So why not lean into it fully and make the GM responsible for the whole balancing?

Because having things balanced properly in regard to the myriad options that are possible in people imaginations is hard, especially related to combat. Improper balacing leads to people having a bad time, while having an established, fair ruleset lets the DM and the players focus on other things.

No need for the framework to do balancing, because a good GM will do that.

But at this point why even have rules? A "good GM" can just entirely improvise a system.

While GM decides what monsters to throw into players, they still need to know what they could use without it being either underwhelming or overwhelming. You dismiss this simply by saying: "just be a good DM".

• New DM's will want guidelines to start from.
• If combat is important having written rules help to use consistent ruling on same situation in different instances.
• Story focused DM might reduce amount of effort needed to plan combat, since there is no need to build it from scratch.

Disadvantage of having to look up rules then you don't remember them could be mitigated by just saying: Look guys, I don't remember ruling now, so not to break the flow, I will rule it this way, and look it up later.

So while for most players rule heavy systems are less accessible, they are actually more accessible for many DMs, and since mastering have much higher barrier of entry, such systems at least should not be dismissed outright.

Balancing is the job of the GM.

And some systems make that job easier for the GM than other systems. Winning all the time without challenge is boring. Getting TPKd every other session does not feel good. A good GM should hit somewhere in-between. So you either have a system that helps you do that or you really need to have a lot of experience.

Yes, the GM balances - they decide what type and how hard the encounters will be. But after that decision is made, it's the job of the system to provide the GM with tools to build that encounter and help me balance things: How much skeletons provide the difficulty I want? Is a lich too much? Red dragon or white dragon?

In 5e, you don't have the proper tools imo - the challenge rating is next to useless. In PF2, you have something akin to point buy for encounters - and if it says the encounter will be "moderate threat" - then you can trust that in 99% of the cases.

But at the end of the day, as a GM, if I want to provide my players with a hard, but fair fight, I don't want to have to guess what will work and what won't. Yes, with a lot of experience I will have an idea of that, but why would I pay for a system that just offloads the hard part of their game design to me? Good encounter-building tools don't get in the way of your creativity.

So why not lean into it fully and make the GM responsible for the whole balancing

Because they should have fun too? Having to rule and improvise everything makes for a harder job for them, needing to keep track of everything to make it consistent, and it's also bad for players too, since they don't really know what to expect.

The framework helps the GM be able to do so, its another tool.

I mainly say that in such a way that if a character is thought to be a pusher the player would know that to push I have to be this close and cant push something thats x times bigger (or some other thing). A GM can (and should) adjust and change things if it would make it more fun for the table but the framework helps understand the world better. for some parts the GM is not directing but explaining what happened.

I push the rock off the edge of the cliff, bar something else, it should fall. The GM at that point is giving the results of that action, what is the result? the GM could simply say that it fell and hurt someone it landed on nearly killing them. the issue comes when the same situation comes up and the GM does something different because they think it should do differently (a different GM more then likely ) this breaks flow if things are different. (assuming all things are the same in both situations for simplicity of course). The Frame work put that a rock fall would deal X amount for how far it fell and the players would have the knowledge (while it would be slightly meta, it would be a "world" known if the rock would deal less damage then the pointy sword XP )

That's what I meant by balance across areas, expectations are known on what some cause and effects are. Frameworks are great ways to help guide things through HOWEVER, a giving framework/gamesystem is not perfect nor a fit all for all game types and tables. A Group also shouldn't need to "go into the weeds" constantly (or at all during a session), something made on the fly or close enough is good to keep things moving.

I mean, pen&paper RPGs aren’t a players vs GM game, but instead the GM plays together with the players to create an interesting experience where everyone has fun.

Full Agree, but the next part is that the Framework helps the GM do the balancing

But if you have the tools that tell you how to make differently balanced encounters, it makes the job of balancing the game waaaay easier.