Skill Trees, and related problems.

[Alfryd]

Like it says, I've been trying to come up with an RPG mechanic that revolves around the notion of skills being arranged in a hierarchy, going from vague, flexible and relatively weak abilities that progress down to more refined, specialised, and powerful skills.  A category of related skills would thus form a tree structure, with 'parent' and 'child' skills diversifying to model a character's portfolio of strengths (and perhaps weaknesses.)

To give an example- which will probable be needed to explain the problems I've been having, a straightforward warrior might have the following development of skills. (parent skill is listed in brackets after the child skill name.)

Code Select Expand



Life[]  2
  Body[Life]  4
    Brawn[Body]  7
    Vigour[Body]  5
    Foritude(injury)  6
  Mind[Life]  3
    Artifice[Mind]  6
      Combat[Artifice]  10
        Melee[Combat]  14
          Parry[Melee]  16
            Fencing[Parry]  19
        Proficiency(sword and dagger)[Combat]  13
          Proficiency(rapier)[Proficiency(sword and dagger)]  17
    Knowledge(general)[Mind]  5
      Knowledge(humans)[Knowledge(general)]  9



You can see what I'm getting at here.  Child skills are more powerful than their parents (which can substitute for them in a pinch,) but come at the price of specialisation.  This allows you to effectively abolish the distinction between attributes, skills, attack bonus, etc. etc. and permit the player to specialise solely in whatever set of qualities are best suited for the setting and style of play.

So far, so good.  Where I've been having significant problems is in the area of support skills- skills which aren't direct ancestors or children of eachother, but nonetheless benefit eachother in application.
For example, the above warrior should be able to use his proficiency with the rapier and knowledge of humans to increase his effectiveness in combat- when fencing against human opponents with a rapier.  (This problem is further complicated by the fact that he (obviously) won't get this bonus if he's hunting lions with a blunderbuss, but that's another issue.)  How would you model this kind of synergy?  Let them make several checks and keep the best result?   Add 1/2 skill level?  How do you exclude the effects of common ancestor skills?  Am I just going about it the wrong way?

[Hereward The Wake]

Is it just me but I can't see your example just and empty grey square!
Best
JW

[masqueradeball]

Um, just do it. Have a rule saying that if a character possesses a skill that would be relevant to the situation but is less central than the skill their rolling with, add a bonus based on the rating of the related skill. If this is too cost beneficial (its easier/cheaper to wrack up secondary bonuses than improve the main skill) you might want to make the player pay for these synergizes some how. Like for X points (of whatever) you can add 1/5th of your Knowledge (Humans) to your Proficiency (Sharp Pointy Things) rolls when attacking humans...

Also, I imagine melee here means what it does in D&D, but just so you know, it doesn't mean that outside of that particular game, so maybe there's a better term. Also, "proficiency" smacks of D&D-ness. Of course, that's just my tastes, so feel free to ignore this last paragraph.

[Alfryd]

Is it just me but I can't see your example just and empty grey square!
Best
JW

Strange... what browser?
Life[]  2
  Body[life]  4
    Brawn[Body]  7
    Vigour[Body]  5
    Foritude(injury)  6
  Mind[life]  3
    Artifice[Mind]  6
      Combat[Artifice]  10
        Melee[Combat]  14
          Parry[Melee]  16
            Fencing[Parry]  19
        Proficiency(sword and dagger)[Combat]  13
          Proficiency(rapier)[Proficiency(sword and dagger)]  17
    Knowledge(general)[Mind]  5
      Knowledge(humans)[Knowledge(general)]  9

Um, just do it. Have a rule saying that if a character possesses a skill that would be relevant to the situation but is less central than the skill their rolling with, add a bonus based on the rating of the related skill. If this is too cost beneficial (its easier/cheaper to wrack up secondary bonuses than improve the main skill) you might want to make the player pay for these synergizes some how. Like for X points (of whatever) you can add 1/5th of your Knowledge (Humans) to your Proficiency (Sharp Pointy Things) rolls when attacking humans...

Yeah, but... consider the case of Proficiency(sword and dagger) boosting Combat rolls with that weapon(s).  They have 4 ancestor skills in common, which accounts for about half the total skill level, so adding a 20% bonus is like saying the skill level is, at minimum, 10% higher- once you take the skill.
I could add a rule that says 'exclude mutual ancestor skills', but that seems a little tedious to work out on the fly.  I could also work around it by divorcing the proficiency and knowledge skill trees from the main trunk, so to speak, but that might not work with all cases.  I mean, consider the Jump[Brawn] and Gymnast[Artifice] skills, which both demand active use (so I can't just split them off the main trunk,) but should probably support eachother in some fashion.

Also, I imagine melee here means what it does in D&D, but just so you know, it doesn't mean that outside of that particular game, so maybe there's a better term. Also, "proficiency" smacks of D&D-ness.

Well, I come from a largely D&D background, so, yeah.  I have no problem with different terminology if you think it's less confusing.

[Lance D. Allen]

Least important comment first: Screw terminology. Explain what you mean, and worry about what to call it later on in the development process. Just go with what's easy.

Secondly, if you structure your skills a little bit differently, it might get rid of the problem. Try something like this:

Life[]  2
  Body[life]  +2
    Brawn[Body]  +3
    Vigour[Body]  +1
    Foritude(injury)  +2
  Mind[life]  +1
    Artifice[Mind]  +3
      Combat[Artifice]  +4
        Melee[Combat]  +4
          Parry[Melee]  +2
            Fencing[Parry]  +3
        Proficiency(sword and dagger)[Combat]  +3
          Proficiency(rapier)[Proficiency(sword and dagger)]  +4
    Knowledge(general)[Mind]  +2
      Knowledge(humans)[Knowledge(general)]  +4

So this way, you're fighting with proficiency rapier (Life(2) + Mind(1) +Artifice(3) +Combat(4) +Sword and Dagger(3) + Rapier(4) = 17) against a human (+4) for a total of 21. (fencing and proficiency rapier are kinda redundant, unless there's something I just don't get) If you did add fencing it would be 24.

Now maybe instead of adding the full 4 points of any synergy skills, you could halve them, so in the above case it would be a total of 19.5 (round up or down) with knowledge: humans and fencing thrown in.

[Adam Riemenschneider]

Hello Alfryd,

Well, you have a lot of options, depending on how you want the setup to work.

a): You could, as has been mentioned, allow the two to come from the same ancestor areas, and roll both as full Skills. This will require you to have a sub-mechanic that spells out what level of success on the knowledge aspect means as a form of bonus on the melee aspect.

b): You could separate the trees so that they *dont* share ancestors. Following this option, if I were in your position I would base combat skills off of a physical Parent. You will still need to work out a sub-mechanic, as in "a," so players will know what kind of bonus they get.

c): You could make specific Knowledge a kind of Specialization/bonus that is bought separately. The real plus with this option is that it will reduce the amount of rolling required (no need for a sub-mechanic). You also don't have to worry about ancestor skills or parent skills. Also, your Knowledge (humans) bonus of +4 (or whatever) could then readily be applied to other Skill trees, without having to worry about worming it into the actual tree structure for each one. "First Aid check final total 15, with my +4 Human Knowledge bonus, is 19. What's my difficulty on this roll?"

Good Luck!

-a-

[Alfryd]

(fencing and proficiency rapier are kinda redundant, unless there's something I just don't get)

Well, in theory you could attempt to fence with a stiletto or broadsword (you'd just suck at it,) and throw or cleave with a rapier (at which, again, you'd suck.  But less so than if you weren't familiar at all with the weapon.)  But you do raise an interesting point.

So this way, you're fighting with proficiency rapier (Life(2) + Mind(1) +Artifice(3) +Combat(4) +Sword and Dagger(3) + Rapier(4) = 17) against a human (+4) for a total of 21. ...If you did add fencing it would be 24.

Very true, but since you don't incorporate ancestor bonuses directly, you now have the computational hassle of adding up a half-dozen subskills for every check.  I have no intrinsic problem with this, but I've had complaints about the system being overly complex, so I was hoping to streamline.
(On the other hand, the example represents a moderately advanced warrior, by which time the player would (presumably) be used to adding up subtotals.)

In your experience, would you be more or less comfortable with doing this kind of math pretty often?

Now maybe instead of adding the full 4 points of any synergy skills, you could halve them, so in the above case it would be a total of 19.5 (round up or down) with knowledge: humans and fencing thrown in.

That, or allow direct ancestors to count double, whichever is less trouble.
However, this results in *very* generous bonuses for support skills that don't have much in common- such as Jump[Brawn], Gymnast[Artifice] and Balance[Acuity(tactile)].  Which might actually be fair, but it should be borne in mind.

c): You could make specific Knowledge a kind of Specialization/bonus that is bought separately. The real plus with this option is that it will reduce the amount of rolling required (no need for a sub-mechanic). You also don't have to worry about ancestor skills or parent skills. Also, your Knowledge (humans) bonus of +4 (or whatever) could then readily be applied to other Skill trees, without having to worry about worming it into the actual tree structure for each one. "First Aid check final total 15, with my +4 Human Knowledge bonus, is 19. What's my difficulty on this roll?"

Hello Adam.  I agree, by and large, but a few points-
1.  I probably still have to worry about parent and ancestor skills, since knowledge is just as hierarchical as any other field of expertise.  Science ->  Biology ->  Medicine, for example.
2.  This means this would be similar to divorcing the knowledge tree from the main trunk, so to speak.  Which is fine in this case, by and large, since knowledge doesn't have a lot in common with the combat tree regardless.
3.  I'm hoping to keep the skill tree as unified as possible, so that characters who haven't expressly trained in, say, Ornamental Basket-weaving[Weaving], can still have a reasonable proficiency at the task by substituting appropriate ancestor skills.

But what I'm mainly worried about is cases like the Proficiency subtree, where there are a lot of mutual ancestors.  I'm wondering if the simplest thing to do would be to actually limit performance in a given skill to 2x the level of all 'support' skills.  i.e, if your Knowledge(humans) is only 9, your skill level when fighting them can't exceed 18.  It's easy to apply, but it might be a little harsh.

[Velcanthus]

Like it says, I've been trying to come up with an RPG mechanic that revolves around the notion of skills being arranged in a hierarchy, going from vague, flexible and relatively weak abilities that progress down to more refined, specialised, and powerful skills.  A category of related skills would thus form a tree structure, with 'parent' and 'child' skills diversifying to model a character's portfolio of strengths (and perhaps weaknesses.)

I think that the first order of business is to try and work out what you want this system to do. Are you looking for something that will naturally focus a character's development as they advance? Are you, perhaps, looking for something that is an understandable model of character progression? Are you looking for a system that is quick and easy to administer?

[Eero Tuovinen]

Unless I'm missing something, you just need a suitable resolution mechanic that can factor support ability bonuses fast and easy. Then you can just go case-by-case while playing, picking any abilities that suit the situation, pick the most important one as primary (if you want to have that distinction, that is) and check the others for bonuses to the primary. That's a simple, fast and flexible system that's also difficult to unbalance. The Shadow of Yesterday, incidentally, works like that.

Of course, picking the right resolution mechanic might be difficult, especially if you have some other requirements for it to fulfill as well. For example, if you're doing percentage roll-under, this is all very easy: just roll the support abilities and halve the chance of failure for each successful support. But it all depends on how your exact resolution mechanic works: how those ability scores are used to determine success and failure?

[Lance D. Allen]

My example was admittedly somewhat simplistic. You wouldn't have to add up the score for Rapier every single time.. You'd have the current total for each subskill listed next to it's bonus.. So it'd be more like this:

Life[]  2
  [...]
  Mind[life]  +1 = 3
    Artifice[Mind]  +3 = 6
      Combat[Artifice]  +4 = 10
        Melee[Combat]  +4 = 14
          Parry[Melee]  +2 = 16
            Fencing[Parry]  +3 = 19
        Proficiency(sword and dagger)[Combat]  +3 = 13
          Proficiency(rapier)[Proficiency(sword and dagger)]  +4 = 17
    Knowledge(general)[Mind]  +2 = 5
      Knowledge(humans)[Knowledge(general)]  +4 = 9

Then you choose the primary (for my previous example I chose Proficiency Rapier) and add the individual sub-bonuses of each applicable support skill.

Also, to expand upon my comment about the redundancy of rapier and fencing.. If you're proficient with a rapier, you're not going to be proficient in using it in ways it's not meant to be used. You won't have learned how to hack with it like it were an axe, or bash with it as if it were a club. You will be proficient in using it as though it were a rapier, which is fencing. So, my personal, outsider's perspective on this is that you should have one or the other, but not both.

Now what does this mean if you have the proficiency rapier, but you use a broadsword? In that case, you'd just use your sword and dagger skill, assuming that this refers to skill with both swords and daggers (slicey-stabby implements) rather than a particular combat style utilizing both a sword and a dagger. If that assumption is wrong, then you'd just use combat.

[daeruin]

You might try something fairly simple like counting the number of related skills that apply, with higher numbers of skills contributing a diminishing amount to the current roll. A single related skill adds so much, two skills add a little more, and so on, without considering how related the skill is. Very simple. Not sure if that's what you're looking for, but I thought I'd throw it in there.

[masqueradeball]

would my synergy bonus model work if you included both the total and the amount that particular aspect was raised? For example Knowledge (Humans) could be listed as +9/+6. 9 is the skill level when used alone, and 6 would be the synergy bonus when used with another Mind Skill, or, if you want synergies to be less effectual, that second number could be halved. So it would be listed as +9/+3... this seems simple and intuitive, is there something I'm missing that wouldn't allow it to work?

[Alfryd]

Then you choose the primary (for my previous example I chose Proficiency Rapier) and add the individual sub-bonuses of each applicable support skill.

Ah, thanks.  That's much clearer now.  I'll have to consider the issue of proficiencies/weapon styles a bit more later.
I think, on balance, your suggestion is probably the most viable at present.

For example Knowledge (Humans) could be listed as +9/+6. 9 is the skill level when used alone, and 6 would be the synergy bonus when used with another Mind Skill, or, if you want synergies to be less effectual, that second number could be halved. So it would be listed as +9/+3... this seems simple and intuitive, is there something I'm missing that wouldn't allow it to work?

I suppose, yes.  There's still the problem that common ancestor skills grant too much of a benefit.... eh.

I think that the first order of business is to try and work out what you want this system to do. Are you looking for something that will naturally focus a character's development as they advance? Are you, perhaps, looking for something that is an understandable model of character progression? Are you looking for a system that is quick and easy to administer?

Of course, picking the right resolution mechanic might be difficult, especially if you have some other requirements for it to fulfill as well. For example, if you're doing percentage roll-under, this is all very easy: just roll the support abilities and halve the chance of failure for each successful support. But it all depends on how your exact resolution mechanic works: how those ability scores are used to determine success and failure?

My main concern was, for wont of a better word, realism.  I wanted to use skill hierarchies and support skills because it's a natural mimcry of how skills actually develop and interact.  So, I want to keep that if remotely possible.

By extension, I'd also like to be able to handle scaling (i.e, high-level skills and gaps between skill levels) reasonably well.  That is, a contest of skill level 20 vs. skill level 50 should have (roughly) the same odds as skill level 4 versus skill level 10.  I've considered a mechanic where you multiply your skill level by (2d6-2) (higher is better) to obtain a check result, but the math gets a little tedious for higher skill levels.  Another possibility is to roll and add different dice depending on your primary skill bonus.  I hadn't considered a percentile roll-under system, but on balance I'd like a more clumped probability distribution.

Handling scale is also fairly important here because of the synergies from support skills.  Because support skills can rise in proportion to the primary skill, any fixed range for dice rolls will eventually be overwhelmed by bigger and bigger bonuses.

I have no problem with using multiple rolls if that's less hassle than totting up situational subbonuses.

[Velcanthus]

My main concern was, for wont of a better word, realism.  I wanted to use skill hierarchies and support skills because it's a natural mimcry of how skills actually develop and interact.  So, I want to keep that if remotely possible.

It seems to me that this model is not realistic because each member of the taxonomy contributes its full point value.

Take, for instance, fencing and rapier skills. Fencing is a sport based on the use of light swords, rapier is a light sword. In the real world, I doubt that your ability at fencing adds anything to your ability with a rapier. If you didn't have the rapier skill, or if you were using a light sword that wasn't a rapier, then one might see some advantage.

[Alfryd]

Wolfen has suggested that seperate skills for fencing and rapier proficiency may be redundant, but that can probably be fixed.

What I wanted to redress was the problem that, in many RPGs, two closely related skills/attributes are essentially treated as largely independant.  For a simple example, consider Strength and Constitution in D&D.  In realistic terms, it is extremely unlikely that a person could develop high strength (muscle mass through exercise) without substantially improving their constitution in the process (organ capacity, aerobic fitness.)  In GURPS, there's an informal solution using... what are they called, talents?- well, specific skills that boost all checks from a closely related group- 'smooth operator' for social skills, machine expertise, etc.
But then it occured to me that the simpler thing to do would be to abolish the distinction between skills and attributes entirely.  Any skill could have any other skill as a 'key attribute'.  (Maybe even several, if it's performance was dependant on/benefitted from multiple factors.)