I don't know enough about 'real time' netcode for games. However I have read several HN articles over the years so I've got at least a basic understanding.
Why can't the servers distrust the clients? What should a 'client side anti cheat' actually prevent?
The way I think I'd tackle such things is to have multiple copies of each character model moving in different locations and different ways. Such that trying to spy on the state of the game from one client's viewpoint yields mostly false data. New 'threads' would fork off of the existing threads and would only be culled when there are too many or they're about to make a side effect that would be visible if they were real. In that way the server would be responsible for feeding misinformation to clients but maintaining the state of the true game as a secret to itself.
Why can't the servers distrust the clients? What should a 'client side anti cheat' actually prevent?
The way I think I'd tackle such things is to have multiple copies of each character model moving in different locations and different ways. Such that trying to spy on the state of the game from one client's viewpoint yields mostly false data. New 'threads' would fork off of the existing threads and would only be culled when there are too many or they're about to make a side effect that would be visible if they were real. In that way the server would be responsible for feeding misinformation to clients but maintaining the state of the true game as a secret to itself.