Roll 3d6 .... What is this?!?

[Nathan P.]

To condense my understanding of the thread thus far...

We ask "What is the goal of your game?". You answer "Realism".

This is like saying "What do you want to study in college?" and you answer "Liberal Arts."

Well, ok. Fair enough. But if I happen to be a historian, I don't have any reason to help you out even if you happen to want to study history.

You need to narrow it down for us some more to get some valuable input. That is, (setting aside the question of what a realistic model even means), what is it, in the theoretical game that you're talking about, that you want to model realistically?

[John Kim]

Let me suggest something: What really matters when we're simulating situations in a game is not so much how many factors and real-life physics we include, but that the outcome seems plausible. When I fire a gun at someone's chest from point blank, I expect some serious injury. But back in the days of playing Mechwarrior, my players would have shootouts at 10 feet with fully automatic weapons and shotguns. I fire, you fire, we subtract saome HP and continue. That is frustrating.

So, when I run a game, I don't care so much whether I've simulated all of the factors. I care that the outcomes seem plausible and that character behavior is influenced by the possibility of these outcomes.

Well, that tells me what you're looking for in a game.  But that's not the only possible goal.  In particular, most people's idea of plausibility for guns comes from action movies, which are different from reality.  From what I've read about statistics in police shootouts, most shots are going to miss or graze, even at pointblank range.  Now, since you're only concerned about plausibility, it's fine to expect pointblank shots to hit and do serious injury.  You find that plausible and it makes for a fun game.  

On the other hand, other people find it interesting to learn from their games.  i.e. Rather than simply matching their preconceptions, the game can push players to question these and go beyond them.  Now, I don't particularly recall Mechwarrior.  It's been a long time since I've dealt with guns in a campaign, but when I did I used Friday Night Firefight, which worked fairly well for these purposes.  More recently, I used a variant of RuneQuest for my Vinland game, which I thought also worked well.  It helped resist our instincts to recreate action movies, and instead gave a grittier and less glamorized feel for the violence.  

To get back to the topic -- Lev, you asked about attributes.  

But what are they supposed to represent? Genetic traits? Learned traits? Should this represenation be consistent? If not, why not?

What is the "right number" and distribution of such "stats"? GURPS has four; Rolemaster has ten - why such variation? Most games don't differentiate between manual dexterity and bodily agility (and thus end up with cheetahs who can pick pockets and mend watches?).

What's their relationship between stats and skills?

I've got an essay on my http://www.darkshire.net/~jhkim/rpg/systemdesign/">System Design page entitled "Nature vs Nurture in RPGs" which tackles the issue.  I'm mildly opposed to attributes in this sense (i.e. Dexterity, Agility, etc.), because I don't think this topic is particularly fruitful -- i.e. there is nothing to be learned.  I would suggest discarding abstract categories (i.e. "Dexterity"), and replacing them with functional abilities (i.e. "Covert Skills").

[lev_lafayette]

I've got an essay on my http://www.darkshire.net/~jhkim/rpg/systemdesign/">System Design page entitled "Nature vs Nurture in RPGs" which tackles the issue.  I'm mildly opposed to attributes in this sense (i.e. Dexterity, Agility, etc.), because I don't think this topic is particularly fruitful -- i.e. there is nothing to be learned.  I would suggest discarding abstract categories (i.e. "Dexterity"), and replacing them with functional abilities (i.e. "Covert Skills").

Can I just say I really like your comments on that page (short as it is), however I do think the conclusion is problematic, as it doesn't dispense with the problem that a well-trained individual can do things which are simply beyond human capacity. Of course, a way around that is to have species/age specific limits, but you know what players are like about limits...

Another issue I can arising is the scaling and differentiation between ability and effect. I gave a somewhat tongue-in-cheek example of a rat with a jumping skill of 100% (brilliant!) but an AGI of less than 1. Said rat would do really well at the act of jumping but wouldn't be amazing in terms of distance. How would a skill group resolve that issue?

[xenopulse]

John,

Interesting points. Let me reverse my argument: I don't think anyone enjoys the game if the result of an action is completely implausible. If you learn from the game, that means you figure out why the result is plausible. But there's just nothing plausible about someone standing right in front of you, hitting you (as neither of you are moving) directly with a shotgun or a burst from a fully automatic weapon, and you mostly shrugging it off. There might be freak accidents--but the mechanics make it a rule, not an incredibly unlikely event, that this happens. Therefore, the rules create implausible expectations, i.e., the expectation that I don't have to fear for my life and limbs when someone with a shotgun aims it at me from 5 feet away.

And you might be right, it could be a personal preference. But for me, it then matters little whether the actual shooting and injury is determined by calculating wind factors, organ damage, and all kinds of other stuff, or is much more abstract but still provides a plausible result (strong possibility of serious injury or death). The result can still be plausible even if the method is not highly detailed.

And my idea of "plausible" is somewhat malleable, of course, depending on whether the game is expressly cinematic or more "realistic" (have to be careful with this word now :). I have my own ideas on gun plausibility mostly from having been a marksman for many years, but that doesn't tell me much about actual combat. Except for the experience of one of my neighbors being shot on the street and me watching him bleed out.

Anyway. Maybe it all just shows that I'm slowly realizing that task resolution ain't doing it for me anymore, because the modeling of outcomes is either too time consuming or not plausible enough.

[lev_lafayette]

I don't think that I'd enjoy this game.  Why?  Because I like to play melee combat specialists, but I like to play combat specialists who are physically small.  It's some crazy aesthetic preference, it's just what I do.  You're saying that if my tiny character comes up against someone with equal training then my character will lose.

Not entirely so... A GM can always fudge results on the basis of good roleplaying. That, in a nutshell, is my way of using narrative to influence sim. Good actors just seem to be lucky. Bad actors face the full effects of reality. Can be a bit Hollywood, but hey, we like our screen heroes to do well, don't we?

In other words, for the purposes of sim, yes the rules reflect realism (I'm not even going to enter the debate vis-a-vis there being an objective reality or not). Yes, for the purposes of nar, the outcomes reflect drama.

But this is getting way OT from the initial propositions. What is the sim basis for attributes? What is their sim effects on outcomes? They're really the two questions I'm posing.


So, once again, let's try to define the goals of your specific game.  "Realism" is a huge goal, one that's so broad it's probably unattainable.  What, specifically do you want to seem "realistic"?  Psychology?  Physical actions?  Combat?  You could probably say "everything" and even design a game to handle that, but no one would play it because you need a computer to accurately model real-world physics, and I'm not willing to dedicate my play time to discover the precise striking force and armor penetration for a weapon.

Thomas[/quote]

[lev_lafayette]

Let me reverse my argument: I don't think anyone enjoys the game if the result of an action is completely implausible.

Well, they can, but there's a phrase for that. It's "munchkin gaming", right? I mean a system which effectively has a skill-based (i.e., you learn over time, and rather rapidly compared to your base) hit points does lead to results where...

But there's just nothing plausible about someone standing right in front of you, hitting you (as neither of you are moving) directly with a shotgun or a burst from a fully automatic weapon, and you mostly shrugging it off.

Exactly. You are completely correct. Sure, I don't mind a bit of Hackmaster every so often. Heck, such games are such bad sim that you may as well make them good gam or nar.

But ultimately it is preferable to have good resolution methods for all three, and I think it's nonsense, as is occassionally implied that an improvement is S must mean a decrease in N or G.

[quoteThe result can still be plausible even if the method is not highly detailed.[/quote]

More words of wisdom.

Anyway. Maybe it all just shows that I'm slowly realizing that task resolution ain't doing it for me anymore, because the modeling of outcomes is either too time consuming or not plausible enough.

This is not a bad thing however. It's how design evolves. People have a model, they test the model, they find the model is lacking in some aspect.

Hopefully, rather than giving up the process, they attempt to improve the model - and that's all I'm trying to do. I share your sentiments entirely - what's available out there just ain't good enough anymore.

Onwards!

[John Kim]

Interesting points. Let me reverse my argument: I don't think anyone enjoys the game if the result of an action is completely implausible. If you learn from the game, that means you figure out why the result is plausible. But there's just nothing plausible about someone standing right in front of you, hitting you (as neither of you are moving) directly with a shotgun or a burst from a fully automatic weapon, and you mostly shrugging it off.

Sure.  I agree that there are results which are both implausible and unrealistic -- and indeed many games produce such results.  I could quibble a little about realism, but I basically agree.  My issue with what you said earlier was that you removed realism from the picture, and substituted only plausibility.  

Most traditional as well as most narration-based mechanics would be plausible but not realistic.  There is no particular effort at simulation.  The game designer/GM/player may make up modifiers and results, but they are based on feeling rather than real-world knowledge.  So, for example, you can just tell the players to make up something that fits and it is fairly easy to be plausible.  

Realism is distinct in that it requires research.  i.e. Mechanics with are both plausible and realistic go beyond common sense and preconceptions.  For example, games like FVLMINATA or Sengoku incorporate historical research into character generation.  Games like Traveller have science in their starship and star system mechanics.  Games like Millenium's End have research in their gun damage.  These are uncommon but not non-existant.  

And you might be right, it could be a personal preference. But for me, it then matters little whether the actual shooting and injury is determined by calculating wind factors, organ damage, and all kinds of other stuff, or is much more abstract but still provides a plausible result (strong possibility of serious injury or death). The result can still be plausible even if the method is not highly detailed.

And my idea of "plausible" is somewhat malleable, of course, depending on whether the game is expressly cinematic or more "realistic" (have to be careful with this word now :).

I agree that both realism and plausibility are generally easier with less detail.  The less that you specify about a situation, the less chance that an imposed result will contradict what is there.  i.e. The less you say, the less chance you'll be wrong.  And I'd basically agree that "plausible" may be relative to expectations.  i.e. Results which are plausible in Toon may not be plausible in Dog in the Vineyard.  

Anyway. Maybe it all just shows that I'm slowly realizing that task resolution ain't doing it for me anymore, because the modeling of outcomes is either too time consuming or not plausible enough.

Well, to quibble a bit, something can be task resolution without having any modeling of process or outcomes.  For example, take Theatrix or Paranoia XP or Over the Edge or Sorcerer.  But anyhow, sure, that's fine as a preference.  But if you're simply not interested in that side of things, please don't hijack lev's thread to expound your own preferences.  

Can I just say I really like your comments on that page (short as it is), however I do think the conclusion is problematic, as it doesn't dispense with the problem that a well-trained individual can do things which are simply beyond human capacity. Of course, a way around that is to have species/age specific limits, but you know what players are like about limits...

Another issue I can arising is the scaling and differentiation between ability and effect. I gave a somewhat tongue-in-cheek example of a rat with a jumping skill of 100% (brilliant!) but an AGI of less than 1. Said rat would do really well at the act of jumping but wouldn't be amazing in terms of distance. How would a skill group resolve that issue?

Well, actually I wouldn't entirely.  I'm fine for using stats for directly measurable things like height, weight, and perhaps bench press force.  These sidestep the abstract nature-vs-nurture problem of abstract stats like "Dexterity" and "Intelligence".  You seem to want two different stats: one for jumping distance, and one for jumping accuracy.  You suggest that jumping distance should be based purely on attribute (implied inborn or genetic), while accuracy is purely skill (implied trained).  I don't agree with the latter.  Both distance and accuracy improve with training.  

In realistic terms, jumping distance is based on leg strength, leg length, flexibility/skill, and weight.  Leg strength determines force applied.  Leg length and flexibility/skill determine the distance the force is applied over (and how effectively).  Weight determines how far you can go.  

Personally, I would have "attributes" only for measurable things like Height, Weight, and Strength.  I would tend to have only one jumping skill which provides a modifier for both distance and determines accuracy.

[lev_lafayette]

I'm fine for using stats for directly measurable things like height, weight, and perhaps bench press force.  These sidestep the abstract nature-vs-nurture problem of abstract stats like "Dexterity" and "Intelligence".

Well, I agree with you there in most cases. Most RPGs tend to potray DEX as an attribute when what they mean is hand-eye coordination (largely learned), ditto INT. I mean there are attributes (mostly inborn) for DEX (manual dexterity) and INT (abstract reasoning), but that's not what RPGs potray. More their fault really.

You seem to want two different stats: one for jumping distance, and one for jumping accuracy.  You suggest that jumping distance should be based purely on attribute (implied inborn or genetic), while accuracy is purely skill (implied trained).  I don't agree with the latter.  Both distance and accuracy improve with training.  

Accuracy and distance can improve with training, but the former much more so than the latter. Indeed, I would say the former is pretty damn close to perfectable, whereas the latter hits some very obvious limits no matter how much a person (or a rat) trains.

I would tend to have only one jumping skill which provides a modifier for both distance and determines accuracy.

Sorry, but in a word; yuk! It allows for all the problems I've suggested in this thread. In a nutshell, that characters can "train" themselves to the point of implausibility. Again, I refer to our hypothetical rat... What does your single skill translate to in a real world example.

In the simulation system I have suggested which differentiates between effect and ability, the rat could jump less than 1m, but with exceptional accuracy. How would you simulate that wiith a single skill that determines both accuracy and distance?

[John Kim]

I would tend to have only one jumping skill which provides a modifier for both distance and determines accuracy.

Sorry, but in a word; yuk! It allows for all the problems I've suggested in this thread. In a nutshell, that characters can "train" themselves to the point of implausibility. Again, I refer to our hypothetical rat... What does your single skill translate to in a real world example.

In the simulation system I have suggested which differentiates between effect and ability, the rat could jump less than 1m, but with exceptional accuracy. How would you simulate that wiith a single skill that determines both accuracy and distance?

Ah, a bit of miscommunication.  I said that the single skill would be a modifier to distance.  i.e. There would be a base distance based on stats (at least height and weight), then a modifier for skill applied.   Obviously, it should be scaled such that maximum realistic skill results in maximum realistic human distance and accuracy.  You claim that maximum skills are unenforceable, but I think that obviously if you allow skills to go up infinitely then you are going to have unrealistic affects.  Accuracy can be just as unrealistic as distance (i.e. millimeter accuracy from a 8.5 meter long jump).  

Your solution means that training has zero affect on jumping distance, which I don't think is realistic.  It is not terrible, though, and if I was modifying a traditional attribute system I might go with it.  My typical nit is that systems tend to base on Strength attribute (i.e. HERO, GURPS, D20), which makes weightlifters the best jumpers.

[xenopulse]

But anyhow, sure, that's fine as a preference. But if you're simply not interested in that side of things, please don't hijack lev's thread to expound your own preferences.

If I didn't say anything thread-relevant, why did you respond to so many of my points? You're mistaking the end result of my ponderings (which were an aside) for not contributing to the discussion. You may need to work on your reductionism. And your tone.

Ralph started a new thread on plausibility and realism, so I'll make my points there instead of hijacking this thread further.

[Ron Edwards]

Oh chill out, everybody. Chilly formal asshole-ness is still asshole-ness. The Cop has shown up; behave, you know how.

For this thread to continue, the goal is to help Lev achieve his goals with the topic. If you don't think they're achievable, shuddup and find some other thread; that case has been made and it's up to him to process it, not up to you to repeat it.

Constructive interaction. Do it.

Best,
Ron

[LordSmerf]

But this is getting way OT from the initial propositions. What is the sim basis for attributes? What is their sim effects on outcomes? They're really the two questions I'm posing.

Let's try this one more time.  There is no right answer as far as I can see.  Simulationist play does not mean play that simulates reality (as we know it).  Instead Simulationist play is play that tries to simulate (or perhaps emulate) some specific thing.  If I want to simulate a game in which the characters are all cartoons a la Looney Toons, then it's perfectly logical to have skills be the only determiner of jumping distance.  Ditto if I want to simulate The Matrix, or Jedi who can use the Force to jump farther.

Now, let's just assume that the specific thing you want to simulate is reality (which I believe you've said).  What about this...

Attribute are the maximum attainable result, you never ever use them directly.  Then you have skills for how well you can utilize your attributes.  So you have (say) a long jump skill which can never be higher than whatever attribute determines jumping distance.  So the skill can continue to rise right up to the point where it matches the attribute, but you just can't get any better than that.  Would that do what you're talking about?

Thomas

[lev_lafayette]

Ah, a bit of miscommunication.  I said that the single skill would be a modifier to distance.  i.e. There would be a base distance based on stats (at least height and weight), then a modifier for skill applied.   Obviously, it should be scaled such that maximum realistic skill results in maximum realistic human distance and accuracy.  You claim that maximum skills are unenforceable, but I think that obviously if you allow skills to go up infinitely then you are going to have unrealistic affects.

Ahh, double miscommunication (and my fault for not elaborating!). I would also use the skill as a modifer, but not as a determinant of distance. Distance would be limited by stat.

"You've trained as well as you possibly can - but your body just can't do it".

It is not terrible, though, and if I was modifying a traditional attribute system I might go with it.  My typical nit is that systems tend to base on Strength attribute (i.e. HERO, GURPS, D20), which makes weightlifters the best jumpers.

Well it varies a lot in all those systems. In D20 for example, the attribute and skill have equal value.. So a semi-skilled individual can overcome a serious attribute problem - and without no real limit either (i.e., a +10 skill jumping advantage is the equivalent of DEX 20!).

What you've really identified here (and I agree entirely with you) is a lack of differentiation between STR (upper body, arms), AGI (lower body, legs) and general coordination (a learnable skill which is invariably wrongly attributed as DEX- that's use of one's hands!).

[lev_lafayette]

Attribute are the maximum attainable result, you never ever use them directly.  Then you have skills for how well you can utilize your attributes.  So you have (say) a long jump skill which can never be higher than whatever attribute determines jumping distance.  So the skill can continue to rise right up to the point where it matches the attribute, but you just can't get any better than that.  Would that do what you're talking about?

I love it when people get it right! ;-)

That's exactly what I have in mind. By the same token, I would like to see variation (random dice-rolls!) in the application of a skill. For example, a character with say a jump of 35% (pretty poor) and an AGI of 15, suddenly makes an amazing leap (i.e., rolls a critical success).

"That's my girl!", says her aging mentor, "I knew you had it in you".

Now, having (more or less) nailed down the principles for the differentiation between stats and skills... Onward! - To defining the stats themselves!

[LordSmerf]

Now, having (more or less) nailed down the principles for the differentiation between stats and skills... Onward! - To defining the stats themselves!

Again, what's the game about?  Do you want to differentiate between mathematical aptitude, artistic aptitude, linguistic aptitude, etc.?  Or do you just want to roll them all into "Intelligence"?  Now, you seem to be focused on physical things, so let's leave the mental/spiritual stuff aside and do body work.

Do you want to differentiate between:

Weight lifter and Long jumper (upper body/lower body)
Sprinter and Distance runner (fast twitch/endurance)
Beefy bruiser and Numble dodger (strength/nimbleness)
Sprinter and Long jumper (running fast/jumping far)
Push up master and Pull up master (triceps/biceps)
Ballet dancer and Wood carver (gross muscle control/fine muscle control)

Some others?

The point here is that there are limitless differentiations that can be made physically.  What your game is about determines which ones you want to make.  From your earlier posts you seem to indicate a desire to handle at least Upper/lower strength, Gross/fine muscle control, Strength/nimbleness.  Now you can roll Nimbleness into Gross muscle control so you have:

Upper body strength
Lower body strength
Gross muscle control (possibly with an upper/lower split)
Fine muscle control (also with a possible upper/lower split)

Then maybe toss in a catch-all endurance/fortitude of some sort...

Anyway, the answer to "What stats should I use" is just as dependent upon what the game is about as anything else...

Thomas