Traits + Skills

[Valamir]

Split off from the "Do you believe" thread and the Dreamspire thread is a topic near and dear to my heart.

Traits and Skills.


I think most of us are willing to take as a given that game designers who stick to the Trait + Skill paradigm simply because "thats the way RPGs have 'always' been" need to broaden their horizons.  It is likely that an initial adoption of this design as the automatic default is creatively limiting.

However, there is a fairly common attitude that a game which sticks to a system where both Traits (attributes, etc) and Skills are present is, if not exactly wrong, at least missing an opportunity to do something better.

Many related comments have been made on this topic before, but I think it might be good to get them all in one place for future reference.

I'm thinking that a thread which highlights the pros and cons of the Traits + Skills paradigm as well as discusses the pros and cons of alternative ways to define character capability would be both useful and an enjoyable topic.


To get things started, I'll begin with the traditional Trait + Skill paradigm.  In this you have a list of Traits (attributes, whatever) which all characters have in common, which then is related to a list of skills (proficiencies, abilities, whatever) which are different from character to character.

How the lists are determined and how they are related to each other was for many years the only differentiation between character creation systems.

We've all seen these so I won't define or give a myriad of examples for them all.

What are some advantages of such a design.

1) Trait lists can be constructed to focus player attention on features and attitudes of the game world.  Skill lists (sword vs gun, etc) have always done this, but one of the best examples of focusing attitudes by using Traits is Legend of the 5 Rings.  The organization of the Traits into Rings and way players are encouraged to keep them in balance (at least in the 1ed I'm familiar with) does a great job of mechanically supporting the setting flavor.  

Sorcerer Traits also do this but it isn't really a Trait + Skill system, more of a Trait OR Skill system with only 1 Trait thats very skill like (Lore) and 1 Skill thats extra broad (cover)


2) Familiarity.  While some game "snobs" might look down their noses at this as an advantage, the long time dominance of D&D clones, current rash of d20 products, etc provides ample evidence that gamers are attracted to things that are familiar.  Familiarity at least with the concept (Trait + Skill is a fairly widespread concept) makes learning the specifics of how a given game employs the Trait+Skill principal much easier.  

IMO, unless there is a clear benefit to be gained by an innovation, innovation for innovation's sake just leads to a smaller target market.  In other words unless ones goal is specifically to see how far the envelope can be pushed, "when in doubt, stick to the familiar".


3) Flexibility + Granularity that few other design techniques can match.  When you have a game where the scope of the game is broad there are a wide range of potential character activities.  In a game which focuses on a specific limited range of activities other activities outside of that focus can be discouraged, avoided, or handled very abstractly.  For instance, given Sorcerer's focus the game isn't going to provide much in the way of mechanical support for a couple of non Sorcerous characters having a water skiing contest.  The mechanical resolution for this would be very abstract and this is acceptable because all players should realize how far outside the scope of the game that activity is.

However, sometimes there is a legitimate desire to have a broad scope of activities in a game, and I don't mean the oft heard claim of "well characters could do anything they want".  For example, Dreamspire sounds like it will have a pretty tight focus on setting and overall situation, but within the setting a wide range of activities are possible.  If the desire is to handle those diverse activities with equal granularity than a Trait + Skill system provides a great way of covering a lot of potential situations needing the fewest amount of special rules.


4) Such systems are easy to expand to cover new situations and easy to modify where desired while maintaining the same level of granularity.    Many alternatives to Trait + Skill are equally expandable, so this isn't so much of an advantage as a matching feature.  However, many alternatives that are easily expandable, are so because they have a very low level of detail.  Trait + Skill systems have the ability to set a desired level of detail (low or high) and then extrapolate that same level of detail across.



There are probably some more advantages, but that enough to get started.  I'd love to see this thread continue with what are some disadvantage, what are situations where Trait + Skill isn't appropriate, and for those situations what are the pros and cons of alternatives...i.e. what are you gaining and what are you giving up by using an "alternative" to the traditional system.

[Ron Edwards]

Hello,

First off, here's a thread that's worth reviewing, I think:
Attributes or skills, but not both?

Second, I'd like to clarify something about my argument that most people miss: what I am criticizing is not the concept of an "attribute" or the concept of a "skill," but that they are often separated into mathematically distinct entities - either in terms of how they're "paid for" or otherwise derived, and/or in terms of what they actually do, structurally, in play, in most designs.

Functionally, what happens are breakpoints. The more mathematical derivation, especially using ratios, of skills AND attributes, and how they factor into an Effective Value (target number, e.g.), the more the Effective Values turn into "clusters" of strategically effective points interspersed among a range of "dumb" values.

That's the problem I am isolating. Games like Zero, in which both Brawn and Ranged Combat are featured as equivalent game mechanics entities do not have this problem.

So it's really not "attributes vs. skills" at all. It is, needlessly complex degrees of layering in order to derive an Effective Value, which usually result in strategic breakpoints. I've listed lots of games which use more-or-less attributes like Amber, Sorcerer, and Everway; as well as lots of games which use more-or-less skills, like Castle Falkenstein, The Dying Earth, and Zero. All of these are entirely functional.

Some more traditional attribute/skill systems have solved this problem in interesting ways. In L5R and The Riddle of Steel, the attribute is a number of dice, and the skill is some aspect of using what the dice roll (the "keep" mechanic and the target number, respectively). In Shattered Dreams, one rolls a d12 for skill and a d12 for attribute; the effect is determined as one of four possible outcomes (awesome flub, almost made it, almost missed it, awesome success). I have found all of these to be functional as well.

Best,
Ron

(editing this in) An early thread that's very relevant here is Currency. Please excuse the cranky beginning between Jack and me (we are friends now!) and also note that this thread precedes the big GNS essay; some of my points about Currency are made better in the essay. However, most of my attributes/skills points are made in the thread.

[Valamir]

Certainly have reviewed that thread, but (where it doesn't venture off topic which is pretty early on) it is focused explicitly on applying the issue to narrativist games and I was going for something broader.

Beyond that I think you've combined two seperate issues into the problem you're isolating.

Discussing a system that uses Traits and Skills both vs a system that uses one or the other or something else entirely is IMO a completely different topic from how those Traits and Skills are paid for / rolled for.

Having min max breakpoints is a function of point based character design which has nothing really to do with Traits + Skills as a resolution mechanic.

Further you are defining the difference between Traits and Skills as a mathematical difference in the way they're resolved and pointing to games that have 1 mathematical method as being superior.  We could quibble about exceptions, but I wouldn't disagree with the principal.

But again, I think that is a different issue (or perhaps more accurately a sub set issue).  

The key difference between labeling something a Trait (attribute etc) vs labeling it a skill in almost all systems which make the distinction is that a Trait is something which all characters have in common, while a skill is limited to characters who know that skill.

[Ron Edwards]

Hi Ralph,

If I'm not mistaken, the only trait/skill "problem" occurs when the mathematical hassles I described (and Mike has also described) arise. If said hassles don't arise, due to various design decisions, then all is good.

So I guess I'm not seeing what you're raising as the issue. If we're not talking about (a) Currency/derivation hassles and (b) how the most common concept of attribute/skill feeds into that ... then there's no "over-riding distaste for attribute + skill design" to talk about. Except for the (a + b) that I just described, I don't have or see your stated "... fairly common attitude that a game which sticks to a system where both Traits (attributes, etc) and Skills are present is, if not exactly wrong, at least missing an opportunity to do something better".

For those who aren't following me: if a game does use distinctive traits and skills in its design, and if it does not fall into the Currency traps that Mike and I are both heartily sick of, then ... um, it's cool. It's a good game. I'll happily play it, attributes and skills and all.

I do think that such games can be counted on the fingers of one hand.

Best,
Ron

[Mike Holmes]

Having min max breakpoints is a function of point based character design which has nothing really to do with Traits + Skills as a resolution mechanic.

This is an excellent point (and something I should have included in my thread). I agree that chucking point based creation is a good start to fixing many of these problems. But I also feel that CharGen does need structure. How to give structure without getting the problem back? For example, one could just use HWs Chineese Menu method, but that's efectively the same as the Split Pool method.

I do agree with many of your points in the original posts. I personally find the Stat/Skill split to be very intuitive. I really like the notion of Stats as defaults. That is, roll skill if you have it, else roll the Stat. I've always liked that. In most single level games, traits are so broad as to be nearly universally applicable, or there are times when no trait applies, and you are left rolling a default that is the same as any other character that is undefined in that area.

Interestingly, Story Engine, though it wants to be a single level system, is actually a two level system (Aspects/Descriptors), one in which the "exchange rate" is fixed. I believe that they came back to this (Story Bones does not have it) as they found that desire to have defaults.

Mike

[Valamir]

Hmmm, for some reason this is going in circles.  I wasn't raising an issue to attack or defend, I was presenting a topic for discussion.  My initial post was not a rant against the "anti-trait+skill brigade".

It was an attempt to start a discussion on the pros and cons of different ways of measuring character effectiveness in a game.  I started with a discussion of the advantages of including Traits + Skills in a game simply as a natural place to begin.

Pendragon uses both Traits and Skills in a way where they don't relate to each other and are never used in conjunction.

D20 uses both Traits and Skills in a way where the values aren't related to each other but are used in conjunction

GURPS uses both Traits and Skills in a way where the values are related to each other and are used in conjunction.

Sorcerer doesn't really use Skills at all (Cover is really no different than a Trait with a choice of profession as a descriptor).  All of the scores are held in common by all characters.

The Pool uses "skills" but not "traits" in the sense that there is no common package of traits which all characters share.

What I was looking for is a discussion of the pros and cons of these different measures of character effectiveness, when is one way more appropriate to a games goals than another, when is it just a matter of personal taste, what does a certain alternative do better than another, what does it do worse.

Apparently when I said

I'm thinking that a thread which highlights the pros and cons of the Traits + Skills paradigm as well as discusses the pros and cons of alternative ways to define character capability would be both useful and an enjoyable topic.

I wasn't clear enough in my presentation.

[Le Joueur]

The key difference between labeling something a Trait (attribute etc) vs labeling it a skill in almost all systems which make the distinction is that a Trait is something which all characters have in common, while a skill is limited to characters who know that skill.

In the 'back office' at Impswitch, while we were working on Scattershot, we started referring to traits and skills as abilities.  Stats (Traits as you call them) were originarily compulsory and Skills voluntary.  As work went on, and we began to consider both scope¹ and user-defined abilities, we realized that the way Scattershot addresses them is exactly the same.  Each is simply an ability that your character has a notable rating in.  Further as we began considering formalizing 'infra-Scattershot,' we had to throw out the 'compulsory' part too.

The reason we kept the titles (Stats and Skills) was to attract, or at least be conversant with, 'old school' gamers.  Hidden in the text about character generation is a line that suggests leaving any Stat blank that isn't notable about your character.  Further still, we realized as the scope¹ of an ability went up, the whole system began to approach those 'completely user-defined' games.  In 'infra-Scattershot,' you could actually take a character whose only rating was 'Private Detective,' and then use it per the permutation mechanics for the lesser, but related, abilities.

Since Scattershot has practically no 'trade' between ratings and there are no 'break points' to calculate, it evades the currency issue completely.  One thing that becomes quickly obvious when reading the Mechanix is where we found problems with the 'Stats are the same as Skills (and having both)' style of game.  Or rather, you notice what we left out.

Most of the 'tools' of play employed in Narrativism are completely absent.  You'd think a game featuring Transition would have them, but where are they?  Well, that has to do with the structure of Scattershot and the problem with ratings.  It's all being rolled out as 'techniques;' the mechanix of Scattershot are meant to be usable across all playing styles.  (Ultimately, if I do that right, the mechanix will look 'abashedly BLANK;' fill in the BLANK with whatever style you prefer.)

Now back to what Stats and Skills don't do.  Fortune in the Middle takes things a step out of the 'trap' set by numbers, but only a small one.  For years many role-playing games have given all these numbers, but they never clearly say when to roll and when not to.  I mean do you roll a cooking skill every time you make yourself food?  I don't know about you, but when I cook, I succeed so often that by gaming statistics I ought to be a master chef.  (Oh sure some systems give modifiers for things like time and complexity, but those are patches to cover the problem I'm failing to illustrate clearly.)

The problem is those ratings are included in a game to create a feeling of suspense.  You never know how the die roll will turn out.  That's the point.  But in common situations dicing is ludicrous.  Scattershot include a fair amount of material on the appropriate time to bring out the dice, but even that's limited.  I like the way they handle things in Baron Munchausen; you don't have any ratings, but there's a system that introduces both suspense and impartiality (the other reason I believe dice are used).  Basically, you narrate your tale and failure only occurs when someone 'calls your bluff.'  And instead of pass/fail (as most attribute/skill systems stock and trade in) these challenges only result in complications to your narrative.

No, I know I haven't done a very good job at illustrating when attributes and skills fail, but I hope I've at least moved the conversation away from simply singing their praises.

Fang Langford

¹ Scope, as used here, refers to the breadth of a skill's application.  For example one die roll resolves a sword strike, but in most games one die roll also resolves a hunting trip or a diplomatic mission.  To equate something as simple as hitting someone with something as complex as negotiating a treat misses the whole point of scope.

In Scattershot, if you have a skill in Hunting and none in moving silently, you can 'default' to the Hunting skill with penalties based on the situation because part of hunting requires quiet movement.  Likewise the reverse; if you have an appropriate weapon skill, you can default 'up' to Hunting although it might work better if bolstered by stealth, tracking, and game dressing skills.

Theoretically, you could also default 'down' from something as abstract as 'Politician,' provided that was the common form of play with your group.

[Mike Holmes]

An obvious con to two level systems is that they are more complex, surely. Almost certainly there will be some additional math to be done, often required to be done in play. And certain versions make understandng the odds of an outcome less intuitive, which some people do not like.

These are not things that bug me, personally, but I hear a lot of grumbling about complexity out there all the time.

Also, to extend the argument, it a two level system is good, why not a three, or four, or Nth level system? Why is two seen as optimal?

Mike

[Valamir]

Also, to extend the argument, it a two level system is good, why not a three, or four, or Nth level system? Why is two seen as optimal?

Well, I don't know about optimal, but I think its the default due to the way they're defined.  A trait is something all characters have in common, a skill is something not all chracters have in common.  Unless you do some wiggling that's a pretty binary choice.

As an example of such wiggling, I could concieve of a system where one of the primary goals of play is the cycle of self improvement over the course of several ressurections.  One could then have one set of "spiritual" traits which represent the character's karma, cosmic balance, progress towards nirvana etc.  These traits could then feed into the traditional traits on the idea that you'll come back "better" in your next life if you raise the spirit traits in your last, and then top it off with skills that start fresh with each new life.  voila...three tier system.

Riddle of Steel kinds sorta touches on this in a simple fashion with the Insight Trait, however, I don't believe the Insight Trait is actually used for any purpose during play.

[Paganini]

An obvious con to two level systems is that they are more complex, surely. Almost certainly there will be some additional math to be done, often required to be done in play. And certain versions make understandng the odds of an outcome less intuitive, which some people do not like.

A bit more math yes, but it can be done before play if you are clever. :) Again, D6 is a good example. Attributes and skills are rated with six sided dice. When you give a value to a skill, the skill's value is added to the attribute associated with that skill. So, if you have an Awareness attribute of 4D, and you put two dice in the Search skill, then your search skill is actually 6D. Since attributes are defined at chargen and do not change, you never have to do math during play. You just look at your character sheet and use the skill value if it's listed, or the attribute value if it isn't.

And, this same principle could be used for any system, not just a dice pool one. For example, you could use a numerical rating system that uses attributes / skills as target numbers rather than the number of dice to roll.

Also, to extend the argument, it a two level system is good, why not a three, or four, or Nth level system? Why is two seen as optimal?

It actually isn't, if you talk to people who are hardcore skill fans (that is, people who actually take psychology classes to find out how skills are spread, how people learn, and so on). Those folks tend to like skill tree systems, which are definately N level systems. :)

I think the traditional 2 level system is used simply because it's percieved as the best balance between playability (2 levels are easy to handle) and realism (skill dependancies). Whether or not this is true I don't know.

[Mike Holmes]

A bit more math yes, but it can be done before play if you are clever. :) Again, D6 is a good example. Attributes and skills are rated with six sided dice. When you give a value to a skill, the skill's value is added to the attribute associated with that skill. So, if you have an Awareness attribute of 4D, and you put two dice in the Search skill, then your search skill is actually 6D. Since attributes are defined at chargen and do not change, you never have to do math during play. You just look at your character sheet and use the skill value if it's listed, or the attribute value if it isn't.

And then subtract dice for penalties and bonuses due to situation. Now I'm doing math in terms of thirds of dice in the middle of combat or something. My D6 game didn't last long, and I think that was one of the reasons.  I would have been happier if they had simply trippled the number of dice rolled. In any case, this system is an example of Split Pools (If it weren't it would be a nightmare of min/maxing), which are limiting in CharGen.

All systems suggest doing as much math before hand. Some don't allow that as the character is allowed to back a particular Skill with multiple Stats depending on the situation. This trades off doing more math in-game for increased flexibillity. In any case, Single tier systems don't have that added compexity before or after.

It actually isn't, if you talk to people who are hardcore skill fans (that is, people who actually take psychology classes to find out how skills are spread, how people learn, and so on). Those folks tend to like skill tree systems, which are definately N level systems. :)

Yes, I've used skill trees of varying depths, including infinite. I should have said "seen as optimal by so many" or "are more frequently used" instead of "seen as optimal". Obviously not everyone sees them as optimal.

I'm asking why two level systems are so often chosen; is it a justified approach. Besides tradition, you and Ralph have provided two more reasons, Playability, and Intuitive place of splitting. OK, any others?

Mike

[Bankuei]

My favorite use of Attributes + Skills is DreamPod 9's use in Heavy Gear and Tribe 8...  They justified their view as :  Skill means success more often but talent makes higher degrees of success.  

For example:  The artists that you see do portraits and caricatures on the street do that all day.  They are extremely proficient at producing that style of art in very short periods of time.  You could say that they have a hight skill level.  Another artist is hit or miss, spends months doing preliminary paintings, sketches, half done versioins, and every so often kicks out a masterpiece.  Lower skill, more talent.

This made more sense to me for the reasoning behind splitting the two and their use(as opposed to the straight "add'em together" approach used by most games).

Chris

[Paganini]

And then subtract dice for penalties and bonuses due to situation. Now I'm doing math in terms of thirds of dice in the middle of combat or something.

Hold on cowboy, I'm not talking about Star Wars D6... that's a whole different ballgame. I'm talking about D6 Legend, which is the success / failure dicepool system used in the Xena / Herc game and the DC Universe RPG. You don't ever have to worry about 3rds of dice again. :)

I guess this is getting outside the scope of the thread, but in any case, even in the original D6, the preferred method was to incorporate the situation penalties into the difficulty number chosen by the GM. They included a lot of modifiers and such for people who wanted to use them, but they were very optional.

All systems suggest doing as much math before hand. Some don't allow that as the character is allowed to back a particular Skill with multiple Stats depending on the situation. This trades off doing more math in-game for increased flexibillity.

Yes, this is true. Such a system does require more math during play. However, it does eleminate the breakpoint problem that we were discussing.

Yes, I've used skill trees of varying depths, including infinite. I should have said "seen as optimal by so many" or "are more frequently used" instead of "seen as optimal". Obviously not everyone sees them as optimal.

I'm not actually sure that they are even frequently seen as optimal. It strikes me that it's often one of those "taken-for-granted-elements" being talked about in another thread. People tend to use that model just because "that's how it's done." The folks who go against the system and ignore the two tier system tend to come up with some really interesting and innovative ideas.

So, I'm not sure that there *are* very many reasons beyond tradition. Successful games used it, so everyone else did too. :)

As a matter of fact, I myself tend to prefer single tier systems. Or rather, two teir systems where the "attribute" level is replaced by a default mechanic that's hardwired into the game. When I play I tend to focus on characterization. Having common sets of attributes tends to make characters resemble each other (especially if there are breakpoints in the system). I prefer to focus on what sets characters apart.

[Bailey]

Well, Men of Teak uses stats and skill for the reason of familiarity.  My personal philosophy for designing Men of Teak was to not change much from the norm unless it was important to change.  Of course mine is old school gamist fun that mainly trims off things that get in the way of a the action or promote playing that runs contrary to kung fu action.

[Mike Holmes]

Well, Men of Teak uses stats and skill for the reason of familiarity.  My personal philosophy for designing Men of Teak was to not change much from the norm unless it was important to change.  Of course mine is old school gamist fun that mainly trims off things that get in the way of a the action or promote playing that runs contrary to kung fu action.

How do you get around the currency problem? Is it one of the solutions I mentioned, or have you found some new and innovative way to solve the problem. If so, you can't keep it a secret. The world must know!

Mike