The quest for a Wiki-less game

Isn't that even more immersion- and game-breaking? Figuring out what kind of attacks have what kind of effect on what kind of enemy is generally part of the game.

* Give good in-game info. Players resort to wikis because they don't feel like they know what they're doing and there's nothing more awful than making a choice you aren't ready for. Roguelikes tend to be varying levels of this; Risk of Rain for example will absolutely give you the exact percentages for its items in the collection menu, but during a run you don't need those percentages to know what you should pick. The bad side of this is The Binding of Isaac, which just has items that do not work at all, for any player whatsoever. Isaac is a game where you open a wiki/cheatsheet just to make sure you aren't picking up a run-ending item. CRPGs are universally terrible at this, something only amplified by their dice roll system. Disco Elysium is the only CRPG I've played that made choices work without pulling up a wiki because it makes it very clear early on that most consequences are yours and it will not lie to you about how far your choices will reach.

* The Wiki Will Not Save You. This is for games like Nethack. Nethack has a wiki. Nethack also has so much depth that a single human will struggle to not just learn it, but might screw up even when they have all knowledge they need. Very difficult to design without alienating your players, but if it works, it's probably the best way to avoid wiki syndrome since there's so much info that most players will just end up mentally arranging the information they themselves need. (Noita I'd consider to be a partial success at this, where the wiki contains all valuable info for achievements without learning Swedish, but the actual spellcrafting and core loop is something with so much depth that every player I've seen try it ends up with their own preferred combos.)

This is one of three things I would time travel to tell my childhood self.

The way that devs determine where the line is for "most people that play get frustrated and give up" and "most people figure it out and keep playing"

It's certainly possible to design a game where the intuition of the player matches a less contrived looking world, bit it's a lot harder. On the other hand, having a convention that climbable surfaces will be blazed yellow, and teaching it early in the game is comparatively easy.

There's also aspects of pacing and focus to consider. A game that ebbs and flows in it's intensity often feels better to play than one that's high intensity all the time. So if your game is mostly about combat, you'll want to break up big fights with lower energy experiences. But those lower energy experiences really aren't "the game" in a sense, they're a kind of filler that makes the high energy experiences feel more fun. It can't be completely boring, but it should be easier than the primary gameplay experience and also have lower stakes. That's a big reason why do many actiony games have blatantly obvious climbing and puzzles. It keeps the player lightly engaged while letting them catch their breath before the next set piece.

It's not the only way to build games by any means, but it is a generally effective, and consistently reliable, template.