Dungeons & Design 4: Skills
I keep trying to come up with a well-organized, coherent post for this, and it always turned into a mess, so I'll just make a long, rambling post about all kinds of crap and hopefully you find it interesting and informative. (^_^;)
First, a confession and a note about skills. For a long time, I was...well, biased against a 2nd edition-style implementation of skill systems, where tasks are basically rolls against your attributes and skills (well, proficiencies...) modify the attribute for the purposes of resolution. That's ridiculous, I thought. That makes skills entirely a subset of natural talent, which is pretty much the opposite of the way it works for the vast majority of the population except for a few rare skills that are mostly related to physical activity (gymnastics, for example)! Skills as their own thing is way better. Fie on you, AD&D! Fie!
Then I realized that Warhammer Fantasy Roleplay, which I love[1], does exactly the same thing. Rolls are all percentile under stats, and skills modify the roll, which meant that d20 was closer to my idealized skill-based system than WFRP was. Fortunately, I had enough self-awareness to realize what was going on and think that this proved that maybe a skill-system based on attributes wasn't so bad.
But I still prefer a skill-focused system.
Basic Mechanics
As I mentioned in the post about dice mechanics, I'm currently thinking about using the 2d10 dice pool mechanic to provide a simple method of determining skill level. There would be four different tiers of skill, Apprentice at 2d10, Journeyman at 3d10 keep 2, Master at 4d10 keep 2, and Unskilled at 3d10 keep lowest two. The probability breakdown is here. In addition, there would be a way to buy up to a static +5 bonus on individual skills, with a further +5 available from situational bonuses. Annotation has the bonus first, then the degree of mastery, so "+2M" is roll 4d10 keep 2 and add 2, and "A" is roll 2d10.
So, recapp'd.
Attributes Affecting Skills
Here's the place where I start wavering between a few different options.
The first option is to go the d20 route and tie each skill into attribute, and have the attribute score provide a static modifier to the skills. This is simple and not too bad, but it does have some problems that come in with bounded accuracy. For one thing, it either makes the +5 skill bonus irrelevant due to a large attribute bonus, or makes the attribute bonus irrelevant if it's raised later. There's the possibility of a higher bonus offsetting penalties, but adding up bonuses and penalties before the roll is tedious and annoying, and anyway they tend to be forgotten during play unless they're with +/-2 of the base number anyway. Lots of fiddly adjustments to a roll are a computer gaming thing.
A slightly more complicated option is to come with skills for the attributes themselves. Like, Strength has a Might skill, and Perception has a Notice skill, and Charisma has a Charm skill, that are calculated the same way other skills are, and increasing mastery provides a static bonus to skills based on it. This makes it intuitively easy to determine what to roll for tasks that don't require much special training, like trying to convince someone of something, as well as easy to write down monster stats when they can just default to those attribute skills for everything. The problem there is why even have attributes at all? Why not just say "Strength: +2A" or "Perception: +1J" and so on, and ditch the D&D scheme of attributes entirely? Or just have skills as a modifier to the roll and keep the Apprentice/Journeyman/Master breakdown for the attributes? Just legacy design?
A further option is to go True 20-style and break attributes down to just the bonuses, so a character doesn't have 13 Strength, they have +1 Strength, and then have attributes only add to skills. That skips dead attribute levels (odd levels, in d20), makes it obvious what the effects of attributes are, and attribute checks can just be 2d10 + whatever the attribute is. That makes things intuitively obvious, but means all attributes are just 2d10, so skills are favored whenever they're applicable.
I may be able to draw inspiration from Monsters & Magic, which uses 3d6 and has some modern game design conceits like social mechanics, but claims to be compatible with other pre-d20 D&D games. With dice pools, it'll be easier to hack than basically anything else out there. We'll see how true that is when it comes out, which is supposed to be next month.
Saving Throws
I'm including this here because I had an idea to tie them in to the skill mechanic by treating them like skills. Making Fortitude, Reflex, and Will essentially skills that are advanced like skills will let them keep pace with skills--which is necessary since as the curve of skills rise, the curve of defenses need to rise as well to keep pace with them, and hopefully to beat them in many cases. I'm of the opinion that D&D works much better with 2nd edition-style closed saves than with open-ended saves. The three saves also cover a lot of the stuff that most people would use attribute rolls for. The only thing left is Strength checks, and folding that into Fortitude wouldn't really be amiss. It'd cut down on the skills warrior types need to take,
Breadth of Skills
One of the big development points to worry about when designing a skill system is how broad you want it to be and how discrete the area each skill should cover is. This is a place where I completely break with realism, since no one wants to take "Computer Programming (C)", "Computer Programming (C++)", or "Computer Programming (Objective C)" as separate skills, even if knowing one gives you bonuses to doing the others. A certain amount of abstraction is necessary to avoid getting lost in minutia.
A game like Alternity has a lot more skills (40, plus all the subskills each of those has), but a fantasy game has less of a total number of skills, because there's no laser guns, no computers, no piloting aircraft, no hacking, and so on--at least, not usually, though science fantasy or sword and planet do exist. On the other hand, the AD&D 2nd edition proficiency list does exist and has 229 separate proficiencies. On the other hand, I think we can all agree that stuff like "Winemaking" and "Tea Ceremony" and "Planar Direction Sense[2]" and "Feign/Detect Sleep" are perhaps not suited to be separate skills, and that maybe "Trail Marking" and "Trail Signs" could be folded into a greater "Tracking" skill.
Cheesemaking, though? Totally worth the points.
Without developing a specific skill list (though if there's any interest, I will. So far no one's commented on any of these :p), I think one good thing to do would be to split combat and noncombat skills and come up with a separate way to advance in them. One good thing that D&D and its derivatives do is provide everyone with (varying degrees of) baked-in competence for combat, and you can't run into the problem I've seen in several skill-based games I've run where it's possible, and indeed easy, to make a character who's terrible in fights. That might be good to extend into feats/advantages, too--divide them into combat and non-combat, with separate xp for each. Hmm...
Anyway, weapons skills are best broken down by weapon group, something like Morrowind's skills--bow, crossbow, thrown, long blades, short blades, blunt, axe, spears, unarmed, some other stuff I've probably forgotten. Maybe if "blades" are two skills people will take some weapons other than long swords for once.
Skill Categorization
One thing that Fading Suns (pdf warning) did that I really liked was that it put some skills into a "natural skills" category and gave them to everyone, to represent the kind of things that a person who grew to adulthood would know how to do--charm, dodge, melee, and so on.
So, I'd probably divide skills into three separate categories--Natural, Basic, and Advanced. Natural skills would be (to pull examples from my NEMESIS list): Notice, Persuasion, Lie, Insight, Athletics, and so on, that every player automatically starts with. That prevents the problem that a lot of skill-based games have, where players can easily spread their points too thin and end up with characters who are less competent than they are in real life[3]. Then, Basic skills are skills players need can still roll if they don't have them, but they roll as Unskilled. Stuff like Tracking or Stealth or Search and so on. Advanced skills are skills that cannot even be attempted without buying them, like Ancient History or Witchsight or Alchemy.
This also lets me use skill categories as a way of differentiating cultures or groups, like, "Halflings treat Athletics as a Basic Skill" or "the Dunedain of the North treat Tracking as a Natural Skill" and so on.
That was surprisingly better organized than I thought. I should have just sit down and banged it out earlier. Hopefully it made sense! Next up, I'll take all this stuff and apply it to combat.
[1]: Except for the percentile part, but I mod it into a dice pool system. I <3 dice pools.
[2]: A lot of them are campaign-specific like this, but still, 229?
[3]: This does somewhat depend on the scale used and what kind of task calls for rolls, admittedly.
First, a confession and a note about skills. For a long time, I was...well, biased against a 2nd edition-style implementation of skill systems, where tasks are basically rolls against your attributes and skills (well, proficiencies...) modify the attribute for the purposes of resolution. That's ridiculous, I thought. That makes skills entirely a subset of natural talent, which is pretty much the opposite of the way it works for the vast majority of the population except for a few rare skills that are mostly related to physical activity (gymnastics, for example)! Skills as their own thing is way better. Fie on you, AD&D! Fie!
Then I realized that Warhammer Fantasy Roleplay, which I love[1], does exactly the same thing. Rolls are all percentile under stats, and skills modify the roll, which meant that d20 was closer to my idealized skill-based system than WFRP was. Fortunately, I had enough self-awareness to realize what was going on and think that this proved that maybe a skill-system based on attributes wasn't so bad.
But I still prefer a skill-focused system.
Basic Mechanics
As I mentioned in the post about dice mechanics, I'm currently thinking about using the 2d10 dice pool mechanic to provide a simple method of determining skill level. There would be four different tiers of skill, Apprentice at 2d10, Journeyman at 3d10 keep 2, Master at 4d10 keep 2, and Unskilled at 3d10 keep lowest two. The probability breakdown is here. In addition, there would be a way to buy up to a static +5 bonus on individual skills, with a further +5 available from situational bonuses. Annotation has the bonus first, then the degree of mastery, so "+2M" is roll 4d10 keep 2 and add 2, and "A" is roll 2d10.
So, recapp'd.
Attributes Affecting Skills
Here's the place where I start wavering between a few different options.
The first option is to go the d20 route and tie each skill into attribute, and have the attribute score provide a static modifier to the skills. This is simple and not too bad, but it does have some problems that come in with bounded accuracy. For one thing, it either makes the +5 skill bonus irrelevant due to a large attribute bonus, or makes the attribute bonus irrelevant if it's raised later. There's the possibility of a higher bonus offsetting penalties, but adding up bonuses and penalties before the roll is tedious and annoying, and anyway they tend to be forgotten during play unless they're with +/-2 of the base number anyway. Lots of fiddly adjustments to a roll are a computer gaming thing.
A slightly more complicated option is to come with skills for the attributes themselves. Like, Strength has a Might skill, and Perception has a Notice skill, and Charisma has a Charm skill, that are calculated the same way other skills are, and increasing mastery provides a static bonus to skills based on it. This makes it intuitively easy to determine what to roll for tasks that don't require much special training, like trying to convince someone of something, as well as easy to write down monster stats when they can just default to those attribute skills for everything. The problem there is why even have attributes at all? Why not just say "Strength: +2A" or "Perception: +1J" and so on, and ditch the D&D scheme of attributes entirely? Or just have skills as a modifier to the roll and keep the Apprentice/Journeyman/Master breakdown for the attributes? Just legacy design?
A further option is to go True 20-style and break attributes down to just the bonuses, so a character doesn't have 13 Strength, they have +1 Strength, and then have attributes only add to skills. That skips dead attribute levels (odd levels, in d20), makes it obvious what the effects of attributes are, and attribute checks can just be 2d10 + whatever the attribute is. That makes things intuitively obvious, but means all attributes are just 2d10, so skills are favored whenever they're applicable.
I may be able to draw inspiration from Monsters & Magic, which uses 3d6 and has some modern game design conceits like social mechanics, but claims to be compatible with other pre-d20 D&D games. With dice pools, it'll be easier to hack than basically anything else out there. We'll see how true that is when it comes out, which is supposed to be next month.
Saving Throws
I'm including this here because I had an idea to tie them in to the skill mechanic by treating them like skills. Making Fortitude, Reflex, and Will essentially skills that are advanced like skills will let them keep pace with skills--which is necessary since as the curve of skills rise, the curve of defenses need to rise as well to keep pace with them, and hopefully to beat them in many cases. I'm of the opinion that D&D works much better with 2nd edition-style closed saves than with open-ended saves. The three saves also cover a lot of the stuff that most people would use attribute rolls for. The only thing left is Strength checks, and folding that into Fortitude wouldn't really be amiss. It'd cut down on the skills warrior types need to take,
Breadth of Skills
One of the big development points to worry about when designing a skill system is how broad you want it to be and how discrete the area each skill should cover is. This is a place where I completely break with realism, since no one wants to take "Computer Programming (C)", "Computer Programming (C++)", or "Computer Programming (Objective C)" as separate skills, even if knowing one gives you bonuses to doing the others. A certain amount of abstraction is necessary to avoid getting lost in minutia.
A game like Alternity has a lot more skills (40, plus all the subskills each of those has), but a fantasy game has less of a total number of skills, because there's no laser guns, no computers, no piloting aircraft, no hacking, and so on--at least, not usually, though science fantasy or sword and planet do exist. On the other hand, the AD&D 2nd edition proficiency list does exist and has 229 separate proficiencies. On the other hand, I think we can all agree that stuff like "Winemaking" and "Tea Ceremony" and "Planar Direction Sense[2]" and "Feign/Detect Sleep" are perhaps not suited to be separate skills, and that maybe "Trail Marking" and "Trail Signs" could be folded into a greater "Tracking" skill.
Cheesemaking, though? Totally worth the points.
Without developing a specific skill list (though if there's any interest, I will. So far no one's commented on any of these :p), I think one good thing to do would be to split combat and noncombat skills and come up with a separate way to advance in them. One good thing that D&D and its derivatives do is provide everyone with (varying degrees of) baked-in competence for combat, and you can't run into the problem I've seen in several skill-based games I've run where it's possible, and indeed easy, to make a character who's terrible in fights. That might be good to extend into feats/advantages, too--divide them into combat and non-combat, with separate xp for each. Hmm...
Anyway, weapons skills are best broken down by weapon group, something like Morrowind's skills--bow, crossbow, thrown, long blades, short blades, blunt, axe, spears, unarmed, some other stuff I've probably forgotten. Maybe if "blades" are two skills people will take some weapons other than long swords for once.
Skill Categorization
One thing that Fading Suns (pdf warning) did that I really liked was that it put some skills into a "natural skills" category and gave them to everyone, to represent the kind of things that a person who grew to adulthood would know how to do--charm, dodge, melee, and so on.
So, I'd probably divide skills into three separate categories--Natural, Basic, and Advanced. Natural skills would be (to pull examples from my NEMESIS list): Notice, Persuasion, Lie, Insight, Athletics, and so on, that every player automatically starts with. That prevents the problem that a lot of skill-based games have, where players can easily spread their points too thin and end up with characters who are less competent than they are in real life[3]. Then, Basic skills are skills players need can still roll if they don't have them, but they roll as Unskilled. Stuff like Tracking or Stealth or Search and so on. Advanced skills are skills that cannot even be attempted without buying them, like Ancient History or Witchsight or Alchemy.
This also lets me use skill categories as a way of differentiating cultures or groups, like, "Halflings treat Athletics as a Basic Skill" or "the Dunedain of the North treat Tracking as a Natural Skill" and so on.
That was surprisingly better organized than I thought. I should have just sit down and banged it out earlier. Hopefully it made sense! Next up, I'll take all this stuff and apply it to combat.
[1]: Except for the percentile part, but I mod it into a dice pool system. I <3 dice pools.
[2]: A lot of them are campaign-specific like this, but still, 229?
[3]: This does somewhat depend on the scale used and what kind of task calls for rolls, admittedly.