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

[lev_lafayette]

Ralph, I think you've missed the point. The matter I referred to was the lack of differentiation between stats and skills (the former being effect, the latter being ability to reach that effect). OK, so it's a sim argument and if you don't like Sim-based justifications, then there's no debate...

That said, there are a few things about a sim reason which I think you may have misrepresented a little...

It is not necessary to know exactly what centimeter a character can or cannot jump. But it is sloppy and unrealistic design when a skill rating defies principles of abstract simulation.

See the problem is, using D&D for the example, a DEX 8 character can make a 10 meter jump, without much difficulty or inhuman skill level.  Now I don't think that's on. The defining and limiting characteristic on how far one can jump is not skill, it's their stat (or as I prefer to call them, trait). Now how well they jump is a different matter...

Nor does it mean that this has to be particularly complex or "crunchy" as the vernacular seems to be here.  Simulation, to me, demands abstraction in the first instance, so compelling reasons for particular rules to me is a self-evident truism. Not everything is detailed "for Sim purposes". But dammit, if a rule is just nonsense - or a principle (such as effects being determined by skills rather than stats) - then it should be ditched.

None of this of course, is meant to impede on narrative or gamist approaches. Sim doesn't mean make it impossible to play. One can have good simulation and speedy resolution - and that is a design feature

[LordSmerf]

Lev,

What about long jumpers?  Or high jumpers?  Even if you think that
this is about "traits" it's pretty clear that you actually have two
different traits at work in these two different activities.  And
they'll be different from sprinting, and different from distance
running, and whatever.

So, how do you resolve this?  1) You either lose the granular
difference between the two and roll all of these "effects" into a
single trait that covers them, or 2) You have a large number of "core"
traits to cover the range of differences, or 3) You compromise by
having a small number of core traits and supplementing them with
something else, possibly "skills".

<snip>  But it is sloppy and unrealistic design when a skill rating defies principles of abstract simulation.  <snip>

This is an incredibly problematic statement.  What exactly are you trying to simulate?  If you are trying to simulate a world in which jumping distance is a skill (a la The Matrix), then I'm not seeing a problem.

Thomas

[Valamir]

Ralph, I think you've missed the point. The matter I referred to was the lack of differentiation between stats and skills (the former being effect, the latter being ability to reach that effect). OK, so it's a sim argument and if you don't like Sim-based justifications, then there's no debate...

Actually Lev, I don't believe I've missed the point at all.  I'll try to avoid repeating myself, so please refer to my earlier comments as well as those below.

I have no dislike for Sim justifications.  A few years ago, I would have (and did) make the exact same arguements you are.

I have since come to realize that those arguements...alone and by themselves...are inadequate to justify spending any time at all worrying about things like differentiation between stats and skills (and what effect that might have on things like jumping distance).  Note: I specifically used the word "Inadequate", not "Irrelevant".

see...all of this...

But it is sloppy and unrealistic design when a skill rating defies principles of abstract simulation.

See the problem is, using D&D for the example, a DEX 8 character can make a 10 meter jump, without much difficulty or inhuman skill level.  Now I don't think that's on. The defining and limiting characteristic on how far one can jump is not skill, it's their stat (or as I prefer to call them, trait). Now how well they jump is a different matter...

...is insufficient.  You are not articulating a foundational design principle here.  At best you are articulating a personal preference  - which may actually BE a true preference, or may simply be an artifact of what you're already familiar with and thus most comfortable with.  In other words, right now you're just assuming that these things are a design feature.

I'm saying they're not.  EVEN for a SIM game these are NOT design features...YET

They CANNOT be design features until you first define what the overall purpose of your game is, what your game play goals are, and what specific elements of game play you want the rules to support...and most importantly...why.  Again, you can't start working on a form for your game until you first define its function.  Until you've defined a function...NO FORM can be considered a feature.

"because it makes for a good Simulation" is not an adequate definition of the game's function.  First you have to tackle what it is about your game that you think is best handled by rules that are an accurate simulation.  What are those rules to be used FOR that would make such a simulative approach the ideal.

There is an erroneous assumption that has been made by game designers for years...and you're making it now.  That simply because you CAN make rules that are an accurate simulation* you SHOULD.  The error goes on to assume that once you have your rules set that provides a nice accurate simulation you can then go ahead and use it for whatever game you need to because if the rules accurately model physics they'll work well for anything.

That's simply not true.  Even for sim design.  In fact, I'd say that not only is it not true all of the time, that its rarely ever true.  Rules that provide an accurate simulation of physics (and when you're talking jumping 10 meters or not...you're talking physics) are ONLY ideal for that small subset of games where the most significant action players will be having their characters do is interacting with the physical world in a realistic way.  For any other purpose such rules range from unnecessary at best to down right obstructive.  Again this is not a Sim vs. Nar vs. Gamist consideration.  This is understanding what you are trying to simulate and why...BEFORE you set about writing rules.

For example, if your game is to simulate the struggle of the social underclasses to find a voice against an oppressive government in a world where no one can be certain who they can trust and everything they hold dear can be yanked away tomorrow...an accurate simulation of physics is almost assuredly irrelevant.  An accurate simulation of individual and crowd psychology would be far more useful.  How your neighbor might respond to fear, intimidation and social pressure is far more useful to define mechancally than how far they can jump.


So, to summarize:

Step 1) Define what your game is about.  What characters are supposed to be doing during play, and what play experience players are supposed to have while they play.

Step 2) Determine what elements of game play require mechanical support.  Constantly refer back to step 1 and avoid writing rules about stuff that isn't part of Step 1.  Decide how those elements should be portrayed using rules in order to deliver the desired play experience.

Step 3) NOW start to design the actual mechanics refering back constantly to step 1 and 2.  IF and ONLY IF step 1 and 2 included strong justification for rules that provide an accurate simulation as part of the necessary design aesthetic should you be worried about such things as the relationship of Skill to Stat.   If there is such a reason, great.  If not, don't waste your time on trivium.



* of course "accurate simulation" is itself a misnomer given that few people can even agree on which elements can be simulated and which need to be abstracted let alone how accurate is accurate enough (accurate enough for you is bound to be nonsense for someone else).

[lev_lafayette]

...is insufficient.  You are not articulating a foundational design principle here.  At best you are articulating a personal preference  - which may actually BE a true preference, or may simply be an artifact of what you're already familiar with and thus most comfortable with.  In other words, right now you're just assuming that these things are a design feature.

Well I'll respectfully disagree. I think they are a design feature. Because I believe that a game system must be a simulation (game narrative is a different matter).

They CANNOT be design features until you first define what the overall purpose of your game is, what your game play goals are, and what specific elements of game play you want the rules to support...and most importantly...why.

OK, the purpose is to accurately simulate historical fantasy from the perspective of individual participations. There. Done. Now can we go on to sim-design?

"because it makes for a good Simulation" is not an adequate definition of the game's function.

In your opinion. In my opinion, good simulation is adequte justification. There are already sufficient imaginative setting out there. Assume, if only for the purposes of debate, that the games purpose has already been determined.

That simply because you CAN make rules that are an accurate simulation* you SHOULD.

Which, if it doesn't distract from play, actually supports play. The inverse of the argument (that because we can make realistic and playable rules we shouldn't) is nonsense.

The error goes on to assume that once you have your rules set that provides a nice accurate simulation you can then go ahead and use it for whatever game you need to because if the rules accurately model physics they'll work well for anything.

Physics, and history, and sociology and anthropology etc. So why not?

Step 3) NOW start to design the actual mechanics refering back constantly to step 1 and 2.  IF and ONLY IF step 1 and 2 included strong justification for rules that provide an accurate simulation as part of the necessary design aesthetic should you be worried about such things as the relationship of Skill to Stat.   If there is such a reason, great.  If not, don't waste your time on trivium.

Why don't just assume that the discussion is already at step three, rather than engage in an off-topic tangent?

[lev_lafayette]

What about long jumpers?  Or high jumpers?  Even if you think that
this is about "traits" it's pretty clear that you actually have two
different traits at work in these two different activities.  And
they'll be different from sprinting, and different from distance
running, and whatever.

Excellent questions.

Here's an idea. Yes, you do have to compromise to an extent. I wouldn't want one trait per DNA sequence. Say you're using AGI to represent maximun potential application of leg strength. You could still have separate skills for long and high jump, but the effect would be determined by the AGI. This is significantly different from having the skill determine the effect as is the case in some or having the skill determined by the trait.

It scales reasonably well too; you could have a rat with a jumping skill of 100%, but an AGI of 0.1 - a brilliant jumper, but still not one which makes huge distances.

This is an incredibly problematic statement.  What exactly are you trying to simulate?  If you are trying to simulate a world in which jumping distance is a skill (a la The Matrix), then I'm not seeing a problem.

Good point. But even in The Matrix it was only a skill for those people who had hacked the code, who had  seen through the reality. For everyone else it was still a trait. Imagine a character having a meta-skill if you like - the ability to determine their traits. "I want to have DEX 30", said Kenau... and the next line was "Whoa!"..

[Valamir]

Why don't just assume that the discussion is already at step three, rather than engage in an off-topic tangent?

Because that assumption is what has led us to 2 decades of for-shit game design.  Because there are already about a zillion games which do "accurate simulation" already and the game world needs another one like it needs a hole in the head.

The fact that you want to bypass all of the really important fundamental design questions in order to skip to the trivium...because its the trivium that's captured your interest at this point in time...is just a repetition of the same flawed game design method.

You keep putting your fingers in your ears and pretending that I'm saying "sim design is bad" so that you can come back with nothing more substantial than "sim design is good".  When in reality, alls I'm saying is that you are putting the cart way before the horse.  Its akin to building a ship and debating whether it should have 2 masts or 3 before you even decide what you're building the ship for.  Either answer could be right, but jumping the gun to talk about it is a waste of time.  There is no possible right answer without first knowing what the ship is to be used for.

But I've said all I can say on that matter.  Since apparently you aren't interested in discussing the fundamental logic behind your question, I'll bow out of the thread.


Edited to add:
I did think of one other angle I thought I thow out there to illustrate my point.

Lets say that you wind up having an extended dialog about this issue and a bunch of folks like Thomas jump in and offer their ideas.  Lets say at the end of it all you have 3 different mechanical approaches that all tackle the desire for accurate simulation somewhat differently (or 30, or 300).

How do you decide which is the best one?  What standard do you you use to judge whether idea #1 is better than idea #3?  How do you know that any of the ideas actually fulfill the goal?  THAT's why I've stressed starting from the idea of knowing what your game is about before you start to get into the details.  Because without knowing that, you have no yard stick to measure the appropriateness of the solutions you come up with.

[LordSmerf]

That simply because you CAN make rules that are an accurate simulation* you SHOULD.

Which, if it doesn't distract from play, actually supports play. The inverse of the argument (that because we can make realistic and playable rules we shouldn't) is nonsense.

This is a major breakdown in the understanding of the way roleplaying games work on your part.  The statement itself is true, but the problem is: every point of contact (i.e. every time you access a rule) you are required to spend some amount of time.  The more complex the rule, the longer the time it takes.  This means that any rule that does not contribute directly to attaining the goal of the game (whatever it may be) necessarily is a distraction from the point of play.

Here's an idea. Yes, you do have to compromise to an extent. I wouldn't want one trait per DNA sequence. Say you're using AGI to represent maximun potential application of leg strength. You could still have separate skills for long and high jump, but the effect would be determined by the AGI. This is significantly different from having the skill determine the effect as is the case in some or having the skill determined by the trait.

This is also problematic.  You know that you're going to have to compromise, but there's not a good way to know how much is too much.  The reason is it is different for different people.  Not everyone has the same tolerance for this compromise thing, it's a pretty big deal.  You're basically saying "My way is the right way, because it is."  You're not really saying why your way is right, and I don't know that there's any way that you can...

Thomas

[lev_lafayette]

Because there are already about a zillion games which do "accurate simulation" already and the game world needs another one like it needs a hole in the head.

That's quite an assumption on your part. I would have thought that my original post made it quite clear that certainly is not the case.

The fact that you want to bypass all of the really important fundamental design questions in order to skip to the trivium.

Your "really important fundamental design questions". It may be a view shared by others, but it is not a universal concern - and I may add, nor I am I suggesting that my concerns are universal. I am simply raising them for those who may be interested.

Sorry for not joining your party.

Its akin to building a ship and debating whether it should have 2 masts or 3 before you even decide what you're building the ship for.

I know what a roleplaying game is for, and I know what particular roleplaying games are for. In this case the analogy is more of a question of "will this design float?"

But I've said all I can say on that matter.  Since apparently you aren't interested in discussing the fundamental logic behind your question, I'll bow out of the thread.

Thank you. When I raise a what a game is for, or more to the point, how to modify a sim according to a narrative structure, please feel free to provide your wisdom.

[lev_lafayette]

This means that any rule that does not contribute directly to attaining the goal of the game (whatever it may be) necessarily is a distraction from the point of play.

That's true enough and you'll find no dispute from me on that level. Nevertheless, assuming that a rule does contribute to play (which to me is a trivial assumption), can we state that an accurate rule is better than an inaccurate one, assuming playability is the same? That a rule which does not require additional suspension of disbelief (such as our ten meter leaping old man with a limp) than what already exists?

This is also problematic.  You know that you're going to have to compromise, but there's not a good way to know how much is too much.  The reason is it is different for different people.

No argument from me again. Of course, the degree of complexity is going to vary from gaming group to gaming group. That's what house rules are for. However the principle is a clear distinction between traits (potential effect) and skills (application of trait for a specific action) is one which I think is important and, dare I suggest, somewhat novel.

[Bankuei]

Hi lev,

Let me make sure I'm understanding you:

Baseline assumptions for discussion:

1) We're talking about games that rate, measure and model abilities of the characters- not things like "dramatic influence", "player plot points" etc. as the focus of resolution

2) That these abilities are also being rated by innate ability & trained ability seperately

3) That all the above also is something that promotes the goals of this particular game design.

The miscommunication that is happening between you and Ralph is that these baseline assumptions were not stated at the onset.  There are many game designs that benefit by the above axioms, and there are many more which do not, simply because they were adopted without consideration for the actual design goals of the game.  An example of a functional game that completely lacks any of the above would be Nighttime Animals Save the World.  All Ralph is saying is that not all games require, or benefit by those assumptions.

So, assuming we are going with the above assumptions, your questions were rather hitting quite a few different topics:

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? In RQ a character with a Dex of 30 gains +20% to their attack rolls. Using a similar scale, a GURPS character would effectively gain a bonus of more than 100%. Why is there such variation between two systems both which claimed to be realistic?

What role should ratings, stats, traits, attributes or whatever we're going to call them hold?  

They should rate whatever is relevant and useful for the game design.  

If it is functionally important to measure how big a character's feet are for the purposes of play- then that should be included.  If it's not important, then it shouldn't be included.  That simple.  That also answers how wide of a distribution you will need.  

Should these things represent innate ability, learning, drive, whatever?  Depends on what you feel is plausible for the setting and if that explaination is even necessary.  Few games take the time to really explain all the hoo-hah of attributes- they usually just sum it up to being "ratings, now, to be used in play" without all the navel gazing of whether it was a genetic thing, or a matter of lifestyle, fate, destiny, magic, etc.

There is no "universal answer" for what is best for game design.  Each game will be unique according to its design goals.  The reason RQ and GURPS are different despite both attempting to model realism is:

1) they have very different resolution systems
2) the designers had different research material and beliefs of what "realistic" is
3) the designers took very different approaches towards solving the balance between modeling physics and playability

No one can give you meaningful answers to your questions because the questions need more qualification.  If you went to the drug store and asked what kind of medicine you needed- they would need to know what illness or condition you are suffering from- it would be impossible to answer your question without further information.

Likewise- here we have a lot of questions without the necessary design goals.  There is no way I can tell you how well a vehicle travels if I don't know if its supposed to go on land, sea or in air.

You might want to consider refocusing your question a bit better and everyone will be able to help you more.

Chris

[Callan S.]

Its akin to building a ship and debating whether it should have 2 masts or 3 before you even decide what you're building the ship for.

I know what a roleplaying game is for, and I know what particular roleplaying games are for. In this case the analogy is more of a question of "will this design float?"

Well, what is a roleplaying game for (in your opinion)? Yours in particular?
Then ask yourself, does that answer some questions about what traits/skills you should have?

I think that's all Ralph is saying. And if this doesn't help answer these questions very well, perhaps you could better define what the game will be for, until it does help. Further defining your game may help answer these questions.

[lev_lafayette]

Baseline assumptions for discussion:

1) We're talking about games that rate, measure and model abilities of the characters- not things like "dramatic influence", "player plot points" etc. as the focus of resolution

2) That these abilities are also being rated by innate ability & trained ability seperately

3) That all the above also is something that promotes the goals of this particular game design.

Damn that's a good summary.

There is no "universal answer" for what is best for game design.  Each game will be unique according to its design goals.  The reason RQ and GURPS are different despite both attempting to model realism is:

1) they have very different resolution systems
2) the designers had different research material and beliefs of what "realistic" is
3) the designers took very different approaches towards solving the balance between modeling physics and playability

This is where I must express disagreement. Both use roughly the same scale and resolution method. Both claimed to realistic by "real world" standards. Both have the same approach and balance in modelling physics and playability.

The question is why? Which one is "more true"?

[Bankuei]

Hi lev,

A normal car, a baby carriage, and a mars explorerer robot all have 4 wheels- why do they all work differently?

A similar scale doesn't mean much.  The resolution is VERY different, just from % rolls vs. 3D6 under a shifting number.  We're talking the difference between a linear probability and a bell curve.  

"Realism" as anyone, anywhere understands it is based on what materials they're looking at and consider more credible + their own life experiences.  Check out all the debates over how ballistics work and you have a perfect example of real people, who fire a lot of guns, who research into actual physics- and still come up with different answers.  Even experiments are often biased and under different conditions- despite the ideal of the repeatable experiment.

Which one is more true?  Well, I suppose you could go research into general physics, grab tons of data and match it average results given by any system.  You'll have to check the data and figure out what to keep, what not to keep, which are likely biased("Car accident info from a public study funded by General Motors... hmm...").

In the end, you'll have to do the work of a real researcher.  No one is a master of all fields of physics and human interactions- and no one can give 100% accurate info on everything- AND compare it to the odds that show up in a game system.  The sheer logistical impossibility of being able to test and verify everything means "reality" has limited applications in terms of game design.

Again- even if our goal is realism, we have to ask, "How much time/effort  towards resolution will hit a playable balance between realism and play?"  "What aspects of play are most important to fit what evidence to create 'realistic'?"

Or- for a perfect example- consider ranged combat in GURPS.  It's harder to hit a target moving lateral(or, presumably vertical as well)  to you than one standing still, or moving in a straight line towards or away from you.  This makes perfect sense.  But the modifiers and the odds given to you- do you think the folks at SJ games set up experiments over 100's to 1000's of subjects to discover how much the % to hit goes down based on a target's movement?  And did they average the degree of angle at which a target would be moving to the shooter?  And did they map that accordingly to the bell curve distribution in GURPS?  The idea makes sense, but the numbers were pulled out of thin air.  Even the US military doesn't know the exact % by which lateral movement reduces accuracy.  It's too complex and too minute of detail to bother researching- instead it just makes more sense to consider the general causes and look into minimizing it's effects in real life.

Reality may exist for all of us, but you can see as soon as we're talking about modeling it with the human brain(and not, say, a bank of Cray supercomputers), that we have to set aside reality as it is, and look at what it means to us.  We have to look at what our game design goals are, what gives the "feeling" of realism and then start talking about how the resolution mechanics interface with it.

Chris

[xenopulse]

I agree with Chris. And Ralph, for that matter.

Also, reality has data available. In the game, you make up most of the data anyway. What's the wind speed? Are you going to simulate a whole continent full of weather, or determine what it is at that point in time anyway? Does the character have something in his eye when he's about to shoot? Is he getting a cramp from running while he's preparing to fire? Is he actually breathing right? Has he learned that before and does he remember to apply it? What did he have for breakfast, and is his stomach upset? Is his mind still distracted from breaking up with his girl friend earlier that day? All those things matter for a single fired shot. How many are you going to simulate, and how many are you determining?

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.

As a side note, most modern scientists will tell you that "truth" can only be defined in a pragmatic way (a la William James) nowadays. Gone are the days when we thought that truth was a direct representation of reality. All of our scientific theories are only "true" insofar as they are useful for our interaction with reality through the prediction of future events within the interpretation of the theoretic framework itself.

So--this leads back to what has been said. The system is more "true" that has more cash-value (in James' terms), i.e. is more useful. What's more useful? Whatever satisfies the purpose of the game.

[LordSmerf]

This means that any rule that does not contribute directly to attaining the goal of the game (whatever it may be) necessarily is a distraction from the point of play.

That's true enough and you'll find no dispute from me on that level. Nevertheless, assuming that a rule does contribute to play (which to me is a trivial assumption), can we state that an accurate rule is better than an inaccurate one, assuming playability is the same? That a rule which does not require additional suspension of disbelief (such as our ten meter leaping old man with a limp) than what already exists?

Okay, let's assume that a rule does contribute (note, you say "to play" which is wrong, all rules can be said to "contribute" to play in the sense that they have an effect on the game and the SIS.  What is important here is that they contribute to whatever the goal of the game is, we'll just assume that for your game "realism" is the hightest goal) to the game's goal.  Universally, no we can't say that "accuracy" is better than "inaccuracy".  See xenopulse's post, "accuracy" just means that an outcome you expected has occurred.  You're not actually modelling reality, you're modelling reality as you see it.

If I wanted a game that was totally surreal then it turns out that "inaccuracy" probably meets my goals more than "accuracy".  Now, we'll move on from that because I think you understand that point, and that you're specifically talking about your theoretical game where "realism" is the goal.  Just remember that "realism" isn't something that you can say is really an objective thing...

This is also problematic.  You know that you're going to have to compromise, but there's not a good way to know how much is too much.  The reason is it is different for different people.

No argument from me again. Of course, the degree of complexity is going to vary from gaming group to gaming group. That's what house rules are for. However the principle is a clear distinction between traits (potential effect) and skills (application of trait for a specific action) is one which I think is important and, dare I suggest, somewhat novel.

Novel, yes definitely.  I don't think I've seen a game that models this.  Important?  I don't know, I guess it depends on your design goals.  Let's go ahead and assume that for your specific take on "realism" this is a very important consideration.

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.

Now, that's fine, and going by xenopulse's definition of "realism" (it conforms with what I expect to happen) then it even seems "realistic".  But I don't want realism here, in fact this is a case where choosing realism hurts the game for me because to play my preferred aesthetic I have to take a character who is sub-optimal mechanically.  You are penalizing my desire to play what I want to play.

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