Nah. Progressing from the weaker stuff to the stronger stuff is the fun of it. Having access to everything from the get go would defeat the point of "building a character".

That would depend on many factors. Some crossbows required immense stength to manually wind up and especially strong people to operate, something your average thief might not be able to do. If a person was not able to deadlift the required winding power of the crossbow (could even 150-200 pounds), the crossbow would simply be unusable. That is why you often see crossbows in games as the ranged weapon for strength based classes. Whereas bow can be released even when not drawn to the max potential.

If it were a crossbow with a winding mechanism and not a manual one, or a low power crossbow, then a Thief could use it. But the Bows=DEX and Crossbows=STR distinction was there for that reason.

I'm not sure who misunderstood (though I'm open to the possibility that I am the one).

OP said: I can't think of any game, where the abilities as listed above (health, dialogue options, "new spells") are dependent from each other. Doesn't sound like he is ok with several steps of the same ability - to me it sounds like he wants as many different (and effective) abilities as possible, in as short a time as possible.

Missing the forest for the trees. If I wanted granular progression, I wouldn't even bring up skill trees in the first place, I'd just do an attribute and skill system that allows one to unlock perks if one meets the conditions that I already do in my homebrew system, a la Fallout (well, pre-Fallout 4). The skill table idea is for games that don't need granular progression but still want some upgrade system. It is based on the premise that your character is already competent at various things, the upgrades simply have them go from basic to advanced.

Crazy thing about Gow original, the game can still be beaten even if you dont upgrade as seen with the fact that PAIN runs exist. Never tried it myself personally but the fact that you technically have access to the tools you need to beat the game from the beginning if you understand the mechanics well enough is very impressive imo.

I dont know how I would learn to even start a PAIN run but it does seem interesting.

I dont think skill trees are objectively bad. Narratively they make sense where characters need to learn basic moves and skills before learning more advanced ones. It makes sense that a mage needs to learn more basic fire spells like fire bolt before stronger ones like meteor or fire golem.

The problem with skill trees is how they are implemented in modern games. Modern games use skill tress as an easy way to lock off certain skills from users by making them be far up the skill tree, forcing players to have to grind or specialize in a certain tree and pick up skills they dont really want or need to get to the ones they want.

A combination of both diverse and near free access to all skills and a semibalance of a skill tree does exist which can be seen in Shadow of War. Players can choose to get any of the skills they want Talion to have and once they unlocked it, they can spend another point to advance the skill by purchasing an advanced variant that provides another effect or adds an element to that initial skill to make it more potent, giving both a sense of progression but freedom to choose to build your character that OP is looking for.

My thought is this: all abilities are not created equally. Some abilities do naturally build upon others. Some skills may be game breaking early on, and while you addressed this as bad game design, I say it is not. If you want a sense of progression, being able to select 2 or 3 abilities from what is normally considered an endgame build would make the game boring to me. If you have to balance the abilities to all be roughly equal, it makes the game potentially boring as well. At least to me. It says my ability choices don't matter. At least in a skill tree, my character is getting stronger in a set of abilities, and requires certain investments to reach mastery in that group. This is not unlike how learning abilities in the real world functions. You should learn the basics before getting the better abilities, before becoming a master of your craft after investing many years in the practice,

Personally, my problem with skill trees is when they are poorly balanced. New abilities in a tree should almost always appeal to users of those skills. I like how Grim Dawn and Titan Quest do it. Most of the late game skills power up the early game skills. You're still using many of your starting abilities at the end of the game. But my issue with those games is that leveling up your abilities feels underwhelming, and even endgame characters feel underpowered. Though that could be by design because the world is harsh and unforgiving.

Difficulty wise, how does that compare to the likes of, say:
* A Final Fantasy 10 No Sphere Grid run
* A Final Fantasy 5 low level 0 ABP run (or the slightly easier no abilities run, where the main difference is not having to go out of your way to avoid ABP)

It sounds very much that the thing you dislike is badly designed skill trees. The way games are designed usually, you have a progression, you need to "progress". You go from weaker at the beginning, to stronger at the end.
A properly designed skill tree won't have INCREASED CHOPPING ACTION as a prerequisite to DOUBLE JUMP. That's not a fault of skill trees, that's a fault of bad design.
Likewise, every skill won't always be exactly the same, some are more powerful than others, and require more XP/skill points/magic knowledge gems, so you end up having a progression anyway.

The example list you gave isn't really a progression at all- it's a list of totally disconnected skills that have nothing to do with each other, and again, that's not how most games are designed. What you've created is a list of skills you can acquire in a rogue-lite/roguelike- meant for short playtimes, doesn't matter if sometimes you get combinations that break the game (or make it impossible), you die and then try again. Which of course, is totally valid, but won't work for other types of games.

Like I said, I havent done the challenges so I cant say. By Final Fantasy no sphere grid, do you mean 10? 12 was license board?

I think it depends but from what Ive read, Final Fantasy 10 you can at least beat the main game no sphere grid and players get stuck at dark anima.

I think GoW 2 and 3 have been beaten at PAIN+ but GoW original is really hard due to the clone fight at the end that some consider just getting to the clone fight as having "finished" the run.

Yeah 10 is good but I think 10 no sphere grid, no aeons, etc just becomes Rikku and Mix spam to beat the game since you can do crazy things like trio of 9999 to guarantee 9999 damage and healing although dealing with aeo attacks is probably very difficult, especially since certain skills will be unavailable to you.

These are also two different types of games (action vs RPG) so they are quite different in that respect as well. Action is probably harder since you need to react and cant plan out like X.

If you think NSG is too easy for that reason, try NSGNSNTQ:
* No using the Sphere Grid
* No summons (except the few times when the game forces you to)
* No Trio Quartet (no using Quartet of 9 or Trio of 9999)

This challenge means that you can't just use Mix to guarantee 9999 damage.

(If you want, you could just ban overdrives entirely, which would also prevent Mix use.)

One other difference between action and RPG: In action, it is, in principle, possible to beat the game at starting stats just by being really good at dodgind and having endless patience (there are a few exceptions, like Crystalis and Ys 2, where being too underleveled can result in you being unable to damage the boss in the first place). In RPG, this may not be possible, as you can't just dodge out of the way, and therefore need to take hits; in cases like FF5 and FF10, it's not entirely obvious that such a feat is possible.

In Dragon Warrior 1, for example, it is not possible to beat the game at too low a level:
* At level 6 or below it's outright impossible. You have no way to avoid the final boss's attacks, will die in one hit, your normal attack will do at most 1 damage, and the HURT spell, even if you get it to work (1/16 chance), is not going to do nearly enough damage.
* At level 7, it is theoretically possible, but is simply not going to happen without RNG manipulation. During the final battle, you need to put the boss to sleep (1/16 chance), then each round, you need to cast HURT, have it work (1/16 chance), hope you get a good damage roll (if not, you might not have enough MP to finish the fight), and then there's a 1/3 chance of the boss waking up and killing you, and you can't have that happen.
* Through level 16, the situation is the same. You might be able to get away with *slightly* less luck, but still don't expect to win in this situation.
* At level 17, you get HEALMORE, and now winning is reasonably possible, though apparently you still have to put the boss to sleep once (again, 1/16 chance), and the boss waking up right away could be a problem.
* Level 18, your stats have increased enough that you can kill the boss with normal luck, and without needing to hit any 1/16 chances.
* Level 20, you're likely to win if you avoid using MP on the way down. At hlgher levels, the final boss eventually becomes really easy (but be aware that Red Dragons could still kill you if you're unlucky; I believe I got killed by one at level 29, and the cap is 30.)

My bad, I should not have sounded so affirmative in my post ;)

The explanation I gave made sense to me because I have been playing Titan Quest recently and its progression looks a lot like a "skill table": skills do not depend on each other to unlock, but you still have some gating by having to put points into the mastery (some kind of level of the skill table itself) to have access to better skills.

In the attached screenshots you can see what looks like skills depending on other skills, but they actually are improvements of the base skill. They do not unlock a new ability but add some effect to an already unlocked one.

Titan quest is still skill trees: You learn base skills, and then upgrades for those skills. They're flatter (non-branching) than most games, but that's all. And many of the upgrades very impactfully change the base skill in different ways.

Ex: If you're playing a Warfare + Earth combination, you often might not want to learn the Flame Arch upgrade to Flame Surge -- because you might not want the fear to trigger, making your close-range/melee-range combat harder. (Of course, you might still go that because of the +projectile which makes it better to "one foe got hit with multiple of them" at close range.)

Fear is one status ailment I don't like seeing in a game at all, as it tends to prolong encounters regardless of whether it's used by the player or by the enemies. It's annoying enough to chase down enemies that I don't want to use it, and when used against me, it takes control away, which is not fun.

Adding fear to another otherwise useful effect is a no-no. (Wizardry 8 has this issue with the Omnigun, which at one point gains the ability to blind enemies, which functionally acts as fear. On the other hand, at least fear in that game only has those severe issues when it hits enemies, as fear on allies doesn't make them run away. (Perhaps they got a lot of complaints from Wizardry 7, where one afraid party member could make your entire party run away, denying you all the XP for enemies you already killed?))