In general, this is not the case, though. A metagame (“beyond the game”) adds an extra layer—a higher level—that can often (though not always) make the game more interesting. e.g. Trading in Monopoly could be considered a metagame.
I want to ramble about two scenarios today.
Scenario 1: A magic-user is about to drop a fireball on a bunch of enemies. The character doesn’t know that the party thief, who is out-of-sight, will get caught in the blast. The player, however, does know it. The player also knows that there’s a good chance the fireball will kill (or knock out) the thief.
One might argue that the player should have the magic-user choose a different action because: (1) A player shouldn’t knowingly have their character do something that substantially risks the fun of another player. If the thief character were to be killed by the fireball, the thief’s player is going to be out-of-the-game for an indefinite period of time. (2) There are surely other actions the magic-user could take that wouldn’t be out-of-place or out-of-character.
I tend to argue, however, that a large part of the fun of role-playing games for me is having things play out in a “natural” manner. Trying to prevent yourselves from getting caught in a situation like this and suffering the consequences when you do find yourselves in this kind of situation is important
Scenario 2: Black Dougal fails a saving throw versus poison and dies. (This also applies to characters who are otherwise disabled or simply not present at the current scene for some reason.)
What I’ve typically seen in this situation is that Black Dougal’s player is relegated to quietly watching the game since they no longer have a live and present character to speak through.
These days, however, I tend to think that there is no reason that players should not be able to advise other players.
(This post partially inspired by Grimm Studios podcast episode 19.)