Skill points

[bel]

The XP curve in the game is very weird.

There's also another mechanic in the game: skill points. The curve for the skill points depends on character XL, so it is inherently filled with breakpoints. Also, the formula for dependence on XL is this extremely weird function (the horizontal axis should say "XL" instead of "XP").

The two mechanics also don't gel together. For instance, the XP curve becomes harsher at XL 6, while the skill points curve becomes harsher at XL 10.

[tealizard]

It's true: Skill points are bad. The mechanics of stats are better in general, particularly stat growth. There is no reason skills could not work more or less the way stats do or even be literally implemented as stats. The distinction between stats and skills goes back to the removed concept of "practicing" a skill from Linley's. The current system is merely a streamlined and automated version of the Linley system. I don't think you can justify its particulars or certainly its complexity without reference to its roots in early versions of the game. Ripping it out and replacing it with a system based on stat mechanics would improve the game.

[petercordia]

The curve for skill points actually depends on the total amount of xp invested into skill points, not xp. This difference becomes apparent in wiz-mode, and has minor relevance when comparing species with different XL aptitudes, as well as when looking at the effect of manuals. (Using a manual will slightly decrease the effect of xp later in the game.)
I wouldn't mind if the formula was made more elegant, but if is optimised for good gameplay atm. Any change would have to be playtested and might not be an improvement.

re tealizard: you don't play many wizards do you? Deciding which (& how many) skills to train can be a huge strategical decision. The skill system also allows you to change your strategy mid-game, whereas the stat system greatly discourages switching.

[tealizard]

I have no idea how your comment is supposed to be responsive to mine.

Regarding your theory that the ability to change ones strategy mid-game is a self-evident good, I would liken this idea to saying that Super Mario World should allow the player to have both the fire and cape suits at once. It is good for game features to be mutually exclusive. That is what real decisions are made of. But what I suggest doesn't even go as far as that. One point of some kind of magic skill is one point that cannot go to other skills -- you still have the ability to mix and match, but the decisions are more discrete, weighty, and genuinely strategic. By contrast, crawl skills push the player to get a moderate amount of everything and skill decisions made early mean almost nothing half-way through the game. The specialization of the early game quickly melts into fairly generic strategies.

It would be better if the player were encouraged to build on their existing skill commitments with limited scope for branching into other skills. This would create more varied and challenging play through the mid to late game, as the player would be forced to make do with what they have rather than do whatever they want.

[bel]

I wouldn't mind if the formula was made more elegant, but if is optimised for good gameplay atm. Any change would have to be playtested and might not be an improvement.

Yes, I agree that any change would have to be playtested.

However, I don't know how one can evaluate the statement: "if (it?) is optimised for good gameplay atm". What is the objective function that it is optimizing for? How would one check if it is optimized or not?

From the XL thread I linked in the my OP, that particular function is largely unchanged for about 13 years now. I wouldn't be surprised if this function has stuck around about as long. That is a long time, and it is not even clear what these weird functions are attempting to do.

[bel]

It's true: Skill points are bad. The mechanics of stats are better in general, particularly stat growth. There is no reason skills could not work more or less the way stats do or even be literally implemented as stats. The distinction between stats and skills goes back to the removed concept of "practicing" a skill from Linley's. The current system is merely a streamlined and automated version of the Linley system. I don't think you can justify its particulars or certainly its complexity without reference to its roots in early versions of the game. Ripping it out and replacing it with a system based on stat mechanics would improve the game.

I don't know if skill points are bad or not. The basic idea is that for a fixed amount of XP, you gain less skills if you were lvl 20 instead of lvl 2, which is reasonable on its face. My question is: why choose this curve? What is it trying to achieve?

I don't know what exactly you mean by "system based on stat mechanics". Skills are upgraded every few XLs?

[petercordia]

I wouldn't mind if the formula was made more elegant, but if is optimised for good gameplay atm. Any change would have to be playtested and might not be an improvement.

Yes, I agree that any change would have to be playtested.

However, I don't know how one can evaluate the statement: "if (it?) is optimised for good gameplay atm". What is the objective function that it is optimizing for? How would one check if it is optimized or not?

From the XL thread I linked in the my OP, that particular function is largely unchanged for about 13 years now. I wouldn't be surprised if this function has stuck around about as long. That is a long time, and it is not even clear what these weird functions are attempting to do.

Sorry, I meant to write "it is optimised for good gameplay atm".
What I mean is that for a character who does stuff in order of increasing difficulty, skill point progression feels pretty good.

This might well be due to the xp assigned to various monsters and their stats, which have probably been balanced around the current skill and XL progression curves. Either way, it feels balanced. (And changing it might make it feel unbalanced.)
My only gripe with the current system is that when I do all of D before heading into Lair, all of Lair feels boring. I wouldn't really see that as a skill point progression problem though.

[tealizard]

@bel: Pretty much all formulas in crawl try to achieve reasonable (in no sense optimized) results and the specific forms they take reflect iteration within some model, whether it's dice, a numerical table, or whatever. The answer to this sort of question is a genealogy, not a clear piece of reasoning or design principle.

As for what stat-skills would look like, you would have skill growth choices on level ups, perhaps all level ups. You would make the choice on the spot and that would be it until the next level.

[petercordia]

If I understand you correctly, you could have (almost) the same effect by making the skill-points/skill-level 1 for all skill levels, and decreasing the skill-points/xp you get?

[tealizard]

It would be very different in that there would be no difference in cost between say the sixth and the tenth level of a skill. In particular, the cost of having 2 levels in four skills is the same as having 8 in one skill. This is in stark contrast to the current system where having a large number of skills at 10-12 is cheap compared to, say, a typical conjurations heavy "blaster caster" (lol) type allocation. From an interface point of view, you would spend a lot less time screwing around with skills.

[petercordia]

That's precisely what I said. "making the skill-points/skill-level 1 for all skill levels"
(Sorry if I wasn't clear enough.)

It would be a big boost to 2H weapons, and a big drain on magic users, because they need at least 3 skills to have decent offence (spellcasting, conjurations/summoning & something else). It would also nerf stabbers (who need steath, short blades & throwing/hexes/earth+transmutations).

[tealizard]

The implications of a new system like this in practice goes to the specifics of its design. Obviously, a lot of things would need to be adjusted.