Forum changes: Editing of posts has been turned off until further notice.

Main Menu

When does combat resolution become too slow?

Started by Dauntless, September 29, 2004, 06:40:21 PM

Previous topic - Next topic

M. J. Young

I'm picking on this point because I think it betrays something deeper.
Quote from: DauntlessAs an example, most modern systems that allow you to build a character with a point system and choose skills however you like (a descriptive design approach) is not very realistic.  In this sense, class systems and random rolls to build things like attributes are in fact more realistic because they mirror more closely how reality works.  Descriptive design allows one to create the desired object as long as you "pay the points for it".  Procedural design can get you close to your desired output as long as A) you have the right inputs,  b) you are lucky enough or C) can afford the design requirements/prerequisites (mass, cost, social class, minimum trait score, etc).
There is a part of me that agrees with you wholeheartedly. If you want to create people who are representative of the probable distribution of the world population, then you need to use a system that creates statistically average characters.

But no one plays the ordinary character. (Not no one, but in the main that's not seen in the hobby.) We play the extraordinary characters, the heroes, superheroes, protagonists, great movers and shakers, even the gods.

What a system of randomized rolls does is produce characters whose variation from the statistical norm is predictable. It prevents you from having a party of all, say, fighters with 18(00) strengths (to use a D&D example), or having every character in the party display maximum values in the scores that matter to that class and minimum values in those which do not.

A well-designed point-based system, on the other hand, is designed to let the player build a character he wants to play, who is within the range of what the game allows. It is admitted that somewhere out there there is a fighter with an 18(00) strength. There are probably quite a few of them, since the odds of rolling that are one in twenty-one thousand six hundred (if I'm doing my spot math correctly), and most worlds have a lot more than that many people in them. If you want to play that one guy, the point based system lets you create him, instead of rolling again and again and again and destroying a lot of good characters who aren't what you wanted to play.

There is an inherent assumption in the randomized method that if it weren't randomized certain types of heroes would be underrepresented.

Don't misunderstand. There are some great strengths to systems that include randomizers. I gave some serious thought to various character generation models in my">Game Ideas Unlimited: CharGen and">Game Ideas Unlimited: Negative Points (these are currently for Gaming Outpost subscribers only, but the GO management is planning to eliminate the subscription system, refund money to people who paid for future subscriptions, and make everything available free, probably fairly soon). There's a lot to be said for randomization. However, I've known players who spent hours and hours rolling random characters and discarding them, in the effort to get the one they wanted. How random is that, really? On the other hand, is the alternative to force the player to play a character he really doesn't like at all?

Also, this struck me:
Quote from: Later heAs I gave in an example earlier, what seems a better tradeoff:

1) A resolution system that takes 30 seconds and is within 15% of the true score 90% of the time.

2) A resolution system that takes 2 minutes, and is within 5% of the true score 95% of the time.

It's these sort of considerations that I'm not really sure about how to go forward.
The first thing that struck me is that for the system to be seriously off five percent of the time is outrageous. A system that produces an accurate result at least ninety-nine out of one hundred rolls is the minimum tolerance, I would think.

And then I wondered, within 5% of what? There is an underlying assumption here that someone knows what "really would happen", and is designing a system to emulate that. However, I don't think we know what really would happen in most situations.

When he was in high school, my brother could do a standing broad jump of nine feet six inches, provided he could see just where he had to hit to do that. If he attempted to go an inch farther, he would probably fall backwards and lose distance. He could make that jump reliably enough that his coach told him to get the judges to show him where nine six was before he jumped, and aim for that. Now, what are the odds that he could make nine seven? No one knows. He could make it, sometimes. What are the odds that he would fail attempting to jump eight feet? Pretty slim, I'd guess--but it could happen, if he lost his concentration or slipped or something. All we can say is that if he knew where nine six was and he tried to hit that spot, he would succeed reliably. I never knew him to fail.

Before Columbia, what were the odds that a U.S. manned spaceship mission would result in a fatality? It had never happened. We'd lost astronauts on the ground, but never in space--not even on the ill-fated Apollo 13 mission where so much went wrong. Yet it did happen with Columbia, just as we were getting accustomed to the idea that it couldn't happen.

There's an article in one issue of The Way, the Truth, and the Dice entitled Hitting Them Where It Hurts, which reviews military statistics on injuries and fatalities in combat over the past century or so, and attempts to devise a realistic damage system from them--admirably so, and worth reviewing. Yet ultimately this sort of realism leads us in directions that are so different from our expectations that it doesn't really support play all that well. How do you know what would "really happen"? On most of the things that matter, you don't and you can't. On those trivial details where you can come up with an approximation, attempting to be accurate to the approximation is still an approximation.

Yes, you can have better and worse approximations.

You put a lot of effort into arguing for an approach that builds the results from the inputs. I understand that; to a degree, that was a strong influence in Multiverser--we want this to count, and this, and this, and this, and this, in whether the attacker hits the defender. In the end, though, the test of it all was when the ordinary person attacks the ordinary person, does his chance of injuring him match what we expect? You can talk all you want about how your system builds the chance of success from realistic variables, but in the end if you have absurd results--an ordinary person has no chance of injuring another ordinary person if he swings a pipe wrench at him, or a skilled marksman with a scoped rifle can kill a pigeon from thirty miles away on a lucky roll--all that talk of realism flies out the window. All evaluation of realism must be based on the output, not the inputs, no matter how important the inputs are, and if the output isn't any good, you've got to go back and change the inputs until it is good.

I feel like I'm wandering, so I'll stop here and hope that this helps.

--M. J. Young


Quote from: DauntlessActually, I think it's a very good idea for designers to state the reasons why they designed the game they did because it helps the players understand the games possible strengths and weaknesses, while also helping the player decide if he wants to buy into the system (although my rules are actually going to be I could care less about that).

I just wanted to say that i totally understand where you are coming from with this and to a great degree i agree with you.

Quote from: DauntlessAs for realistic rules requiring consistency, this consistency is based upon the judgment of the GM. ...[snip]...

This sentence really seems to highlight what you are working toward here.  This implies to me that the GM has final authority over what does and does not happen.  As in "Well, it just is not possible for you to hit a pigeon at 30 miles, sorry it doesn't happen."  So two questions arise:

1. Why have a system at all since the GM can just decide what happens?  Is it there as a set of guidelines so that the GM has a little less work?  (Note: that is a perfectly valid reason for a system.  If that is what you are doing it is not a problem.)

2. What do the players and player characters do?  Is this pure Simulationism?  Do the players exist simply to expore the world and their characters?

I am especially curious regarding these two issues because from what i have been reading it sounds like you are setting yourself up for some serious Illusionism.  (Note: Illusionism is not necessarily bad, remember that Participationism is a functional form of Illusionism play.)

I eagerly await your reply.

Current projects: Caper, Trust and Betrayal, The Suburban Crucible

Mike Holmes

MJ actually hit a lot of what I would have said. But I do have another angle.

You are not trying to create reality, we all agree on that, but some feeling that the results are accurate. As MJ said, if the system, despite looking at important inputs gives back bad data, you have to go back and consider the inputs. IOW, you're really starting with the desired output, and working back to the inputs anyhow.

But moreso than this, in a RPG, since we're not creating reality, we can decide what it is that we want to create. Sure it can mirror reality somehow if that's what you want, but what parts of reality it mirrors will depend greatly on what you want to display as the output. See, when modeling true reality, say for the purposes of an economic model, it's important that the output match reality in some useful way. In an RPG, you're creating fiction, and so you only have to match some form of fiction. Which can be anything.

Note that I think the games out there that are said to be most "realistic" are actually quite fictional in many ways. To a large extent this is because you are forced to limit the inputs to a set that produces output that says something about the inputs. For example, if we make calibre important in gunfire calculations, this means that the output says something about calibre.

But what most games notably lack is some way to describe the mental attitude of a person in a combat situation. That is, it's assumed that for some reason, the player gets to decide a combat participant's mental state to a large part. Basically, you are allowed to attack whenever you feel like it, or take aim, or do any such thing. Now, some few games do have "morale" rules, that attempt to cover these things. But that's just another angle, and one that's actually fallacious as it comes at the problem from a "management" perspective. That is, "poor morale" is a descriptive, not a proceedural way of describing any fighting unit, including the individual soldier, who won't perform. It doesn't say anything about why the soldier won't perform. Is he scared to be hurt? Scared to hurt others? Scared to leave his family without a father? Is this based on moral convictions, or is it because the person is immorally not with the goal of the mission?

Now, I'm not saying that ignoring all of this stuff is bad, it's just a choice. Recently, in the games I've played, these sorts of things have seemed to me to be really important considerations. So, to me when I play now, any system that just allows attacks without consideration of the character's mental make up at the time of the attack is critical? Not just because I find this stuff dramatic, either, it turns out that if you read the reports that MJ is talking about, or the works done post-conflict in any war this century, you find out that this mental state stuff is 90% of the effectiveness equation in combat. All the aiming, weapon accuracy, all of those inputs mean crap if the soldier refuses to fire his weapon as it's designed to be fired. Which is what happens 90% of the time.

Now, again, that doesn't mean that I don't find the sort of combat system that you're looking at invalid; what I find it to be is a choice of fictions. The fiction that I produce says that most of those factors of aiming and such are randomized, and come under the effect of the die roll. Your fiction says that the characters are mentally affected whatever way your system says they're affected, narrowing the scope of the output to a certain band. My output says something about the realities of individual mental make up, while yours says something about something else.

Both are "realistic." Even from a procedural POV. They just focus on different things, and, since they can't deal with everything, they are fictions in terms of what they aren't willing to address.

Which is fine, nobody is going to die because our model does not start with reality, but with a subset of the output that we're interested in seeing. Because, on top of all of this small talk about whether or not it's "realistic" to model combat my way or your way, is the thought that in fact, there's no real reason to give combat any more "realism" than any other area. This is my standard combat rant again, yes (you have read it, no?). One of the possible fictitous outputs that we can have doesn't give any more detail to combat than to anything else. Let's say that the game's basic resolution system is to flip a coin, and on heads, the character wins. In a combat situation, this is perfectly viable, and realistic. Because sometimes one wins in combat, and sometimes one looses. The system models this fact well.

If your game is about being a chef in nineteenth century France, then, in fact, there should be no "combat system" at all, and instead there should be an intensely complicated "Cooking system" that deals with the ins and outs of the subject matter.

Now, I'm presuming that you've decided that your game is about combat (if it's not, then we really have problems). And, again, I'm assuming that you've made the educated decision to do this "realistic" modeling. But that all must then mean that you've decided that "realistic combat" is the output that you're looking for.

Now, you can decieve yourself that you're working from the inputs. And, again, you can create a tailored output from the inputs, given that you're saying something about the inputs you choose. But if you actually choose the inputs based on some idea that "X input is neccessary, because it's crucial to the actual effect in real life," then you're doing yourself a disservice. Because any input can be ignored, and the result still remain "realistic." It's just a different realism. There is nothing that says that one input is better than another a priori for a RPG. Because the result has to real world implications. In the end, the result is to produce a fiction in the minds of the players. Yes, a potentially realistic one, but a fiction nonetheless.

So, given that the output doesn't "need" to be any particular thing, given that it's not a model that's going to be used to save lives in the real world, or put men on the moon, given that you can't be "accurate" for all the reasons that MJ states, given that all that's sought is a "realistic fiction," ...

Start with the output you want to see. That doesn't mean that you have to describe it descriptively, you can describe it as procedurally as you like. But even in the procedural method, you have to select which inputs make for the output desired, what the output talks about. And abstract the rest, as you must.

Second point, people don't know what reality is really like. They think, for example, that guns knock people down. I remember, for example, someone saying how they felt that the movie Shane was a breakthrough in that there's this special effect that they use when Jack Palance is hit by Shane's shot in the saloon, and he flies back into the other room like he was hit by a Mack truck. The person's comment was something to the effect that a Colt peacemaker was a powerful weapon. Which it was, actually (and highly innaccurate, too). Just that it couldn't possibly knock anyone over. Even if the collision were completely elastic (meaning that the bullet didn't penetrate, but instead just pushed him back), the force of a bullet is equal to the kick of the weapon. Newton says so. Which, if it doesn't knock over the firer isn't very well likely to knock over the target of a highly inelastic collision.

I remember later seeing the director had said that the reason he'd done the effect like they had was to reinforce the emotional impact of the bad guy getting shot.

So, I'm not sure how many people want "realistic" realism. I think they want that verisimilitude mentioned above more often, when they care about realism at all. Now, you may be writing to an audience which wants real realism, and that's fine. Just be aware of the difference. I think that that crowd is pretty small, actually.

Further, given that our understanding of the realities in question is limited, the best you can hope to do is to achieve some realism that relates to our current understanding of things (I think MJ might have said this). That is, who knows, maybe what I've read from the FBI about stopping power being a myth is incorrect, and they just haven't analyzed it enough yet. The point being that we're often better off abstracting details like this, given our lack of understanding overall.

The point here being that 50% accuracy is "good enough" because people won't know the difference. That is, if you make them go through more hoops to get a supposedly higher accuracy, they'll not see the point, because they'll neither be able to confirm or deny whether or not the output is actually accurate. In fact, as long as it's internally consistent, they'll assume that the rules accurately portray the game world in question. And as long as this doesn't produce any glaring errors, you're fine. So it's often best to aim for generally realistic. Which again means more abstraction.

Member of Indie Netgaming
-Get your indie game fix online.


MJ Young-
You're right in that sometimes you don't know the expected theoretical value.  In fact, it's very rare to know the expected theoretical value.  Sometimes it's even hard to come up with an observed or measured value.  Since you're a statistician (I I confusing you with someone else??), you're probably familiar with stochastic analysis and determining how we calculate probability itself.  Sometimes, trying to figure out an algorithm is impossible or too difficult to be useful.  So sometimes we just observe the events n-times, with n being a usefully large sample and see if a trend emerges (if the lim as n-> infinity converges to some limit).

If it does, then we can use this as a baseline....but unfortunately, it still doesn't help us with creating a function or algorithm to let us come up with this number.  But at least we now have a baseline number to examine.

If our observed results  don't seem to converge to a certain number with our given dependent variables (our inputs), then we seem to be stuck up a creek without a paddle.  It's in these cases that we resort to verisimiltude.  Or rather, I should say we create an appearance of reality that suits the intended atmosphere of the game system.  If you're trying to crate a grim and gritty game system, then you should have results which reflect that (i.e., grandstanding and trying to look good will get your ass shot off).  If you're trying to create a swashbuckling system, then leaping onto a chandelier to land on a foe should have a reasonable chance of success given the character abilities.

As for your point about playing characters which are larger than life.  I agree.  Games are (in some ways unfortunately) an escapist pleasure.  We want to be better than what we are.  So I've devised a "grading on the curve" scheme to character design, that will attempt to satisfy most players needs for some kind of balance or desire to have a character that fits their vision.   The more points you've spent to create or guide the character along on his path, the more "Karma" you've accumulated.

It becomes in essence a gauge for the GM to decide on whether or not experience should be awarded.  As the old adage goes; "For those whom much has been given, much is expected" applies here.  If you're powerful, then you have to stick your neck out more often and tackle harder problems to advance.  So it may turn out that someone who started with less karma may eventually surpass a more powerful character who wasn't challenging himself...not just against physical foes, but also in introspection.


Quote1. Why have a system at all since the GM can just decide what happens? Is it there as a set of guidelines so that the GM has a little less work? (Note: that is a perfectly valid reason for a system. If that is what you are doing it is not a problem.)

The system is there as a guideline for the GM.  The GM should have the right to break the rules, but doing so consistently will break the faith (as displayed by their credibility of the results) of the players.  In other words, it'a "good faith" system.  So yes you're right...the rules are there so the GM can let the rules do the work.  If he doesn't agree with the rules, then he can veto them so to speak.  Sometimes this is necessary because a game system can never be perfect, and the results you get out of the rules may not be very credible, or because the result leads to a direction that is undesirable (perhaps an incredibly lucky punch kills a character).

Why bother having rules if the GM has this power other than as guidelines?  Think of it this way.  The rules gives you a 2nd version of what happens under a given event.  The rules in effect says, "This is mostlikely would happen given these sets of conditions".  Then you compare this to the GM version...if he has one.  Two heads are always better than one...

Quote2. What do the players and player characters do? Is this pure Simulationism? Do the players exist simply to expore the world and their characters?

That's my primary idea.  But one of the reasons I'm aiming for a high degree of versimilitude is because I want the players to be able to relate to the characters, and internalize (to introspect on) some portion of what they explore.  IOW, it's a simulation where the exploration is not just of the game world, but also of the self.  This will hopefully be achieved by the more realistic rules granting a sense of plausibility and hence immersion, while also having a setting that allows the player to easily transfer the metaphor of the game world to not just the character experience, but his own.  That's why my game setting (to coincide with the grittiness of the rules), is based on an extrapolation of human society and culture about 100 years into the future.  It's a mix of military science fiction, transhumanism, and post-apocalypse settings....with a hidden but very deep reservoir of transcendental thought, like buddhism ,taoism, hinduism, esoteric Christianity (gnosticism) Judaism (Kabbalah and Islam (Sufism) that will be revealed in an overarching storyline.

A second game background also deals with a pan-asiatic pseudo-historical setting which has some of the same themes, but is set in alternate earth of approximately the 1880's time frame.


Mike H-
I agree with you entirely.  When we decide on the output of a system, we inherently give credence to whatever the  inputs are that decide the outcome.  Because reality is complex, we can't hope to account for every little variable.  So instead, we have to pick and choose which ones to account for.  By doing this, we emphasize the importance of those inputs.

Even in procedural design, you have a forward thought that says, "I want my final object to be something like this".  You plug in the inputs, and hopefully get close to what you were hoping for.  Sometimes even coming up with the algorithm in a procedural design has to be forward looking...or we have to make an assumption about a certain result.  Procedural design is also very hard to do simply because of the other point you mentioned....we sometimes have no real value upon to which to work.  Having no expected value means we can't build a function with the inputs we think are important.  So quite often, we must work in a descriptive manner.

So a large part of the art of design is picking the inputs that will "flavor" your game system.  The science of game design deals more with figuring out how those inputs affect one another under given conditions to produce a result.

It's funny you mention the aspect of morale on combat.  I hadn't thought of it before as a descriptive solution to a problem, but you're right in that it is.  In my game, I have Psyche Traits, which are analagous to the Passions in Riddle of Steel, or the chivalric codes in Pendragon, or if you've looked at Politically Incorrect Game's Active Exploits (which is freely available and highly recommended btw) their convictions and beliefs.  It's an attempt to justify a rather nebulous concept that I've called Emotional Stability.  Stability can take a +/- value with 0 being calm.  The psyche traits tell you how your stability became unbalanced, but the Emotional Stability track itself determines the affects.  In effect, my Emotional Stability is just another word for morale.

I guess my view of realistic game systems tries to match reality where it can, but I'm quite content with plugging in and filling the gaps with the "fictional" reality where required.  Another nice aspect of realistic design is that when you do have a theoretical or expected value you can use, then you don't have to worry about play balance too much.  If your game system mechanics come very close to the expected value very often, then you shouldn't need to worry about play balancing (unless the other parts of your system that may relate to it are really out of whack).  My fictional reality simply tries to make sure that the denial of plausibility never or rarely happens.


Quote from: DauntlessThe rules in effect says, "This is mostlikely would happen given these sets of conditions".  Then you compare this to the GM version...if he has one.  Two heads are always better than one...
I disagree with this in the strongest terms.  Two voices are not clearer and less confusing than one.
Just published: Capes
New Project:  Misery Bubblegum


In my experience, combat is one of the worst possible places for GM fiat to surface: it always seems like you're either trying to hose or to protect the character in question.


I didn't say they were clearer, just better :)

Take a look at democracy.  Having all these voices and differing perspectives is bewildering.  But at least a democracy allows for the expression of a different viewpoint to be taken into account.

That's what a system like this does.  It gives you one other output on which to reflect on.  Most of the time, the GM isn't even going to question the rules, and he'll go along with them.  There will be cases however in which going strictly by the rules either doesn't make sense (because no set of rules are perfect), or because something extremely lopsided happens due to pure chance.

If we only ever have one voice to listen to, then there are no needs for rules, or there are no needs for GM's (both viable approaches to roleplaying btw).


This isn't about a "perfect" rule system.  If your rule system is giving outputs that don't make sense for your goals then your rule system is broken.

You will seldom see a more lopsided case of bad dice-luck than this actual play post.  If we had been playing the way you intend to, the GM would promptly have overridden the dice to rescue us.  What a pointless waste that would have been.

You might take a look at this thread, which discusses more generally how fortune and rules as a whole have a role that should be respected.
Just published: Capes
New Project:  Misery Bubblegum

Mike Holmes

Tony, I think that the question of GM fiat is an interesting one, but perhaps something for a different thread. That is, the subject deserves to be discussed on it's own, without it interfering with, or being interfered with by, the other material in this post.

Dauntless, I'm the statistician/programmer/analyst (MJ is the Lawyer/Theologian).

We seem to agree: combat (or any other part of the game) becomes "too slow" when it produces output that doesn't match the goal. More precisely, when the output makes statements about the inputs, as we agree they must, and this output doesn't match the input, then the effort of putting in the inputs is not worthwhile. Pretty simple principle, when you get down to it.

Now, what you're saying is that you're goal is to be merely plausible. But I put it to you that it must be more than that. Because if it were to merely be plausible, again, you could just flip a coin to determine the winner (given two humans with everything else abstracted out, this is accurate). Your goal has to be that you want to make some statements about things like Psyche, weapon contribution to damage, character skill, whatever things you're worried about in terms of whether or not they'll make the resolution take too long.

Basically, your overall question can only mean one or both of two things:
1. You're worried about the elegance of your system. Does it provide the sought for feedback in such a way as the efforts involved do not outweigh the value of the output?
2. Are the things that you're using as inputs that are creating more work worth considering in terms of the output that's created?

The answer to the first is that only critical thought and playtest analysis can tell you if the system is elegant enough. It assumes that you are comfortable with the output, and only seek really to reduce the time needed to do the input.

The answer to the second is that you really have to consider whether or not a particular input is important or not. Because if your goal is "only plausibility" then that means that every possible input has the same quality for you. The Coriolis effect of the earth spinning is as potentially important as anything else. Yes, it's statistically unimportant to the outcome, but shouldn't it be in there as a -0 just to show that it's being taken into account?

Rather, what I'm saying is that you are already picking and choosing what's important at a level that's beyond "just plausibility." So discover what that goal is that you're trying to achieve, and drive for that. Because until you do, the problem is that a player playing will not see what the system is driving to provide, and so can't get "into" it. The product will seem random.

Further, this is the micro view of what's going on. It seems to me that you're saying that combat (perhaps all resolution, I don't know), isn't linked to the overall output of the game. That is, I'm guessing that your game isn't merely about "realistic combat." Or is it? What's the game in question? Who are the characters, and what do they do?

For example, if the game is about the Vietnam experience, then perhaps it's reasonable to want a realistic combat system. But the system should at that point have the overall goal of portraying the realities of Vietnam. Which is a different subset than the realities of WWII. Or any other conflict.

If the game is generic, then we have that as an entirely separate issue.

But you can't say that it's OK for any part of the game to simply be plausible, and support the overall output that the game is designed to produce. This would be trying to confirm the statement that System Doesn't Matter, that as long as the part in question is plausible that it'll fit in with whatever other thing is going on. Which is not to say that this is a broken design method. Simply that directed design of these elements is superior.

Discuss the relevant parts of reality to the output, not just the most statistically pertinent in terms of real world outcome.

Member of Indie Netgaming
-Get your indie game fix online.

Ben Lehman

Hi again.

Okay, now I'm a little more clear on your goals.

Have you ever heard of">Engle Matrix Games?  I'd be interested in your opinion.  They use dice to indicate "deviancy from the expected result" and can take a (theoretically) infinite number of inputs.  I'd be interested to hear your thoughts, at least, on the topic.



You can constrain your output results so that you have a limit or range of values that can be produced.  In effect, this will prevent any oddball results you might get.  However, such constraints are often artificial or they exist only because you can not account for every input variable that would go into the system.

I agree that 99% of the time, good design will prevent paradoxes or blatantly statistically deviant results.  But often being able to exclude these "bugs" in your design is done at the expense of factoring out or generalizing some of your inputs or the algorithm itself, thereby losing some detail.  So it's a balancing factor for the game designer, and sometimes it just doesn't quite fit right.


2. Are the things that you're using as inputs that are creating more work worth considering in terms of the output that's created?

I think this is the issue that I'm most concerned with.  I'd like to have the detail and tactical choice, and I'd like it to have an actual impact on the results.  But I'm wondering when and if it becomes too time intensive to be worth it.  As long as the playes have fun going over the crunchiness and also realizing that their tactical input has a decided effect on their chances of success, then I'm willing to sacrifice more of the playability factor.  

The realism/accuracy is still important of course.  But it's an auxilliary goal to immersing the players in the world that I've created.  The plausibility is there not for its own sake, but to increase the visceral influence it will have on the imagination (and here, I sharply disagree with those who feel abstracted generalized systems allow for more vivid imagery in the imagined space).

Mike Holmes

Sounds like we're talking diminishing returns here. That is, if we're not concerned with elegance per se, but with the number of inputs (or relative work to include them) versus the improvement they represent to the quality of the output, then it can only be a question of diminishing returns.

Because given that any particular input can be interesting in theory, if they're all delivering a constant amount of benefit to the output, then you could add as many as you like, and the game would have the same work:beneficial output ratio. But I'd agree that eventually there's a curve here. I think that at some point what really makes an input have less of a return is if it's "more of same."

For a stilted example, if you have one roll to determine whether or not a character's agility allows them to dodge an attack, and another to see if their reaction time saves them from being hit, and another to see if their intuition saves them, any of these alone are theoretically interesting. Having to do each of them would diminish each, however. Meaning that you'd have a lower effort:output ratio.

Yes, at some point, this becomes unplayable. The thing is that this seems to be input driven design to me again. Looking at the output, we wonder if each of these might be interesting as part of the output, and we decide to find a simple method that includes them all. Most systems will take the three stats in question and amalgamate them into some "defensive value" or something. This doesn't provide as much detail as the original system, but it provides enough to get a feel that the stats in question are being incorporated.

This is a key part of output driven design, and a subtle point that I'm making. It's not a question of elegance here - I could decide that the latter method is equivalent of the first, and make a change here based on the fact that it's more elegant. But let's assume for a second that the first method is seen to be more "accurate" from a procedural POV. In that case, from an input POV (which I argue is incorrect), you'd include the harder method. From an output POV, you note that the details are getting lost in the use of the system, and see that using the simpler system, though less proceedurally accurate, provides the more appropriate outcome. That is, despite the input method being in some way "innaccurate" to the real world, it's accurate to the output that's sought.

Again, it's choosing where to make your abstractions and fictions.

And, again, the system "takes too long" precisely when the added inputs are no longer adding proportionally to the output. Where they're getting lost in the jumble. Where the complexity is such that one can no longer weigh at all the impact of the individual elements.

Yes, this is damn realistic - in real life we have to play everything by ear. But these are RPGs, and not understanding what's happening, means that the system is returning feedback that's not satisfying in any way. We have to be able to process the feedback in order to be able to happily want to loop back to the system.

This is why people say that games like Go are great designs. Simple to learn, impossible to master. Because the feedback is simple and digestible. That doesn't make it any less intense in terms of strategy, or complexity of situation. It just mean that there are only a few well chosen inputs considered.

Again, when precisely you've put in too many inputs is a matter for playtesting and such - I don't think that it's possible to give a formulaic answer given that we can't know how interested people will be in the output value of any particular included input. But I will say this. Once you've gotten a largish set of inputs decided upon, I think that it's definitely a good idea to go through them and find ones to eliminate that serve essentially redundant purposes. And, yet again, what's redundant for one will not be for another, so you'll have to decide on your own.

On the subject of "genralized abstracted systems" first I think we're talking preferences here, so there's not much to say. But the support that a particular set of rules gives to "vivid imagery" depends very much on what it is that people are interested in imagining. If what they're interested in are the feelings of the people involved in combat, then what they need are complex rules that deal with that, and any details that speak to the physics involved will just get in the way. Because they're not looking for the system to make any statement there. But, generally, yes, systems can provide more or less support to any part of the imagination. I don't think that anyone here is for a lack of support. I think that they're for focusing support where it's desired.

Meaning that if you're focusing detail on physics that you're supporting visualizaton of things like bullets flying around, and swords clashing, in very precise ways. More precise than any other medium would care to offer, interestingly. That is, the "generalization" or "abstraction" of combat in some systems provides precisely the same sort of support for visualizing action that, say, action films do. Which is fun for some.

Member of Indie Netgaming
-Get your indie game fix online.