Response to Danc’s Loops and Arcs

I wrote a really long response to Danc’s post Loops and Arcs. It was so long his blog rejected it. So I’m posting it here. Please read his excellent article first if you want this to make any sense.

My Response Follows

Cool, Danc. A few thoughts I had while reading.

First thought:

Interesting approach to the topic of systemic/predefined. I conceptualize this differently – not as a dichotomoy of kinds of experience, but as a spectrum between predefined experiences and experiences which are generated wholly on the fly.

Going from left to right: Film, Call of Duty 4 single player, StarCraft II single player, World of WarCraft, Call of Duty 4 multiplayer, chess, The Sims, MineCraft, real life.

Since some aspect of nearly every game is authored – see the names of the pieces in chess. And since the possibility space of any system (besides real life) is constrained by an author’s decisions about how to build the system, I don’t think it’s right to try to categorize pieces of games cleanly into these two kinds. It seems destructively reductive.

The experience a player gets is always a combination of what the systemic elements of the game generate on the fly (as constrained by their design) and the pre-defined elements authored in (whether they be a character model or an intro movie).

Obviously the experience transforms itself on repeated plays to the degree that it is generated on the fly.

Second thought:

I don’t think that “loops”, as you’re calling them, are nearly as disliked in the mainstream game market as you’re saying. Call of Duty for example, is marketed with videos from single player, but people play it for MP. Ditto Halo. MineCraft, The Sims series, World of WarCraft, StarCraft II, and lots of other really popular titles are very heavily driven by extremely strong systems. Hell, we can throw Draw Something on the pile.

It’s true that systemic design is often underemphasized by the paymasters of the game industry. Publishers and businessmen typically cue into single-threaded story records instead of thread-making story generators. It’s what they know, and it looks good in a pitch meeting. But I think the market itself responds quite well to a well-designed system. Always has. If they want arcs they’ll watch movies (and most of them do).

Third thought:

I don’t think that your description of religious activity as a “loop” is really fruitful. I think what you’re calling a loop is an authored, self-contained system that, when interacted with, spits out experiences. At least that’s how I read it. Religion is an organic cluster of activities, which are not self-contained or authored in a clear way and are designed more to propagate themselves than to spit out experiences. If you’re going to include religion in the definition you might as well also throw in every other interactive part of human life. That is, everything. See my first thought on the spectrum from loop to arc.

Fourth thought:

My favorite thought in the article was this: “Of the two, loops are rarely discussed in any logical fashion. People note the arcs and comment on them at length while being quite blind to the loops driving the outcomes.”

Now that’s interesting. To me this has to do with the human mind’s obsession with single-threaded stories and lack of natural ability to think in terms of possibility spaces and statistics. This prediliction expresses itself in, say, politics with an undue emphasis on individual situations (“I met a kid who couldn’t afford college…”) instead of the broader statistical reality (“X% of kids can’t afford college…”). And there’s not a chance in hell of hearing about the systemic reality behind those individual or statistical outcomes (“The college affordability rate is driven by…”). Basically we love images of things happening one by one. It’s why 9/11 scared people but car crashes don’t, even though a half million Americans have died in their cars since 9/11.

So everyone talks about games in terms of the threads they generated in their particular playthroughs, instead of the broader statistical reality of what threads the game will tend to generate over all playthroughs with all players. Good stuff.


Overall, a thought-provoking article. Feel free to rip me a new one if I misinterpreted you anywhere.


4 thoughts on “Response to Danc’s Loops and Arcs

  1. Danc

    Good comment.

    Three semi-random comments sparked by your thoughts. :-)

    1) You can indeed define a spectrum based off the ratio of loops / arcs in a particular work. However, that often isn’t the most fruitful usage. Instead, think of them as ingredient or components that can be mixed and matched. Component/Inventor thinking vs Categorization/Curator thinking. How are the arcs used and to what purpose? How do the loops work in a particular project? Are the arcs necessary? How might you close them? How might you make them evergreen? This goes far beyond a simple categorization tool. In the end categories are often overly broad and do little to bring a specific game to life.

    2) The reason I introduced religion is because loops and arcs are common structures throughout human activities. Games are one activity where one can spot loops, but it is by no means the only one. You are completely right that you can look at almost any interactive experience.

    3) I don’t know if I’d consider games as things that ‘spit out experiences’. That’s a very media-centric view of games that I don’t really subscribe to. That is indeed one result, but not necessarily the only goal for games. A game can also be a tool that lets players learn, experiment or build social bonds. Experiences result from this, but then again experiences result from any interaction with the world. You could just as easily claim that a car spits out experiences, but that is not its only utility.

    All the best,

  2. Tynan Sylvester Post author

    From 1)

    Okay. This is interesting. So we’re splitting a game into pieces and calling them loops or arcs. I’m trying to do this in my head and coming up with difficulties. Take a look.

    What is the role of the authored limits or tendencies of a dynamic system? e.g. Chess tends to play out with a certain pacing, and the same kinds of things tend to happen. They’re not arcs because they’re not written down event by event. Yet at the same time these patterns of repeated similar situations in the possibility space are authored in through the design of the system.

    As another example, Left 4 Dead is a pretty systems-heavy game, but the situations it tends to generate almost always fall into the same few categories (survivor down, needs help, etc). Yet, those situations are never actually written down anywhere. Where are the loops and arcs?

    Another thing I’m having trouble on: the representational layer. The king is chess is called a king. The marine in SC2 is called a marine and looks like a marine, even though he’s just a circle with numbers attached. Where are the loops and arcs? Is the king’s abstract entity a loop, while the king’s representation and name is an arc?

    I’m sitting here trying to break games apart according to these pieces and finding it difficult. Not saying it’s impossible, of course. Perhaps I just suck at it.

    What would really help crystallize this is a detailed breakdown of a nontrivial game into loop/arc pieces. Left 4 Dead would be a good challenge case, I think.

  3. Jacob Rummelhart

    Sorry for coming into this a bit late, but I just stumbled onto it today.

    Tynan, I think you’re trying to break things down into pieces that are too small. The loop/arc only deals with the interaction of playing a game, not with the individual pieces that are included in the game. Every loop or arc must consist of a model, an action, a system, and feedback. A king in chess is neither a loop nor an arc, because it is only a small part of the interaction.

    Model: Based on what you know of the game, “if I move the king here, it will be safe.”
    Action: Move the king to that spot.
    System: Each chess piece has a rule about where it can move. Based on those rules, a bishop can move to a place that puts you in check.
    Feedback: The other player puts you in check, and says “check”. This is feedback that you probably didn’t make the right move.
    Model: Incorporate that new information to see if you can move somewhere better on this next turn.

    That’s a loop, because it feeds back into itself. (It wasn’t authored, either, because it’s up to the player(s) to determine what the outcome will be.) Loops/arcs in L4D could be broken down like this as well, but it would be a much longer example, due to of all of the overlapping elements involved.

    Let me know if that made no sense at all. And hopefully, I’m accurately representing Dan’s ideas. Good luck!


  4. Tynan Sylvester Post author

    JR, Indeed, you’re correct. I talked to Danc about this a bit more and he clarified that the loop/arc distinction is intended to apply at the subsystem level and is mostly focused on how players interact with the game, not necessarily on the design itself.

    So I suppose in chess, there is really only one subsystem, the “piece moving loop”, which you identified here. The game has lots of predefined elements and even semi-authored tendencies (e4 d5 opening), but these aren’t arcs/loops because they’re not really subsystems of the player experience, if that makes any sense.

    I still think there are definition attacks you could make against the distinction. e.g. How would you classify a quicktime event sequence? There are a lot of subsystems/sub-experiences in game that mix the concepts of feedback-driven system and predefined output quite freely. That’s not to say I don’t think there’s a lot of truth in what Danc wrote. Just that, as in all game design theories, we find caveats in this one. And I like stressing theories to see if they hold up, just for funsies.

    By the way, thanks for dropping by.

Comments are closed.