I'm nearly done revising my game On Mighty Thews for release as a for-pay product.  I started the project thinking "Oh, it's already mostly done.  I've just got to tweak a couple of rules, playtest a bit more, lay it out, and I'll be ready to go." How wrong I was.

It's been a bit of a marathon, but the finish line is in sight.  There's one hurdle (hurdles in a marathon?) left.  I think I know how I'm going to deal with it, but I think it's an interesting issue for design, and brings up things I hadn't really thought about before.

In the original rules give characters three "Attributes" which define how well the character fulfills three roles.  I borrowed these from Jason Morningstar's "Dungeon Squad".  They are:

Warrior: Fighting, knowing about fighting, and interacting with military types.

Sorcerer: Magic, book learning, and interacting with magical beings.

Explorer: Physical actions, exotic skills, and interacting with foreign cultures.

Each of these attributes was asigned a die type.

There are a few problems with this system though.  "Explorer" gets used way too much, and "Sorcerer" gets used very little.  They're kind of vague, and sometimes it means you can get away with not describing specifically what the character is doing.

So I tried a new system.

There are four stats:

Thews: Muscle power, agility; running, jumping, climbing, throwing.
Instinct: Reflexes, perception, acting quickly, noticing things, sensing danger and reacting appropriately.
Cunning: Planning, being sneaky.  Detecting falsehoods and getting away with them yourself.  Application of skill.
Lore: Knowing things about stuff.

Stats are rated at 2, 4 or 6.  When in conflicts, add two appropriate stats together and roll a die with that many sides.

This system means you have to describe what the character is doing in more detail so you know which stats to use.  It means there is less often a dominant stat. It also communicates better what's important for characters in the genre (sword and sorcery).  Also, I get a stat called "Thews".

So I playtested the new system yesterday, and it was... ...ok.

I was worried that it would be slower and more difficult than the other system.  It's marginally more complicated to explain, but it's pretty straightforward.  I think it did result in more attention to the specifics of how an action was being performed, resulting in the kinds of descriptions of actions I was looking for.  It wasn't a huge change though.

But there was a new element that I hadn't really considered, and it's kind of hard for me to describe.

Under the old system, the attributes defined your character.  They told you who your character was, in no uncertain terms.  If your highest die was in Warrior, the character was the best warrior they could be.  I think that focused the attention of play away from whether the characters could succeed, and onto what form their success would take.  The question of play is "what will the characters do?" which is what I want it to be.

The new system gives the players more choices about their characters, but leaves those characters less defined.  They are defined more by their actions during play - your characters success or failure validates the choices you made during creation.  We find out if your character has what it takes to survive as we play.  That focuses play on the question of "are the characters fit enough", which is not what I want.

So I like the newer system, and there are some really appealing things about it, but I don't think it's right for the game.

I wonder whether you can combine the two ideas.

Say you start with the three attributes (Warrior, Explorer, Sorceror) and you assign the dice to them. For instance, Warrior gets a d6, Explorer gets a d8, Sorceror gets a d10.

So that gives you the clear "My character is this type of hero" effect that you want, right?

Now imagine a spiffily-designed character sheet that clearly shows each of those attributes feeding into three out of your four traits. For instance:

+ Warrior feeds into Thews, Instinct and Cunning
+ Sorceror feeds into Lore, Instinct and Cunning
+ Explorer feeds into Thews, Lore and Instinct.

(Or whatever combination you're comfortable with.)

Players then assign points into those traits based on the same "2 points into each trait" thing you're proposing. For instance, if you've got a d10 in Sorceror, then you've got five lots of 2-points to assign to either Lore, Instinct or Cunning.

That's what I've got for you. Hope it makes sense.

Perhaps the answer is specializations. Consider Donjon and how each character chooses a "Primary Speciality." You might be very good at talking, or have regeneration, or can cast mighty fireballs.  What if your best class got a specialization and your second best class got a minor specialization.  If I can cast a mighty fireball by taking Sorcery (instead of a dinky regular fireball), I'll probably do that.  I might want to be explorer secondary, or maybe I want to be a war-wizard who can cast Rune Armor. 

Your current problem is the perceived usefulness of Warrior, Sorcerer, and Explorer... because Sorcerer has the coolness factor, so if people aren't taking it, they don't think it's useful.  I can see who people would always want to take Explorer (especially as a secondary stat), you might as well have called it "Adventurer."  (When I look at your 3 stats, I see "Fight, Fight and Adventure Skills") 

As a side note, I would never play a game that had Thews, Instinct, Cunning and Lore as primary stats. (I specifically don't like Donjon's stats, either)

Steve:  Hmm, sounds complicated.  I'm interested in the idea of using good character-sheet design to overcome rules problems though.  I'll have to think about it.

David: It's not really that kind of a game.  Character abilities aren't defined with that kind of precision. 

It's interesting that you mention "Adventurer".  That's exactly what it would be called if that word didn't already have a whole bunch of extremely counterproductive connotations courtesy of previous fantasy games.  The characters in this game are adventurers in the sense that the word is used by everyone except roleplaying gamers.

A problem which you've helped me identify is that actually the Sorcerer attribute is largely about things other than Sorcery.  Sorcerers can't cast "spells" in the sense they do in games like D&D.  This is a swords and sorcery game, so magic is more about forbidden lore, forgotten secrets, and ancient technology, than it is about whiz-bang spells.

David, I don't understand your aversion to the stat names.  Those are the most accurate names I could think of.  Would you prefer that they were less accurate? I can understand being frustrated by Donjon's stats, which are just obscure synonyms for the D&D stats, but these cover distinctly different concepts.  What names would you prefer?

Note the the two schemes are very different.  Warrior, Explorer etc. describe roles the person fulfills; Thews, Instinct etc. describe objective properties of the person.

Hence the difference you observe in what the attributes mean, and the way the new set have less to say about the identity of the character.  I don't think there is an easy way to resolve this; the two schemes are inherently doing different things.

On the other hand, precisely becuase they are doing different things it is likely that they could both exist in some form.  At least, they don't necessarily contradict or invalidate each other either.  I mean, it would be normal for someone wioth Warrior high to also have Thews high, but that would also be plausible in a character that had Cunning high instead, thus distinguishing soldiers from generals.

Alternatey, you might want to reasses the initial problem.  You say that it resulted in less descriptipon, but I'm not sure thats a serious problem.  That is, I see no reason why one of these sachemes should serve better than another in terms of being referenced for description.  It might be more the case that it leaves the character rather undefined, i.e. a warrior character could be a looming giant with a battleaxe or a lithe dancer with a rapier and this would be left purely toi narration.  But, I'm pretty sure it can be left purely to narration.

As for one skill being called on more than another, I don't see this as a problem, it's just a case of the gamble.  Over extended play, too much investment in one will prove costly when you need the other, and undoubtedly the frequency of their relevance will be heavily influenced by the content of the adventures or whatever that the characters are going through.

It sounds like you need to find a neat identity that grabs some bits from explorer into sorcerer. My recommendation is to shift a lot of the exotic skills into sorcerer, expanding it's concept from obscure arcane stuff to magic + alchemical + medical + crafting/problem solving, making sorcerer closer to "medicine man" without the gender connotations.

This then matches up to social divisions in some prehistoric society, missing out only the category of farmer/herder, rather than having magic be some more western academic pursuit. The warrior's the guy you go to when people are nicking your goats or kidnapping your family, the explorer's the guy for helping you get your goats back from falling down a hole, or to take you to a new town to trade, and the sorcerer's the guy you go to when your goat gets foot rot, or when vengeful spirits keep burning your hut down. That's a bit too "homely" for your adventures, but I like to imagine what the day to day elements might behind the big adventures (what would Conan be up to in his own country?).

The result of this will mean a breakdown of the simple physical/mental/social division you have going on underneath the stats, so explorer has less mental/"sit down and do stuff" skills, but is compensated with more physical ones, with warrior sitting in between.

A few thoughts.

[...] sometimes it means you can get away with not describing specifically what the character is doing.

Jasper, that's the most interesting idea I've heard yet.  I'll have to think about that.

To be honest though, the game already kind of has a distinction a bit like that, in terms of there being four different resolution subsystems - Dangerous Conflicts, Competing Conflicts, Fighting Conflicts, and Lore rolls.  Maybe that's enough and I don't need this new thing at all.

Yeah, I don't know how your game works so I can't say what will work for you. I bet this works against you, sometimes, as well.  I can easily imagine my group going "I am a Sorcerer so I can cast a fireball" and myself saying "Actually, it doesn't work that way..." Then choosing something else.

I really like your first 3 stats because all three have a double R sound and 3 constants. They seem like a unified group. The second group sounds arbitrary to me. You might as well have called them "Strength, Dexterity, Wisdom and Intelligence." 

I find "Warrior, Explorer, Sorcerer" tasty. While "Hews, Instinct, Cunning and Lore" are blah.

I'd be careful of making drastic changes to solve "minor" problems. I really believe your (small) problem is the perceived (or actual) usefulness of sorcerer (especially if there's no magic). Instead, make it  seem more useful or make it actually more useful.

I think you're right that the stats aren't the right solution to the problem.  I'll playtest some more to find out.

But! You're wrong about the names of the stats.  It's academic at this point, but you're wrong.  "Thews" is an old fashioned word for muscles, which Robert Howard used a lot in the Conan stories.  The stat covers strength, agility, and endurance. "Instinct" is reflexes and perception. "Cunning" is applied intelligence, as well as skill, and deviousness.  "Lore" is book learning, knowledge, and experience.  This is not the same thing as Strength, Dexterity, Wisdom and Intellience.  The stats a game uses shape what qualities the game emphasises, so it's not just superficial window dressing. 

The stats a game uses shape what qualities the game emphasises, so it's not just superficial window dressing. 

Fair enough.

To me, Warrior, Sorcerer, and Explorer all sound like generalized classes which are pretty much synonyms to the three generalized classes in the D&D 3.x Dungeon Master's Guide.

Just from your description, I would have to guess that every player has these three classes to one extent or another based on what their character concept is.

I would also say that you should get rid of Thews, Cunning, Lore, and Instinct. They do sound like parallels for stats from a billion and one other RPGs.

Though, I do like the idea of utilizing a ranking system with attributes belonging to social stereotypes. In this case, the stereotypes seem to be slightly altered D&D classes.

Based on your descriptions of the first 3 attributes, I could consider them to be Combat & Tactics, Magic & Lore, and *Everything else* & Diplomacy (based off of definition alone).

If your intention is to use classes, I might suggest breaking out of just these three. Introduce some more to make it a bit more interesting and then allow players to choose 2-3 to define their character.

Otherwise, I would name them something more pertaining to what they do rather than who does them <-- this is what gets me and why I don't like your use of "Attribute" in relation to "Roles". Call them roles because that's what they are. An attribute is Strength, Reflexes, or Skin Color.

What is your actual goal here? To define a character based on the role (s) that he fulfills? That's cool, just don't change the terminology because you don't want your game associated with the current crop of offerings.

I would suggest that you change to a role or class system in which you have more than 3 different, specialized, classes (Infantry, Cavalry, Archer, Bandit, Scholar, Alchemist, Thief, Healer, Leader, Bounty Hunter, Assassin, Monk, Naturalist, etc.). Assign 1-3 things that the class is associated with (Fighting, Climbing, Swimming, Forestry, Exploring...the "ing" words). Then allow the player to choose 2-3 of these classes to define what the character is about and might describe the skill and ability selections the player makes for the character.

Other than that, I just see a standard class-based RPG with 4 stats instead of 6.

I've done a lot more playtesting with both ways of doing since I started this post, and I'm pretty happy sticking with the three attributes, Warrior, Sorcerer, and Explorer.  I'm toying with renaming Explorer as "Adventurer", but the word has so many D&D connotations that I think it's best avoided.

'Scholar' or 'Sage', instead of Sorcerer?  Might have connotations closer to what you're describing for the role.