On loosing faith in simulation

[b_bankhead]

When I started out in rpgs so many years ago I was a died in the wool simulationist.  When my friends decided to write our own rpg we actually did things like jumping over tables and measuring the time with a stopwatch to get our time/action/initiative system just right.
But over the years I have gradually lost interest is simulationism for a number of reasons:
     1.The large number of occasions where various rules, including complex ones fail basic pen and paper reality tests.( For some examples read the book Murphy's Rules is collates a bunch of silly rules gaffes, many of them from complex simualtionist style systems)
           2. A realization that the narrative was what I actually enjoyed.
   3. Skepticism about the Holy Grail of simulation itself.

   To simulate something you have to have two things: a mathematical model, and a set of data about the start-points for the system. Plus the amount of math involved increases exponentially with the number of elements in the system. So the amount of math you have to do to run the simulation rapidly becomes unacceptable for by-hand calculations.(That is to say keeping track of it imposed an excessive delay in the narrative...)  For example, the simple matter of bleeding. How many GMs keep track of blood loss from wounds?  Practically none,because the additional bookeeping is too much of a burden. Yet bleeding is only one of the issues involved in modeling wounding and its effects.

   Or take bullet damage.  Bullet effects are exceedingly difficult to predict,one guy take 4 handgun slugs in the torso and WALKS three miles to the emergency room, another manages to kill himself with a Crossman B-B gun.  Any rpg damage system to really model this would have to be much more complex than can be calculated by hand.

   Back in the 80's the US military was investigating the issue of changing its standard sidearm from the .45 to a 9mm.   An extremely complicated  debate ensued about whether a smaller faster bullet was 'better' than a big slow one. Trauma surgeons,ballistics experts, soldiers,policemen all weighed in with signifigantly varied oppinions.  Complex tests involving shooting bulllets in gelatin and asessing the damage, and in one group of tests they actually shot some cows and then diseccted them.  Given how difficult 'simulating' bullet wounds was for these real experts how much of a chance is some RPG designer who may have never fired a weapon supposed to design 'ex nihilio' a bullet damage simulation?  

   One of my favorite RPGs (KULT) has different wound charts for no less than 15  types of pistols.  You may be sure that no source exists for good data about the different wound effects for such subtly different weapons.  Deep down virtually everyone knows that designers are making these things up by gueswork and intuition.  And that goes for long charts of subtly varied modifiers for diffent conditions, guesswork ,guesswork all.  

     And when we add the issue of 'simulation' of various genre effects the idea of simulationism really hits the iceberg.  Because then 'simulation' really becomes the working out of narrative outcomes rather than calculation of real-world physical effects and thus collapses into narrativism in practice.

     Given all of this, I have concluded that anything but the most 'fuzzy' of simulation in a by hand rpg is effectively impossible. Anybody who claims to do better than 'great honking gun' and 'little dinky gun' is lying to you and waisting your time making you look up numbers that add up to nothing and making you pay to have them printed out (and that goes for sword and spears too).

   So when somebody has taken a slug and wants to fire back and want to know how much it effects his aim what do we tell him?  Has the friendly game designer gone to some secret vault in the Pentagon where the keep the scholarly records on how much a slug to the gut effects your range performance? Or is he going to say " Hmm four hit points must hurt a bunch I bet it would effect your aim a whole lot, lets call it -5 .....?

   So think of this next time you examine some complex simulationist style system filled with charts and rules.  Big columns of numbers arrayed all nice and neat can produce an illusion  of precision (and realism, whatever that is to you...) entirely unearned by any real world considerations....

[Jeremy Cole]

Add to all that the fact that the incapacitating effect of a bullet is as much its psychological effect on the victim as the actual physical damage...

Simulationism, though, isn't all about realism, or fine levels of granularity.  Its about ensuring cause and effect, at whatever level of abstraction you want, this is an important distinction.

For my mind, the most interesting ruleset to me is the hybrid of light simulationism with narratavist elements*.  I'll probably piss someone off for saying this but I think a lot of the narratavist games out there are simulationist in many ways, forex they still base a lot of resolution around character ability and task difficulty, a simulationist desire to have real world cause and effect.

For my own game ce, I found it was important to limit the player's actions by the inclusion of a simulationist resolution mechanic (whether it would address premise or not, that guy will kick your ass, you can't join that exclusive faction etc).  This does address premise, because the premise I have chosen is in part about dealing with real world limitations, but it is still 'rules light' simulationism.

*Of course, this hybrid is very easy to turn to gamism with narratavist elements, something I am interested in seeing work.

Jeremy

[JMendes]

Hey, :)

A die-hard simulationist is what I am now. :) I do agree across the board with your post. I also think that there is another side to consider.

For instance, I have spent many an hour in utter amazement as I stared at endless charts of Rolemaster hit results, pondering your very point. Where the hell are they getting their data? And then it hit me: playtest!

Ok, so that may not be an accurate model of the laws of physics of the real world. (Or it might be, by some incredible stroke of luck. Shakespeare and the monkeys and typewriters and all that...) But so what? It is an accurate model of the laws of physics of a world. A world that can be explored from a simulationist standpoint.

By the way, you seem to be more heavily against system-based simulationism. But what about setting-based? Or character-based?
(I confess, I have no great interest in situation-based, and I can't quite conceive of color-based simulationism...)

Anyway, I have reached the point you are at, moved beyond simulationism (into gamism, imagine that, narrativism flies over my head) then decided I wasn't looking at it right and came back to it.

Cheers,

J.

[Ron Edwards]

Hi there,

b_bankhead, I totally agree with you on the personal level. You have never seen a more dedicated Illusionist GM than me, from 1982 through 1992 or so. But ultimately it came down to simple personal priorities.

Analytically speaking, though, the topic of successful Simulationist play fascinates me.

One of the several drafts for essays I'm working on at the moment (in between endless scribbling for the Sorcerer supplement), is a rigorous parsing of Simulationist game design. It's really hard, mainly because few of the Simulationist players/groups that I've known have seemed happy. Some have - and upon lots of discussion and observation, they seem to be the ones who can pick What to Explore, get really focused on cause-and-effect for that, and "let be" for the other stuff.

I'll give some very brief and sketchy summary notions.

Situation-Exploration (e.g. Call of Cthulhu) seems to be easiest to reach a satisfactory plateau. Character-Exploration and Setting-Exploration do all right if the game and game-prep and group can keep the two elements in a consistent relationship (e.g. where conflict/tasks arise from).

System-Exploration usually adds Setting-Exploration during early prep (e.g. GM-prep, usually) and then it does well for a while ... but you know, that's where people have trouble, as it continues, and that System-Exploration continually has to live up to its own standards. If you can be confident that the axe "can" chop the guy's head off, that's great. But then the lackadaisical distinction between a hatchet and a tomahawk starts to glare out at you during play ... and you have to factor in the relative skill levels of the characters, even as they constantly improve ... and ...

Anyway, I think that a few of the System-Explore games out there really need to be examined closely in terms of actual play among many groups, to see when and how they really pay off in terms of enjoyment. I know they can and do, but it's worth some real observation and discussion.

Best,
Ron

[Andrew Martin]

For instance, I have spent many an hour in utter amazement as I stared at endless charts of Rolemaster hit results, pondering your very point. Where the hell are they getting their data? And then it hit me: playtest!

Just a minor diversion, but I believe that the tables in Rolemasters were based on the results of rolling several damage dice at once, combined with a skill roll. Note that this is only a belief, I don't know if it's true or not.

[Andrew Martin]

For example, the simple matter of bleeding. How many GMs keep track of blood loss from wounds? Practically none, because the additional bookeeping is too much of a burden. Yet bleeding is only one of the issues involved in modeling wounding and its effects.

In a realistic game I haven't found this to be a problem, because bleeding is quickly first aided or magically or miraculously cured by PCs after a combat, and if not, it's no longer a cause of concern, usually because all the PCs are dead or dying.

Or take bullet damage. Bullet effects are exceedingly difficult to predict, one guy take 4 handgun slugs in the torso and WALKS three miles to the emergency room, another manages to kill himself with a Crossman B-B gun. Any rpg damage system to really model this would have to be much more complex than can be calculated by hand.
...
Given how difficult 'simulating' bullet wounds was for these real experts how much of a chance is some RPG designer who may have never fired a weapon supposed to design 'ex nihilio' a bullet damage simulation?  

I respectfully disagree. My combat system S on my site and discussed on these forums can achieve both results described. Also, "Guns, Guns, Guns!" by BTRC contains realistic gun damages based on the energy of the round.

Deep down virtually everyone knows that designers are making these things up by gueswork and intuition. And that goes for long charts of subtly varied modifiers for diffent conditions, guesswork, guesswork all.

I agree for the most part.

Anybody who claims to do better than 'great honking gun' and 'little dinky gun' is lying to you and wasting your time making you look up numbers that add up to nothing and making you pay to have them printed out (and that goes for sword and spears too).

I agree for the most part, provided you're talking about hand guns or pistols for the first part. For the second part, there's a difference between sword and spear, which is related to their length. Spears will strike first because of their length, and swords later; and this makes all the world of difference in a good combat system with skilled players/characters, as it usually gets to decide who lives and who dies.

So when somebody has taken a slug and wants to fire back and want to know how much it effects his aim what do we tell him?  Has the friendly game designer gone to some secret vault in the Pentagon where the keep the scholarly records on how much a slug to the gut effects your range performance? Or is he going to say, "Hmm four hit points must hurt a bunch I bet it would effect your aim a whole lot, lets call it -5 .....?

Modifiers on the victim's performance have to come from how well the slug has affected the victim. Doing it as you describe is just a silly rule present in too many RPGs. One can use real world tactics and then make sure that the combat system obeys those, and if it doesn't, redesign it so that it does work according to real world tactics. Of course, this usually means throwing away things like hit points, and then realising that most characters are much the same under the skin and can be killed by one shot.

[Peter Nordstrand]

Hi,

I am sure that mr Edwards will correct me if I'm wrong, but is "realism" necessarily a factor in Simulationism? Take, Pendragon or Feng Shui for example. In my humble opinion neither of these games are realistic, yet both are clearly simulationist; Pendragon simulates Arthurian romance, while Feng Shui simulates over the top Hong Kong action flicks. Granted, there are no tables in either game.

Cheers,

/Peter Nordstrand

[M. J. Young]

I'm guilty of that sort of abstraction of base damage values of which B. Bankhead complains. It is a useful abstraction, and in Multiverser it is a simplification that actually avoids detailed rules and charts:

[*]Weapons are placed in a "damage category"; there are seven of these, ranging from annoying to obliterative. Annoying weapons are those with which an inexperienced user probably will not hurt anyone, but might; obliterative weapons are those which will almost certainly destroy you if they don't miss completely. Most guns fall in one of three categories: damaging for small slow bullets, dangerous for bullets that are faster or larger, and lethal for those that are faster and larger. Rather than attempt to list hundreds of weapons, the system in essence tries to provide guidelines for placing a weapon in the appropriate category.
[*]A skilled user, whether through raw ability or through specific training, can bump a weapon up to a higher damage category. Chopsticks are not generally a very useful weapon, but a skilled martial arts expert can fairly reliably kill someone with them.
[*]Certain aspects of a target can reduce the damage potential of a weapon as well.
[/list:u]

But I understand that a significant part of any game is guesswork. It is probably not possible to design a game that 1) accurately simulates everything, 2) is run by a human without computer assistance, and 3) can cover an hour of game time in less than a day of real time.

Still, I wrote this post because I disagree that it is always guesswork. Charles Franklin, a retired marine, wrote a wonderful bit of simulationist wound mechanics for the first issue of http://welcome.to/cggzine/">The Way, the Truth, and the Dice (I believe it's entitled Hitting Them Where it Hurts but it's been a couple years since I read it), in which he worked from a century of military combat wound records analysis, recognizing that the levels of wounds have been consistent despite the advances in technology, and providing a system which ignores hit points and works instead with the nature of the wound. If you're looking for a way to better simulate combat outcomes, this is a great resource. I haven't used it because I'm comfortable with the abstraction.

--M. J. Young

[b_bankhead]

Before I comment on the individual responses it might be useful to mention my thoughts on the GNS typology.

  I think GNS is an exceedingly useful mechanism for discussing the qualities of different styles of gaming. But I have noticed a disparity between my understanding of it and many others.

  Frequently when people talk about Simulative,or Narrative, or Gamism they discuss them like they were discrete positions. My take on the matter is that ALL  rpgs contain narrativist AND gamist AND simulationist qualities. In my thinking the GNS typology is a cube with metrics of the various style running linearly along the 3 dimension of the cube.  In this typology any game is floating somewhere in that 3 dimension coordinate space, because all rpgs have the three aspects to some degree. I also believe there is a certain 'zero-sum' quality in that it is almost impossible for a game to serve all three 'masters' to the same degree...

  For myself the greatest pleasure is to be found along the narrativist axis. My loss of faith in simulatinism is based on the fact that the kind of hard simulation some people try to produce is either impossible or impractical for anything but the most simple of systems.  (I will be posting an article an article on my issues with gamism later). Also in my experience the primary axis in rpgs is narrative because its the element that is always called in whenever there is a failure in the other two (whether is fudging rolls because the simulation produced and accurate but undesireable result or because the gamist balance elements arent working for the particulary campaign element.)

Now that the theory out of the way ,lets look at the responses to my post...

[contracycle]

In a realistic game I haven't found this to be a problem, because bleeding is quickly first aided or magically or miraculously cured by PCs after a combat, and if not, it's no longer a cause of concern, usually because all the PCs are dead or dying.

I would be profoundly annoyed to find such a rule in a game - and I understand it is in one of yours.  You are making far too many assumptions about how people play; I have used blood-loss tracking system - and yes they are painful - but at times they have been absolutely necessary.  This sort of brushing over the issue is very un-Sim, IMO.

[b_bankhead]

Simulationism, though, isn't all about realism, or fine levels of granularity.  Its about ensuring cause and effect, at whatever level of abstraction you want, this is an important distinction.

I would agree however my argument is based on the premise that high levels of simulation are in practice impossible because of the lack of accurate information ,lack of mathematical model or because an model is just to complicated to be usable.

For my mind, the most interesting ruleset to me is the hybrid of light simulationism with narratavist elements*.  I'll probably piss someone off for saying this but I think a lot of the narratavist games out there are simulationist in many ways

I agee completly but that is because I think all rpgs are simulationist to some degree. It's just that I think only 'fuzzy' or abstract simulation is a really practical goal.

*Of course, this hybrid is very easy to turn to gamism with narratavist elements, something I am interested in seeing work.

I would also agee, but only because I think all rpgs have some gamist elements.

[b_bankhead]

For instance, I have spent many an hour in utter amazement as I stared at endless charts of Rolemaster hit results, pondering your very point. Where the hell are they getting their data? And then it hit me: playtest!

Ok, so that may not be an accurate model of the laws of physics of the real world. (Or it might be, by some incredible stroke of luck. Shakespeare and the monkeys and typewriters and all that...) But so what? It is an accurate model of the laws of physics of a world. A world that can be explored from a simulationist standpoint.

Rolemaster is a prime example of what I call a pseudo-simulationist system. It has all the charts and formulae and numbers of a complex simulation with actually simulating reality.  For example a good friend of mine showed that if you allow a housecat to get all the attacks it could theoretically get on the rolemaster small animal  charts. (4 claws and a bite) It works out that cats are actaully more dangerous to an unarmed man than daggers! (How many people would keep cats if this were true?)
I have never been able to see the point in traversing such a complex system that simulates nothing but itself.  That is one of the reasons I rejected D&D and Warhammer 40K.  All the complexity of a sim with all the realism of TOONS....

By the way, you seem to be more heavily against system-based simulationism.

  Thats because it usually doesn't work, or is too much of a pain to use..

But what about setting-based? Or character-based?
(I confess, I have no great interest in situation-based, and I can't quite conceive of color-based simulationism...)

 To my mind setting based simulation really comes down to a species of Narrativism.  And character based sims are usually intuitive and grainy, which is the kind of sim I think can be useful.

[Le Joueur]

Real quick; are you succumbing to the misnomer that Simulationism is in any way about simulation?  It's not.

This has always been a big gripe of mine, this terminology.  I understand the purpose in keeping it as it is, but I can only say this recurrent confusion only underscores the need for a new model.  That being said...

Simulationism is about exploration, plain and simple, of one or a few of Ron's list of Situation, System, Setting, Character, and Color.  The reason this is so confounding with that name is that, while cause and effect support it, they are not the primary interest.  The running guide to what is prioritized in Simulationism goes something like this; anything that doesn't prioritize deliberate 'literary-quality story' or 'player-skill-testing challenge.'  That tends to be just about anything, but gets categorized by the theory's author around the above five groups (or major combinations).

The reason we seem to be confounding realism with simulation has to do with the unusually large attention paid to the cause and effect significance.  First of all, no games I have ever seen simulate, there is almost no attempt to ration results based on the actual interaction of the non-trivial input circumstances.  What that leaves is emulation; kind of 'getting the same results' without figuring out why.  Don't get me wrong, emulation isn't a bad thing; without it, only chaos could ensue.

The problem is, who plays a game because it emulates reality well?  The only purpose in emulation is support (or potentially not interfere with) 'what the game is about.'  And that's where we get back to the SSSCC list.  I mean, sure, at the moment of decision it really looks like you're thinking "because that's how it works in reality," but that is chosen because it supports verisimilitude (or the sense that the game system isn't stealing the limelight over "what happening in the game").

Take Call of Cthulhu for example; it is described as classically Simulationist.  Are monstrous intelligent pools of slime realistic?  Do any actually exist?  Could they?  Speculation about what they could do and how they would act is just that, pure speculation.  You can't emulate it and you certainly can't simulate it, so how is that game a simulation?

It isn't.

What it is (as described by GNS theory) is an Exploration of Situation (and perhaps character for some).  Certainly, until you confront one of 'the great old ones,' you prioritize emulation of reality, but when you 'get there' realism goes out the window and you find yourself prioritizing the Situation.  A 'what would happen here, if I...' way of doing things.  That's why emulation is used; it serves this kind of Exploration, but isn't the end unto itself.

Now, if the emulation served the Exploration of Situation in that it is held to work primarily with an Edwardian Premise (let's be honest folks, we're well beyond Egri here), then it'd be Narrativism.  If the emulation served the Exploration of Situation to put the players to the challenge, then it'd be Gamism.  (Both as I understand them.)  But challenge and Edwardian Premise aren't important so that makes it Simulationism.

Back on topic, what B. Bankhead really seems to be complaining about is the failure of complex or 'highly explicit' game systems to emulate reality.  There seems to be the recognition that 'more rules' usually means 'more loopholes.'  The problem with that for the 'general audience' is not that it is not especially realistic in 'more explicit' systems, but that the verisimilitude fails some of the audience.  For Mr. Bankhead, many of these systems reach a point where they fail to capture his sensation of emulation of reality.  That's fine, but I believe most of the fans of those systems actually take the complexity in place of emulation.  It's not 'realistic' for our world, but it is real in 'that world' because 'those are the laws of physics.'

I think it's endemic amongst game designers, or future game designers, that 'what is already out there' isn't satisfying.  Reaching this conclusion is usually the first step to galvanizing oneself to create one's own system.  Congratulations B. Bankhead, welcome to the club.

Fang Langford

p. s. All of this is with the caveat that GNS is about describing the actual choices of actual people at the time they make them in hindsight at strictly the 'single decision' level.  To characterize anything larger than a single decision means suggesting a pattern stressing one type has arisen.

[b_bankhead]

I think I should also point out some factors in my thinking along this particular issue.
Part and parcel of my increasing embrace of narrativism has been my move toward more and more rules lite systems.

  1. My issue with pseudo simulation isnt just that it doesnt reproduce reality , its that it makes you go to so much work to create unrealistic results.
  2.  Since more complexity doesnt necessarily produce more realism, you can be just as unrealistic with a lot less work.

  3.  Effort spent bookeeping and other management of a complex and unrealistic system could be better spent producing more interesting stories,more detailed gameworlds, better npcs , in other words more creativity and less accounting.

   4. People who fall in love with the outcomes produced my pseudosimulationist systems can produce convoluted and completely spurious arguements for the 'realism' of the most counterintuitive,or just blatantly inaccurate of game tropes. Complex systems breed system fundamentalism in the minds of those taken in the illusion that  more_system=better_gaming. It leads to taking the system as something complete and perfect when it is anything but.

5.Complex systems are also based on the myth, "If I have a big enough book I'll have enough rules."  People who believe this are never dissuaded by the fact that after they have spent say $80 purchasing a rules set as long as DUNE they still run out and spend as much as several hundred dollars on supplements which contain even more rules!  If for example D&D3e doesnt have enough rules (the mere fact that they are buying supplements shows it doesnt) then I am force to the conclustion
YOU CAN NEVER HAVE ENOUGH RULES.

6. If YOU CAN NEVER HAVE ENOUGH RULES,then what you need beyond a basic minimum is a meta system for creating new rules, not a list of rules which by definition will never be enough.

7.Since there is no real utility to greater complexity the focus on the metasystems should be flexibility rather than complexity.
 

Anyway this is how my thinking on rpgs has evolved over the years, and this is where I am now....

[Marco]

So think of this next time you examine some complex simulationist style system filled with charts and rules.  Big columns of numbers arrayed all nice and neat can produce an illusion  of precision (and realism, whatever that is to you...) entirely unearned by any real world considerations....

This has nothing to do with simulationism--read what feng said. This (and your burn out) has to do (by my read of your comments) with demanding that a high-level of abstraction model be a really, really low level of abstraction model. That sounds like a receipe for frustration to me.

I find simulationist gaming to be the mode of play I prefer most of the time (and I don't mean on a game-by-game basis--I mean on a minute by minute basis--a gaming session can (and often does, IME) have elements of all three modes). If a given system is really weak at one of those (say, a hypothetical system where the player must always narrate the results of any setting exploration) it wouldn't work as well for me.

-Marco
[Note: Dunjon is, to a lesser degree, such a system--and I think it would be a lot of fun--but would be more along the lines of playing Munchkin or Frag during a gaming session than a replacement for the type of gaming I normally do.]