In post 454, OnTheMark wrote: In post 452, Bronya Zaychik wrote: In post 449, OnTheMark wrote:A person with a track record of being 100% right during that game, gets NK'd, flips town, should be given a lot more weight, regardless of their prior track history.
But a counterargument with that would be the gambler's fallacy: Just because a random player has 2 correct scumreads does not mean that he has the 3rd one correct; what if it's just luck? If a player who always uses an RNG to perform scumhunting somehow had 100% correct reads in a game, would you still trust his reads even after they're dead? With his track record, it'd be pretty easy to tell, but what if we're following your ideal of not using prior track records to judge a player? What then?
It doesn't necessarily mean that the player is correct. However, gambler's fallacy in this case doesn't apply.
Mafia is not a game of luck, it's a game of skill. If mafia was a game of luck it would apply.
However since mafia is a game of skill and logic, the scientific theory applies. Meaning that you have a hypothesis:
Player X (who was NK'd) had good reads in this game.In fact, Player X hasn't been wrong yet. Until there is evidence that is presented to counter them being wrong, they should be listened to. The evidence has been presented that they are right. Evidence therefore needs to be presented that they are wrong before they quit being listened to.
The track records, are a separate matter. In general Player X could be good or atrocious. That's not relevant to this game. However if they are atrocious and are doing the same atrocious things then it WOULD be relevant. However, a track record ALONE is not enough. It has to be combined with a pattern of some evidence in this game. What the problem is that someone goes "Player X has a 75% win rate as town" therefore their reads are right and should be listened to. And that is the real gambler's fallacy as it's not based on any evidence in game.
Take a look again. Because of the very fact that mafia is a game of skill, all the more should track records apply. It's the same reasoning why you'd ask the class's Math genius for help on your math homework instead of the perennial failures: the Math genius has a track record of actually being good in a skill-based subject of expertise, and is therefore more trusted to be correct in a situation where the answers are unknown.
If Player X was an entirely anonymous and random player with no track record, then, sure, the only thing that we can go off to judge his level of skill is purely by his performance in this one game. However, if he is a known jester with a track record of reads that are totally off the mark (pun not intended), then it would be foolish to listen to his reads regardless of his performance in any game, because there is concrete data that Player X is no more accurate than coinflips. There are consequences to making mistakes in this game; it isn't a science experiment (well, usually... I guess?) where we can make mistakes and move on -- there is an active penalty for getting "Evidence to prove them wrong". And, this brings to light another problem: What if Player X is Correct, Correct, Wrong, Correct on his reads? His final correct scumread would quit being listened to, just because he was Wrong once. Worse, Player X goes Wrong, Correct, Correct, Correct.
Your ideals presented effectively prioritize the order of correct reads instead of general accuracy, which would be a flaw that would be corrected if players take into account track records, which, appears to be the current system, so I'm happy with that. The only issue that I may see being presented is a misinterpretation of the track records, for example Player Y is overrated, and is actually worse at the game than advertised, but that's the fault of the enactment in that current situation rather than the system itself. So, I cannot agree with you here.
