)
Reads I would rather have them NOT be revisiting.
And because they died inaccurate, players who do NKA will be looking at the wrong spot, the end where they were inaccurate.
This is how I would do it, and obviously, my way is not the only way.
My way
is
, however. The best way, and it doesn't take a mafia genius to put together why.
People who believe in NKA look in the wrong spot.
People who don't believe in NKA don't look at all.
The player who proved they
could
be a threat is removed, before they become a threat again.
The player who is currently a threat, via not dieing, thinks that there's a good reason they're alive, and reevaluates to the point where they are no longer a threat.
And so on and so forth. There are a million different reasons why it's the strategy to use.
And yet it only takes one of said reasons for the kill to be the best one to be made.
My point from the ramble on MD theory being.
Most players in the game would, as scum, be smart enough to know that "proof of past threat = need to eliminate, before past threat becomes present again", more or less.
So I stand by Yuurei/Nibbui being killed for their
early
content, not the later content.