In Defense of Complexity

[Sidhain]

I think that's a good question:

[...]how do you design a simple combat system that still has sufficient tension for the player who can see through the simplicity, and yet is still reliable?


In general I don't know that you can design a completely transparent system, but you can work on making it as transluscent as possible. The factors that matter are often /really/ diverse and complex, and often way too huge to factor into a simple combat---essentially why we often reduce them to a die roll: Could the arrow be pushed aside by a freak gust of wind? Yes? Is it important enough factor to put into every game system? No. Thus we often just default to a die roll and let it fall out as it may.

But it depends on what you want from combat system, in some games (Hearts and Souls) I care more about the hero and his actions than any random factors so I rely on a fairly morphable result. While in my FRPG I do care somewhat about the heroes actions, but also on the mechanical feel--that wounds can kill you and are terrible, that some creatures are /tough/ and not simply fodder for your attacks. (One player commented on the difficulty of fighting Myrk, my goblin analog, I pointed out that--indeed that was the point, these are things they build huge walls around cities to /keep out/ because the average person can't fight one, let alone many.)

[Bankuei]

I think one important factor that we need to look at is what degree of Karma, and where it applies, goes into a resolution system.  

Take for instance, character creation, where you apply character effectiveness in certain areas.  This Karma is chosen before play, and usually isn't available to be manipulated in play.  

Second, you'd have in game Karma, which is the options and strategies you could apply during play to decide the outcome.

Third, we have Fortune, which affects the previous two(if they're part of the system at all), to varying degrees.

Amber focuses primarily on pre-play decisions for Karma, and then in play, trying to make sure you're in your field of expertise or politically manipulating others into handling your opposition.  In games such as Otherkind, Dust Devils, & TROS, a lot more emphasis is placed on the in-game decisions and strategy.

When you emphasize preplay decisions more in your design, the combat becomes more predictable, and more obvious to anyone who has the information(Amber-style Karma).  When you emphasize in play decisions, you have more gamist strategic play(Dust Devils).  When you emphasize fortune more, both Karma type decisions are reduced in effectiveness.

If you're talking about games that have simple resolution, but enough strategy to make a difference, try looking at Zenobia, TROS(without any manuevers), Hero Wars, Otherkind, and Draconic.  All of these focuse primarily on in game strategic choices, although pre-play choices can also make a big(but not overwhelming) factor.

Chris

[Ozymandias]

I don't mean that I want a system to be unreliable. I mean that I want a system that affords the seeming underdog a chance to overcome the obstacles because he knows or does something that the seemingly superior opponent can't or doesn't counter. If all our two fighters can do is swing away at each other, we've got a rather predictable outcome; but if one of them has the ability to sweep the legs of the other and so render him helpless, and the other has a dodging ability which allows him to avoid most blows, we've got a much more interesting and unpredictable--but still completely reliable--system.

I think I may have been working from an incorrect premise that "Complexity/Upredictability = Lots of Fortune". What you're talking about here could be done with a completely Karma based system, assuming that it includes some type of abilities/feats/etc that can be utilized to adjust the effectiveness of the attacker/defender up or down. This results in characters being conistently good at the things they're supposed to be good at without every result being pre-ordained.

Your mention of Amber though got me thinking though that perhaps there are some other factors to be considered here, especially in regards to tension....

What attracted me to posting is that, hidden amongst all this rhetoric, we are actually talking about tension. There's that moment, when the dice roll across the table, when nobody, not even the gamemaster, knows what's going to happen next; I call that tension. I find it highly engaging. Too often or for too long and it dissolves into chaos. To me, it is the primary reason for including dice in a game in the first place.

My question on this would be - "Is Fortune therefore the only or even the best source of tension in every game?"

If tension derives from not know what's going to happen next, I'd argue that it occurs all the time without rolling any dice at all.

In every mode of play, exploration plays an important role ergo players should be regularly be faced with the unknown and thereby not know what's going to happen next, is this tension?

On the flipside for the GM, everytime he faces his players with a difficult moral quandtry or a tough tactical situation, he doesn't know what's going to happen next, is this therefore also tension?

I would argue that both of those can result in as much if not more tension than simply rolling the dice, and for certain types of games they would provide a much more satisfying sort of tension.

I would advocate that tension derived from Fortune is primarily tension based on the possibility of incompetence, and that the more competent a game assumes the characters to be, the less Fortune that really needs to be involved. Perhaps this why on the really high-end of the PC competence scale (ie: Amber and Nobilis) it tends to dissapear completely. Tension can certainly exist in these games, it just has to come from somewhere else than the possibility of outright failure.

[Le Joueur]

...That is, provide any two opponents, and in a moment of reflection I could tell you which would most likely win under the system.

...As an overall argument I think saying that unpredictable=good and predictable=bad is a highly debatable proposition. In fact, what you seem to term predictable and unpredictable, I would call reliable and unreliable.

[About more karma based systems]...The answer is a foregone conclusion.

I don't mean that I want a system to be unreliable. I mean that I want a system that affords the seeming underdog a chance to overcome the obstacles because he knows or does something that the seemingly superior opponent can't or doesn't counter....

Thank you, Fang, for bringing out the aspect of tension. That's probably where the secret lies--how do you design a simple combat system that still has sufficient tension for the player who can see through the simplicity, and yet is still reliable?

I'd really have to say you guys aren't even talking about reliability or predictability; it really sounds like it's entirely a tension issue.  Let's frame a couple of new terms; how about obvious and unobvious?

In an obvious system, in the majority of cases, when two combatants face each other the result will be, well...obvious.  And that's what is sounds like you are arguing against Mark.  Certainly some simpler systems can yield that, but so can some complicated systems (often as a result of min-maxing).  A reliable system can be obvious or not and so can a predictable system.  (That's right, you can have an unobvious, predictable system.)

Mind you, it is quite common for a more complex system to be unobvious, but as this discussion is bearing out, this is not necessarily the case.  Likewise, in most cases to be looked at (such as tic-tac-toe), a simpler system is obvious, but there are clear counter examples (like chess).

So, Mark, is your leading article on this thread actually a suggestion that complexity is a good thing if it makes a system less obvious?  Or that you prefer systems that are more unobvious and that that usually comes not in a simple package?

What attracted me to posting is that, hidden amongst all this rhetoric, we are actually talking about tension. There's that moment, when the dice roll across the table, when nobody, not even the gamemaster, knows what's going to happen next; I call that tension. I find it highly engaging. Too often or for too long and it dissolves into chaos. To me, it is the primary reason for including dice in a game in the first place.

My question on this would be - "Is Fortune therefore the only or even the best source of tension in every game?"

Never the intent I had.  Dice rolling tension is simply the clearest and most easily explained; it is far from the most common or best source of tension.  However, it is by far the easiest for a designer to affect.

If tension derives from not knowing what's going to happen next, I'd argue that it occurs all the time without rolling any dice at all.

In every mode of play, exploration plays an important role ergo players should be regularly be faced with the unknown and thereby not know what's going to happen next, is this tension?

On the flipside for the GM, every time he faces his players with a difficult moral quandary or a tough tactical situation, he doesn't know what's going to happen next, is this therefore also tension?

I would argue that both of those can result in as much if not more tension than simply rolling the dice, and for certain types of games they would provide a much more satisfying sort of tension.

Absolutely, but sadly these are far from what a designer can 'build' into a game system (which is the subject of this discussion).  However, I believe that far too few games even actually address these in any substantive fashion (or at least didn't through the eighties, when I stop being able to afford to collect systems).  That's what I am trying to address with Scattershot and the Techniques like "The Suspense is Killing Me" and "Mystiques and Intrigue" and Sequences (listed at the bottom).

I would advocate that tension derived from Fortune is primarily tension based on the possibility of incompetence, and that the more competent a game assumes the characters to be, the less Fortune that really needs to be involved. Perhaps this why on the really high-end of the PC competence scale (ie: Amber and Nobilis) it tends to disappear completely. Tension can certainly exist in these games, it just has to come from somewhere else than the possibility of outright failure.

Since we're separating tension as a result of Mechanix and tension as a result of interaction, I'd like to say that the main difference (in traditional gaming) is how 'objective' it feels (and how this grants a sense of fairness or sharing).  If tension is derived from not knowing what the gamemaster has planned, it feels arbitrary.  If tension is derived from not knowing how the dice will fall, it feels objective.  (I say "feels" because, through manipulation of when the dice are used all objectivity can be eradicated if so desired.)

What 'Fortuneless' systems do is recognize that the group playing already has plenty of randomness; the behaviour of the people playing.  Amber isn't really about maneuvering situations so that you can win, it's about dealing with what the other players throw at you; that's where the unpredictability, the randomness, the unobviousness come from.  And just the same, it is unclear how to keep that from appearing arbitrary.  (I wish 'Fortuneless' games made more of an attempt at explaining that.)

How much the 'degree of incompetence' affects play with a Fortune mechanic is an attractive, but ultimately inaccurate way of measuring 'how much Fortune' a system has.  (It is quite easy to create systems that have lots of Fortune that impacts competence very little and vice versa.)  Once again you are actually talking about how obvious a system is.  It is quite possible to play a Fortuneless system in an unobvious fashion (but none I have read explain this very well) and the opposite is true too.  The other problem is that this also makes the assumption that mechanics are some objective 'laws of physics' mechanism that in some games they clearly are not.

I have to go back to my original point that how a game (not just the mechanics, but the 'how to play' part as well) handles tension is much more important than 'how complex,' 'how reliable,' 'how predictable,' or 'how obvious' it is.  All of these factors are completely subordinate to the 'when to employ tension' issue and, as such, are almost a meaningless comparison.  (Mainly because I can run any system, as written, with a completely different approach to 'how to handle tension' than you and give the players a totally opposite impression of the game, like or dislike.)  Yet how many games actually talk about not only 'how to handle tension,' but 'how your group likes it?'

Fang Langford

[M. J. Young]

So, Mark, is your leading article on this thread actually a suggestion that complexity is a good thing if it makes a system less obvious?  Or that you prefer systems that are more unobvious and that that usually comes not in a simple package?

I'm trying to distinguish the nuance of difference between those options. I think that obviousness is a flaw in a system, and that complexity can be a fix for that. I read a lot of arguments in favor of simplicity, but I've played simple games that were so obvious they weren't terribly fun. That's not to say that you couldn't design a simple game that is not obvious (and yet is reliable--I suspect unreliable games are never obvious). It is not, I think, easy to do.

As I was reading Ozzymandias' comments, I wanted to jump up and shout yes

What you're talking about here could be done with a completely Karma based system, assuming that it includes some type of abilities/feats/etc that can be utilized to adjust the effectiveness of the attacker/defender up or down. This results in characters being conistently good at the things they're supposed to be good at without every result being pre-ordained.

I immediately thought of chess, which you, Fang, mentioned. I don't know that chess is exactly a simple game system--each of six types of pieces have individual movement rules, at least two of them with rather complex special rules (the king, who cannot move into check, and the pawn, who cannot move diagonally except to take and cannot take when moving normally, although en passant will confuse even intermediate players). There is no fortune in it; it is all determined by the strategic abilities of the players. It is its complexity that makes it more challenging than checkers, in part because it is less obvious. Even if I know that David will always beat me at chess, I have no idea how that will happen--and with most players, I don't know who is going to win until very near the end of the game (although perhaps someone with better insight into the game would be able to recognize the outcome sooner).

I'm saying that people are dismissive of complex systems and supportive of simple ones without considering whether complexity has value; and that it often does have value in making outcomes reliable but not obvious.

Is that the answer to the question?

--M. J. Young

[Le Joueur]

I'm saying that people are dismissive of complex systems and supportive of simple ones without considering whether complexity has value; and that it often does have value in making outcomes reliable but not obvious.

Is that the answer to the question?

If you mean that you think that complex systems get a bad reputation because of the obvious ones and when people mishandle the tension, then yes.  If you mean that people support simple systems out of hand, despite the fact that so many of them tend towards obviousness, which you don't care for, then yes that answers it.

My turn for a question; is the 'obviousness' term what everyone else is talking about?  Or tension?  Or are we confusing reliability/complexity/predictability unnecessarily?  (And I suppose finally, are we just talking personaly preferences then?)

Fang Langford

p. s. I agree that unreliable yet obvious games are almost impossible.

[M. J. Young]

My turn for a question; is the 'obviousness' term what everyone else is talking about?  Or tension?  Or are we confusing reliability/complexity/predictability unnecessarily?  (And I suppose finally, are we just talking personaly preferences then?)

In WarGames, the girl (was it Ally Sheedy?) says of Tic Tac Toe that nobody plays it because you can't win (to which the answer is that the computer doesn't know that). In our terms, we would probably say that once you understand the game, the outcome becomes obvious; and that once the outcome is obvious, it's not fun.

So I take obviousness to be an inherent flaw in a game; but perhaps there might be priorities under which it would not be so, and I'm willing to consider that. (A narrativist might prefer a game in which there were no surprises, although I'm not terribly inclined to think so.)

Obviousness and tension would also seem to be related, although I know that with skill you can make it seem as if there is some doubt in regard to the outcome of a forgone conclusion. I seem to remember some Roger Ebert review of an historical movie in which he said the movie attempted to suggest that what we all knew happened might not actually have happened, but that in his mind this just sort of fell flat. I think it's the same in games--if you can clearly see what's going to happen, it's very difficult to maintain any real tension. It doesn't necessarily follow that unobviousness creates tension; that also requires that the players actually care about the outcome, which is something the game and referee have to cultivate for it to happen. But without doubt there can be no suspense, so I think the correlation real.

--M. J. Young

[Alan]

Hi all,

Can we make a distinction between complexity of rules verses complexity of the resulting interactions?  Current science is researching how complex phenomena result from a simple set of rules.  Wouldn't that be an game-system ideal to aim for?

[Andrew Martin]

Can we make a distinction between complexity of rules verses complexity of the resulting interactions? Current science is researching how complex phenomena result from a simple set of rules.  Wouldn't that be an game-system ideal to aim for?

Alan makes a very good point here. The science is called Chaos Theory, IIRC. Simple rules which result in complex behaviour which isn't obvious from those rules. Chaos. Emergent behaviour. Life (the computer game and the real thing). The edges of clouds, countries, and the sediment at pool edges. The behaviour of the flow of air around a wing, the water flow around a yacht's hull and the flow of product along a production line. Just to give people an idea of what it's like.

I feel that interesting games have simple rules with complex behaviour generated through using those rules.

[Christoffer Lernö]

Can I agree with Alan and Andrew here (this might be going off topic)?

From my experience it's a battle-upstreams to create a game which has high complexity directly in the core rules. The overhead to avoid conflicts and exploits grow ever bigger as you try to make it more detailed.

Starting from high complexity one tends to lean towards simplification, right? For example figuring out optimal strategies and combinations just like Andrew already said in his post about 4 levels of complexity.

Since high complexity usually is motivated because it allows the player to do a wide range of things in high detail (that's why I did it anyway), it falls because the player tends to go for a few (optimal) actions.

With simple rules that can generate complex results this problem simply falls away.

Or in short I agree because these are my observations as well.

[Mike Holmes]

OK, missed a lot here so far, but here's how I weigh in.

First, there is no anti-complexity bias here at the Forge. There may be single individuals who dislike complexity, but there are just as many, myself included, who like complexity. In case anyone is interpereting my Combat Systems rant as an argument against complex combat systems, it is not. And never has been. If anyone reads it, and thinks that's the intent, they ought to re-read it again, along with the thread that Shreyas linked to above (thanks, S.).

Ok, that bit of self-centered comentary out of the way, on to the main point. A game needs exactly as much complexity as it needs to accompolish it's goals, and no more. OK, that's tautological, but I'm trying to make a point. Which is that any level of complexity can be suitable for a particular design. The questions that everybody raises above about Tension, Predictability, whatever, can all be seen from different viewpoints. I actually expect that out there, somewhere, is a gamer, Fang, who wants no tension at all in his RPG. And someone ought to design a game for him.

So we can have opinions as to which of these are particularly important to our particular designs, but I doubt that there is some Sumum Bonum that we are going to find by discussing this in further detail.

>>>Huge Tangent<<<
Especially since it's beeen discussed to death by minds much better than ours. What we've delved into here is classic Game Theory. Read your Von Neumann, et al (I'm always amazed that people don't realize that the term Min-Max is from a paper of his from the earlier half of the last century). They talk about "True Games" and "Perfect Games" with much more clarity and precision that I think we could hope to.

For example, one thing I remember from the meager amount of Game Theory that I've retained, is that the most important part of such games, is that the outcome be based, not solely on the outcome of the decisions of one player, and not just on that of multiple players, but on the interaction of those decisions.

Take, for example, D&D. You have the player decision, opposed by the Game's decision (selected by the GM as theoretically neutral interpreter). To a small extent, these things interact over time in terms of strategy. But rarely is the decision of one side to attack affected by any decision of the other side. I like to use bowling as an analogy (as game theorists often use it as an example of something that's not a game). D&D is most often simply a bowling match with each player trying to get a score equal to the opponent's HP before the other does the same. Or, IOW, it's not a "good game" (in the Game Theory sense) on the tactical scale.

As opposed to, say, TROS (or football, for a sporting comparison), where almost all outcomes are affected by both side's decisions. The Game Theorists would say this is a better game.

Note that this is true mostly because it then resembles an economic model, Game Theory is used to model economics as it's practical outcome. And as such, one can see that in something as complex as a market that there is a necessity to have models that involve these sorts of relativvely complex interactions.

So, the question of applicability of Game Theory here is valid. However, we can skip a lot of the bullshit of the discussion, and get right down to the heart of the matter if we were to all just get a little more up to date on our reading of Game Theory. From that basis, one can discuss such matters a bit more clearly and athoratatively, IMO.
>>>End Tangent<<<

Anyhow, it seems to me, again, that a game needs just as much complexity as it needs to accomplish it's goals. If one of those goals is a lack of easy predictability, that's fine. But that still doesn't require too much complexity neccessarily. The sorts of interactive complexity that's discussed in Game Theory can be brought about by a set of rules as simple as Chess.

So it sounds to me like a good rule of thumb is to try to create a game with that sort of dynamic where the classic phrase can be used to describe it. "Easy to learn, difficult to master". Often said of Chess, but applies even moreso to a game like Go, which has even fewer rules.

But that all assumes that this is even a goal. Again, I can see someone wanting a "game" for which it is easy to see the outcomes. For example, a Narativist game, in which the conflict resolution was not the central source of Tension. Indeed, I'd venture to say that the conflict resolution system is not even close to the central part of Tension in most Narrativist games. Take Sorcerer, for instance. Do we hang on the outcome of the die-rolls? No, they are almost an afterthought, IMO. The important parts come well before and after the dierolls, and take the form of the decisions that the characters make. As such, in that case, you want a system like Sorcerer which heightens the Tension by making repercussions for choices. A system that sought to emulate that could easily forget about the whole question of predictability vs uncertainty in resolution systems.

All just seems like style preference to me.

Mike