In Defense of Complexity

[M. J. Young]

Periodically I read threads here at The Forge in which people argue against more complex systems, and particularly more complex combat systems. I take these with some bit of salt; but I hadn't really thought much about them. I think of Multiverser's system as being relatively simple at the core with plenty of room for what might be termed "voluntary complexity"--that is, if the players want to make it more complicated, it is within their power to do so.

Perhaps I was recently looking (elsewhere) at someone's relatively simple combat system, and thinking that it had this factor of the obvious--that is, provide any two opponents, and in a moment of reflection I could tell you which would most likely win under the system. But I can't do that quite so easily with a more complex game. For example, D&D has a certain amount of complexity in combat because it is an interaction between 1) chance to hit (including defensive values); 2) ratio of attacks (how many swings each side gets relative to the other); 3) range of damage; and 4) durability (hit points). Yet even at this level of complexity, given two opponents, five minutes, and some scrap paper, I can work out who will win in an ordinary melee combat with some certainty.

But throw in a few tweaks, and the complexity rises--and with it the unpredictability of the situation. If one of my combatants has a dodging skill that enables him to avoid many of the attacks of the other when it works, probability of a successful attack for his opponent drops precipitously. If there is the possibility of a stun or knockout blow, durability suddenly becomes a less important factor. If my combat style allows me to trade attacks for damage or damage for attacks, I may be thus able to tweak my performance on the fly--that is, whether one attack for d10 points is better or worse than two attacks for d6 points each may well depend on whether rolling twice to attack gives me a better chance to connect at all or if my attacks are all pretty likely to hit, and the ability to make that choice round-by-round becomes an important tactical consideration. Suddenly it is not so obvious who will win; my scratch pad must give way to a computer program.

Thus complexity in the system in itself may be a way to inhibit min/maxing. The player can't easily evaluate what choices will make his the "most powerful" character in play, because there are more variables than he can easily process. In D&D, we can wonder whether the character with better armor class or more hit points is more survivable; but it's something that can be fairly easily calculated. The more complicated the system is, the less likely it is that a player can read it well enough to beat it.

I like simple systems; I like complex ones also. Handling time is a negative of complexity; predictability is a negative of simplicity. In the end, there is balance--each is a defensible choice, depending on what you want the system to do.

Most of you design; I'm interested in your thoughts on this.

--M. J. Young

[Paul Czege]

Hey M.J.,

Thus complexity in the system in itself may be a way to inhibit min/maxing. The player can't easily evaluate what choices will make his the "most powerful" character in play, because there are more variables than he can easily process.

This presumes the player is working toward advantage in isolation. As with Magic: the Gathering, killer combinations become known throughout the player network, and the whole player network ultimately benefits from the separate efforts of all on behalf of advantage.

Paul

[Shreyas Sampat]

I'm not really sure what you're getting at here.

Sure each pole, and the points on the continuum, are defensible choices.

I think that there is an important, and valuable, negative perception against making complex systems without choosing to.  That's a genuine error, IMO.

You might want to take a look at this thread - Anti-Combat Bias - and see if it addresses the points you're thinking about.

[Jack Spencer Jr]

.... predictability is a negative of simplicity

I will disagree with this statement. I would say you can have a very simple system and still have a great deal of unpredictability. What may be the problem, and this is just my guessing, is that a lot of these simpler systems are just -- for lack of a better way to put it -- "dumbed down" complex systems. That is, the simple system shares many logistics with more complex ones, but has fewer variables, and with fewer variables, there is less of a balancing factor and one or two variables can outweight the others and then that's it.

Is this off base?

[Bankuei]

I'd agree with Jack, a simple system could be as highly unpredictable as the dice system in Sorcerer.  What is more of an issue, I think, is the strategizing and tactics factor, associated with Gamism.

A complex game(be it combat, politics, whatever), has several elements which interact in potentially complex manners, requiring a higher level a strategizing.  To give an example, a simple system is more like checkers, with a fairly simple strategy, while chess has several different factors(the different pieces move in different manners), which leads to more complicated strategies.  OF course, a simple game such as Go/Wei Chi has incredibly intricate strategies, but I think few people actually design rpgs hoping to make something as deep as that.

As someone who enjoys gamist elements, I find that any game that has a very limited "success" strategy usually leaves me unsatisfied, while one that has a lot of room to learn and explore, a perviness to it, can keep me occupied for a long time.  I enjoy the elements of strategy and counterstrategy, a need to adapt in game to overcome whatever obstacles may appear.

Chris

[Andrew Martin]

There's four degrees of "complexity" in a game:

[*]Simplistic;
[*]Complex;
[*]Overly Complex and
[*]Simple.
[/list:u]

A Simplistic game requires simplistic strategies to win, for example, charge straight ahead, or always choose All Out Attack (in Usagi Yojimbo RPG).

A Complex game has a straight forward strategy to win, like, all PCs concentrate on the biggest monster first (AD&D), break off if all weapons bounce, fireball first when facing massed encounters, don't bother with shields unless they're highly magical, jump off cliffs to escape encounters. :)

A Overly Complex game has one or more "game breaking" ways of winning the game which munchkins, power gamers and rules lawyers usually find in short order once they get hold of the rules. Examples are things like in RoleMaster (with all the expansions) choosing to play an archmage with a special spell list which allows enormous amounts of spell power to be built up. Or playing the noble warrior class, which allows enormous numbers of attacks per round. In WW's Exalted, there's a combination of two stones which allows PCs to become immune to damage in combat forever; a fellow munchkin found it the first time he generated his character.

A Simple game has a range of strategies, from simplistic, through complex, to simple. The end result is like looking at a fractal; it's basic description is simple, but the expression is complex. Chess and Go have this property, as does real combat (just read descriptions of it in martial arts books, comics and in history books!). My S combat system has this behaviour as well, allowing players to model the behaviour of real combatants. That's because I used a fencing model which is like a four way paper, scissors, rock game.

[Thierry Michel]

The player can't easily evaluate what choices will make his the "most powerful" character in play, because there are more variables than he can easily process.

Well, I don't design, but isn't that somehow a failure ? As a wargamer, it reminds me of the "rule lawyer" opponent who would beat me not necessarily because he was better but because he knew the quirks of the rules better than I.

Wouldn't be better to offer meaningful choices in terms of risk/rewards of each decision ?  

(That said I never really understood how combat sytems could be "tactical", but that's another thread entirely).

[xiombarg]

I'd agree with Jack, a simple system could be as highly unpredictable as the dice system in Sorcerer.  What is more of an issue, I think, is the strategizing and tactics factor, associated with Gamism.

I would like to note, as a quick aside, that I think that it's possible to have a simple system  (in the sense of a small, easy-to-understand ruleset) that provides complicated, difficult-to-predict Gamist tactical and strategic options.

The best example of this, admittedly outside of RPGs, is Go.

So, I'm agreeing with Jack and Bankuei -- M.J. dichotomy between complex=difficult to guess with large amounts of tactical options and simple=easy to guess is false. However, just as it's difficult to design a board game as good as Go, it is difficult to design a simple RPG system that meets the goals of unpredictability and tactical richness. But that isn't to say it can't be done, or hasn't already been done. In fact, I would submit it's an admirable goal, and a lot of the "simple" systems out there are attempting to grasp it in some way.

Edit: I just now saw Andrew Martin's post, which perhaps says what I'm saying, but more eloquently. :)

[Jack Spencer Jr]

Here's a question: What makes a system complex or simple, really? I think this is a slippery slope since it is only in the eye of the beholder and a matter of comparason that any game is simple or complex. I mean, Checkers is pretty simple. Chess is a bit more complicated in comparason, but is is a "complex game?" I don't believe so. I think it's a line subjectively drawn in the sand. and we all put it somewhere else (and I suspect we'd put it in different places at different times of the day)

In fact, I daresay it is more useful to look at Andrew's categories as lines that can be crossed from good design perspective. You don't want to go into simplistic or overly complex in your designs, because the design would be lacking in some way. Or such is my view.

[xiombarg]

In fact, I daresay it is more useful to look at Andrew's categories as lines that can be crossed from good design perspective. You don't want to go into simplistic or overly complex in your designs, because the design would be lacking in some way. Or such is my view.

Oh, I agree. I made my point to say that MJ's dichotomy between simple and complex is false, to highlight the usefulness of Andrew's more balanced perspective. I apologize if I was unclear...

It's certainly the case that simplicity is a matter of taste. A lot of gamers I know consider d20 to be "simple". One only has to look at some of the commentary on the d20 threads here on the Forge to find people who disagree...

[Jason Lee]

Here's a question: What makes a system complex or simple, really? I think this is a slippery slope since it is only in the eye of the beholder and a matter of comparason that any game is simple or complex. I mean, Checkers is pretty simple. Chess is a bit more complicated in comparason, but is is a "complex game?" I don't believe so. I think it's a line subjectively drawn in the sand. and we all put it somewhere else (and I suspect we'd put it in different places at different times of the day)

Maybe we need to define the complexity we are addressing.  Combat systems I take it, but what part?

I agree with this fellow on his definitions of complexity, even if it does suffer a little bit from the seemly innate human need to group things into sets of three (small, medium, large / left, right, observer / GNS)

http://www.rpg.net/news+reviews/columns/elements20sep02.html

Maybe their is a hybrid-ization of terms we can perform with Andrew's terms and those in the above article...so we know exactly what we are talking about.

To hit M.J.'s topic, ever so slightly:  Using the definitions in the above article I favor Simple Implementation and Complex Concept.  I prefer Complex Mass because I prefer campaigns, but if I wasn't playing a campaign I'd lean toward Simple Mass.

[Andrew Martin]

...Andrew's terms...

I based these terms on Sergio's column on RPG.net here: http://www.rpg.net/news+reviews/columns/ruleslaw28jun01.html, in The Design Cycle Approach section. The last step from Over-complex to simple is the important step; it's the refactoring that does the trick. Here's refactoring in computer software: http://www.c2.com/cgi/wiki?WhatIsRefactoring; it's analogous to going from overly complex to simple.

[Ozymandias]

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

But is that really a bad thing? Shouldn't a better combatant generally win?  As an overall 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.

My experience with unreliable/unpredictable systems has been universally bad. The net effect of too much randomness seems to be that a character is stripped of the ability to do even the things they're supposed to be good at with any reliabilty. Now, admittedly in some cases this is contributed to by bad design (ie: The Storyteller system, where dice pools over eight dice start progressively having a lower chance of success), but if there's considerable chance for either success or failure on every action it becomes impossible to create a character who is reliably good at anything.

On the flip side, I've played some really good games with systems with an extremely low unpredictability/unreliability factor (ie: Conspiracy X) b/c it allowed the character to actually do the things they were supposed to be able to do.

So, I'm not sure that unpredictability really provides a good defense for more complex systems.

[Le Joueur]

Y'know, you can't really just jump in assuming a game is either predictable or not.  It's just not that simple; some games are predictable in one area and not in others, some are so-so predictable, some are more predictable than others.

And I simply refute the idea that there is any correlation between predictable and reliable; these two ideas are not so connected.

Simply put, this side of the argument is sliding quickly into "I don't like that system because it's too unpredictable (unreliable) [to me]."  Personal preferences are fine, but not in a discussion about complexity.

Let's take a moment and think here.  We all pretty much agree that it probably isn't any fun to play a game which you know, in detail, what the outcome is from the very beginning; that's one extreme, let's call it 'preordained.'  On the other hand, it probably isn't any fun if the game renders completely unexpected results at every turn; that's the other extreme, let's call that 'dada.'  The problem is this is such a wide range that it not only becomes totally subjective, but pointless to try and define what is 'complex' and what isn't.  People have different tastes (and if that's the point of this thread, let's call it done).

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.

The ultimate problem I see in many designs of games it how tension is handled.  Or rather isn't.  How many games can you count that talk about the subtle social situation that requires or bars a certain level of tension?  Too much, or too little, at the wrong time will wreck any game, but how many designs talk about how you handle this?  So many games give you the option of using dice, 'should you roll or not,' but no instruction as to what is too often or not often enough.

That's why any discussion of complexity in game design is clouded by the fact that there are no 'handling instructions' for tension.  Because it becomes quite crucial to different people's experiences with different games, how badly burned they are by misplayed (and this is both highly subjective and completely relative) tension.

Until you separate complexity from predictability, predictability from reliability, and complexity from 'handling instructions' for tension, I can't really see much value in continuing this thread.

Fang Langford

p. s. Sorry I didn't edit, I gotta run.

[M. J. Young]

"The race is not always to the swift, nor the battle to the strong, but that's the way to bet."--Damon Runyon

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

But is that really a bad thing? Shouldn't a better combatant generally win?  As an overall 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.

I think that what I'm seeing is not the reliability of the system but the degree to which a reliable system can still be unpredictable. Real fights are generally won and lost on intricate combinations of abilities and chance. Looking at military combat, numbers, supplies, position, tactics, and morale at least have impact on the outcome. It is certainly as much so for individual fights.

A simple karma system that said, "compare strengths (or combat ratings, or something like that) and the highest wins" would be totally reliable and totally predictable. My impression of Amber is that the way to play it successfully is to maneuver your opponents into a situation in which you are able to use your strengths against their weaknesses; even so, once it is determined in what realm the battle will occur, 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. 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.

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?

--M. J. Young