Logical Metas
A ramble about the effect of "Logical Metas" and the their effect on making a game stale:
There are plenty of games where over the course of playing dozens of times, a house meta-strategy on the "sensible way to play" emerges. It might not even be a specific localized strategy referencing the particular players at the table, but a set of common sense assumptions about what does and does not work in a title that becomes self reinforcing as players begin to buy into it.
In Chess for instance, for hundreds of years the game was considered somewhat "solved." The competitive scene still thrived, but there was a short list of pre-defined "correct" openings and counters that every player was taught and used, and skilled play was less about creative improvisation and more about memorizing and executing the correct rote patterns to meet and counter the known patterns employed by your enemy.
But! Once in a blue moon a player like Philidor or Fischer would come along, and just play the game different. Wrong. Ignoring the "correct" plays and breaking the meta wide open, forcing the greater community to recognize that many of the staid strategic assumptions commonly held weren't as rigid as believed.
And now AI has entered the picture, tearing up the scene and showing the world completely wackadoodle yet effective approaches that no one has ever seen in the history of the institution.
In less tactical, more social games metas frequently form that are logically sound, yet self-defeating.
In one group I'd commonly play The Resistance with, it eventually became that no one would ever approve of early missions. The logic was, "The most meaningful information in the game comes from how people vote, and analyzing why. If this mission doesn't get approved, a new mission will get created and a new vote called. Therefore good players should vote 'no' to have as many opportunities to watch votes as possible. Only the evil characters begin the game with information, which means any player voting "yes" for a mission knows something more about it than I do, and has a motive that is suspicious."
While sensible, this becomes complete nonsense the moment it becomes a known expectation at the table instead of a quiet inference. If everyone always votes no, whether to watch more votes or to blend in, there's no value to be gained from having more voting opportunities.
I then joined another Resistance group while traveling where I was taken completely by surprise - the house meta was to always vote "yes" on early missions. Anyone who voted "no" must clearly knew something that everyone else didn't, a perfect inversion of my home meta. It too was self-defeating logic.
Bluffing games are tricky; you can't have common rules of thumb about how to get and interpret information without changing the information you get. Information strategy only works if you have a chaotic mix of people behaving in ways entirely of their own design... At which point good luck interpreting the results!
An awkward and ever-shifting balance must be struck between everyone behaving rationally but not uniformly. Forming threads of meaning and patterns of behavior, sharing enough of your logic to compare notes with other potential teammates, but never saying enough to change how others behave. The more explicit you are about what you look for, the more you change what you see.
I love The Resistance. It gives me a headache.