ToME WikiI've never really been a great fan of the current skill system. Sure, it works, but I've always felt there must be a better way to do things. I'm not sure this really qualifies as a better way, but it might work. 
The idea of having various skills and skill points that the player invests in them seems excessively artificial. I'm not an expert on skill allocations and "game balance" but I shall say this: It involves the player in a direct way in the mechanics of the game (perhaps that's not the best way of putting it, but there we go...) and also leads to certain "dodges" that the player must always use to get their skill allocations to work with maximal effect, or traps them into certain choices (though I admit this depends rather a lot on playing style). This all makes me wonder whether the mechanics of the skills can be "hidden" from the player without leading to a great restriction of choices.
I think it is possible to do this, although I've not yet thought out the details ( 
):
- So far, my idea centers on the idea that the player should be able to increase their skills simply by practice, and is rewarded by certain increases by doing so.
- Skills would still exist in something resembling their current form. There would still be skill levels, which would influence the player's ability, however, this would be hidden from the player.
- Skill levels would be increased when the player gained a character level, and the exact increases would depend on how much the player used the skill, as well as on race and class.
I realise this would be extremely hard to "balance", but IMHO, it would be worth it.
This would be possible in 2.3.x, although it would be rather awkward. It might not lose too much of it's awkwardness in 3.0. :-/
Anyway, I plan to do some experimentation in this area eventually, though I'm rather busy ATM.
What do other module developers (and anyone else ) think of skills at present (pros and cons) and what might be done instead?
-- TheFalcon
Comments?
NeilStevens: DG always wanted an automatic skill raising option, that would spend your points for you.
Generating and allocating points based on player actions could work, too. Just ask for any hooks you need that don't exist yet.
ShrikeDeCil: Two methods in pen & paper games that both worked pretty well were the Chaosium- Call of Cthulu system and the ICE Rolemaster system.
In the Cthulu game, you get 'checks' next to anything you 'exercised' while working on advancing. And you can only advance those skills. (And the marks are erased for the next level).
In Rolemaster, the skills costs adjust in two different ways. The first is that the 'benefit' for 'one skill level' goes down as you get good. That is, if you have _0_ weaponmastery (or whatever), you get a sizable bonus for the first point. Modest increases for the first ten levels invested in the skill. _half_ that much for the second ten levels, a quarter for the second ten levels.... In other words - diminishing returns.
The second method is by allowing a character to 'double invest' on a single level advancement. But that second advancement costs (generally) a whale of a lot more. So a fiftieth level character that invested _every_ time in spell power + magic + mana would have essentially the same skill level in Manathrust as a guy that double-invested 25 times. But one of them 1) got there a lot faster, and 2) spent a lot more of their points on the same skill.
A third level Sorceror (in 2.3.3) can have nearly 7 levels in each of those - and have one overbearingly nice Manathrust already. Using the 'practice' rules, its pretty clear he'd still have checks in all of these three skills. Using the 'penalize the _second_ rank of skill per level' rules, you've already got the player behind the Sorceror thinking about breaking the cookie cutter character. Some (probably me) will just barrel ahead and pay the penalty to get to 'max effective sorcery' faster. Others will start thinking "What _else_ do I want? A little mimicry?" or whatever. And saving up 200+ skill levels just to blow when you get a specific FF quest... will affect things.
Sorry if I'm rambling, but of a long list of pen & paper games, the Rolemaster method was one of the very few that really put a clamp on 'cookie cutter' characters. And the 'Cthulu' method would seem to tie in very well. The only question is for skills that are more difficult to practice. (How do you deliberately practice 'Saving Throw'?, Defense?)
RavenRed: Personally, I like the idea of interdependencies. Currently, the "feed" is all one way (subskills increase superskills). I'd like to see the level of superskill cap the level of subskill. Nothing too ridiculous, maybe a 20-30% Superskill -> subskill relationship. This would somewhat rationalise the spending of skillpoints and guide characters into a particular skilltree. I understand that others may see this as restrictive, however...
NeilStevens: No need to worry whether people find it restrictive. A new module that works differently won't hurt anyone.
HarryErwin: The CoC system is also used in most of the Chaosium games. Checks for a skill depend on exercising it successfully in a stressful situation. The role for increasing the skill depends on INT. This would be compatible with a 100-level character system, with an average chance of increasing being 50%. The CoC system has three problems--how do you get started on a skill?--how do you determine that the skill was exercised under stress?--and--how do you account for quantity of experience? This is because having a check in CoC is binary. I'd suggest allowing basic training (although that's also what FF quests give you) to address the first issue and some sort of risk assessment for the second. The third can be addressed by using a probability of advancement that is reflects accumulated risk. Use a biased logistic function--the bias can increase as the skill level increases (creating diminishing returns), and the logistic function can even reflect a chance of losing skill points if the skill isn't exercised regularly.