Monday, April 2, 2018

Juul's Theory of Games Elements 2 and 3: Outcomes

This is part 3 of my analysis of Jesper Juul's (2003) article The Game, the Player, the World: Looking for a Heart of Gameness The first part is here, the second is here. The original article by Juul is online here (page numbers are from the PDF version).

Juul, as noted previously, outlines six criteria that he holds up as being necessary and sufficient to label an activity as a game (or not). Yet, as I noted when discussing the first one (fixed rules), he mentions several 'borderline' cases, including tabletop RPGs, which still seem to be games, but don't conform to one or more of the elements of his definition. My hypothesis is that each element is a cline or gradient, and the dials can be turned up or down along the axis of each of these elements. In this post, I'm looking at his second and third elements, which are related: Variable and Quantifiable Outcomes and Valorization of Outcomes.

Element 2: Variable and Quantifiable Outcomes

By variable outcome, Juul means two things: first the system of the game must allow for chance and/or skill to affect the result; and second the skill level of the players must be such that the outcome is uncertain.

He doesn't spend much time on the first point, only giving the example of ring-a-ring-a-rosies (or as we called it as kids, ring-around-the-rosie). For young children, this is a fun activity. It has set rules (link hands in a circle, sing the song, move clockwise as you sing, fall down at the end). But it only has one outcome: everyone sits down. So, by this definition, it is not a game.

The second point he goes into more detail. He again uses a game for young children, tic-tac-toe. Anyone who knows the system well and plays against another player who knows the system well will ALWAYS result in a draw. Someone who knows the system (an older child or adult) who plays against someone who doesn't (a young child) will always win or draw, never lose. Only two players who do not yet fully understand the system will have a chance of winning, losing or drawing when they play. Hence, tic-tac-toe is a game, but only when played by young children or others who do not yet fully understand it. Once you learn the system, it ceases to be a game for you and becomes more of a puzzle.

He then mentions several games that have handicap rules to allow players of various skill levels to compete. Without the handicaps in place, those instances of play are unlikely to have variable outcomes. Related to this is the phenomenon of a player purposefully not playing at full potential, such as a parent purposefully allowing their child to win a game in order to build interest and confidence in the activity.

This reminds me of my college days. One of the students in my dorm, Eugene, was a chess master. We'd play from time to time, and of course I knew going in that there was no way I could beat him. I just play chess casually. He competed in and sometimes even won tournaments. Yet, it didn't stop me from playing with him now and then. And there WAS a variable outcome. It just wasn't 'win/lose.' Instead, the variable outcome was to see how many moves I could last before the inevitable defeat.

So, again, I think Juul may be being too strict when he posits an either/or dichotomy in his definition of games. This also leads into the second part of this element is what he terms "quantifiable" outcomes. Just as a game has strict rules, Juul says that "the outcome of a game is designed to be beyond discussion, meaning that the goal of Pac Man is to get many points, rather than to 'move in a pretty way'" (p. 6). In other words, outcomes are objective rather than subjective. To me, this seems fairly non-controversial. Game systems define the 'win conditions' but it's also possible for players to decide on their own 'win conditions.' 

The obvious example of computer games without quantifiable win conditions seems to be Will Wright's games like Sim City and The Sims. There are no intrinsic goals for these games. Nothing within the game will tell you that you've won or lost. But players are able so self-assign goals and try to achieve them or fail to do so. Juul puts them as borderline games because of Element 3 (see below), but several other writers I've been looking at, including the creator Will Wright, see them as toys rather than games because of this trait. 

Element 3: Valorization of the Outcome

Valorization is a fancy term to use, but what Juul simply means is that for an activity to be a game, some outcomes need to be better than others. If all outcomes are equally good, then there is no point to the challenge of the game. In Game Theory (which both is and isn't about games), Zero Sum and Fixed Sum games are set up so that one player can only achieve a positive outcome if the other player achieves a negative outcome (or at least a less positive outcome for Fixed Sum games). In cooperative games, all players work together to either beat the game or fail to beat the game. In RPGs, you can slay the dragon/blow up the space fortress/prevent the Great Old One from rising. Or, you can get roasted to a crisp/get disintegrated by droid ships/go insane and join the cult. In many RPGs and computer games, live or die is a commonly valorized dichotomy of outcomes. 

Again, an open ended simulation like The Sims doesn't valorize outcomes. While it's possible to give yourself goals within the game, there's no pressure to actually achieve them other than personal satisfacton, and no award of any kind when you (inevitably unless you choose to give up) achieve that goal. Juul also notes Conway's Game of Life (the computer microbe simulator) and watching a burning fireplace as systems with rules and variable outcomes but without any positive or negative value assigned to those outcomes.

There is a tendency that the positive outcomes are harder to reach than the negative outcomes - this is what makes a game challenging; a game where it was easier to reach the goal than not to reach it would likely not be played very much. (Juul, 2003, p. 7)
This is an important point. Especially when we're talking RPGs. I'll have a bit more to say about this below.

This element is, in my estimation, an actual dichotomy, rather than a cline. I'm having trouble imagining an activity where the outcomes are all somewhere equally positive or equally negative yet still being able to call such an activity a game. If all outcomes have the same value, then where is the challenge? The activity may be play, but it is not, I think, a game. If someone can give me an example of a game with only partially valorized outcomes, I'll be happy to change my stance here.

What these Elements Have to do with RPGs

First of all, I don't think there's any doubt that traditional pen and paper/tabletop RPGs like D&D conform to both of these elements. There are definitely multiple outcomes, and a combination of chance and skill determine what outcome you achieve. There are system outcomes like how much XP you gain, or whatever advancement reward system the game uses which is variable and quantifiable. In addition, in-game outcomes - save the town/defeat the bad guy/complete the quest - are also variable and quantifiable. Basically all of these outcomes, whether looking at system or in-game, break down into outcomes where you advance your character/continue playing it (positive) or else fail to gain XP or die (negative).

Narrativist/Indie/Story games, however, are a bit different (remember, this line of posts was inspired by Zak's close analysis and criticism of GNS theory). Now, my experience playing them is very limited, but from what I've experienced and what I've read about them, the system of those games is probably higher up the scale of Fixed Rules (element 1) than traditional TTRPGs. This is because the rules of these games aren't designed to provide guidance for actions taking place within the in-game fiction. They're designed to assign narrative agency to various players. In these games, when the rules come into play, you're not rolling to see if you can jump across the chasm, you're rolling to see which person sitting around the table gets to decide if you manage to jump across the chasm and what complications or benefits may arise because of the jump. [Again, people with more experience in these games correct me if I'm wrong here.]

The stated outcome of these games is to, through play, craft a compelling narrative about a theme. As Zak posted, Ron Edwards originally believed a morality play to be the highest form of the genre but other game designers have moved away from the explicit moral judgment of the narrative. Which is funny, because Edwards' original intent was an attempt at valorization of the outcome while actual practice (if Zak is correct, and my limited experience agrees with his) has tended to unvalorize the outcomes. The goal is simply to craft a narrative. If a narrative is crafted, the goal is achieved. And I can't help but think that no matter how you play these games, a narrative WILL be crafted. That is not variable.

Juul points out that interactive fiction, like text adventure games and I assume the analog version of Choose Your Own Adventure style books, don't really have variable outcomes...there are numerous outcomes but they are finite and unchanging if I always make the same choices every time I play/read. If I play a story game, but my character dies at the end, it still may be a satisfying narrative - a heroic tale of loss, or a tragedy in the classical sense, for example. Is it quantifiable? I suppose you could make an argument about whether the narrative is 'good' or not, but Juul again suggests that subjective appeals to art (making Pac Man move in a beautiful way) are not quantifiable. Even if we give story games this, that a good story is a better outcome of play than a bad story, how do we judge if the narrative is good or not? It goes back to Element 2, that the outcome must be beyond discussion for the activity to be a game rather than just a form of play.

So one of three conclusions are possible: a) Juul's definition of a game is incomplete [the null hypothesis],  b) story games fail the game test because of a paradox of not being able to uphold both Element 2 and Element 3 at the same time, or c) I'm missing some way to put valorization of outcomes on a cline rather than a binary which would allow story games to exist somewhere in the middle.

I think story games, like traditional RPGs, would likely fall in the borderline zone if they do fail the test. Just like Juul places The Sims/Sim City in the borderline, story games DO have definite rule systems, require player effort, have players attached to outcomes, and have no or negotiable real-world consequences. And they either conform to Element 2 or 3 as well. I'm liking this line of thought because it not only explains the difference between role playing and other games in general, but also easily shows how traditional and story RPGs are different.

If anyone can debunk me, though, I'll be happy to review this again. Next up, I plan to cover the player in relation to the game with Elements 4 and 5. Then I'll wrap this up with negotiable consequences and some final thoughts in the last post.


  1. There is only one quantifiable trait common to all games: they're interactive, i.e. you can play them. For every other criterion that academic folks try to attach to the concept (and I have long experience in this area), there is at least one game that disproves it.

    Now, once we leave the definitions of WHAT a game is behind, we can start to have much more fruitful discussions about WHY people play games and what they like/don't like about them, i.e. game aesthetics. That is a much more interesting subject, IMO.

    It's high time I did a revamp of my Game Aesthetics post from my old blog, for my new one. But if you're interested in seeing where I started along that thought process, here is the preliminary work I did on quantifying aesthetic dyads...

    It's rough, and I've updated it quite a bit over the last 7 years, but the basic concepts are there.

    1. Cool ideas. Thanks for the link. Have you put a lot of work into revising? You suggest going to single label -3 to +3 ratings, but I wonder if that will also prejudice opinions of something is labeled as "Linearity -2" (even though less linearity is often a positive).

  2. Yeah, I noticed some of the negative implications in the early system a while back. So along with changing certain terminology (like railroading into linearity, the latter of which is a much more accurate description, anyway) I now use a scale weighted +3 to 0 to +3, so it would be +1-3 Linearity or +1-3 Agency, or ♎ (the Greek symbol for balance) to represent an equal influence from both.

    For my game aesthetics class, I use the following dyads:

    Theme vs. Mechanics
    Simplicity vs. Complexity
    Determinacy vs. Indeterminacy
    Agency vs. Linearity
    Interaction vs. Isolation
    Player Skill vs. Mechanical Skill

    After I finish my next post on Middle Schoolers experience with Old School ethos (and finish the yearbook, and my syllabus for my college course), I'll do an update on the whole shebang, including what I've been teaching at the secondary and post-secondary level. You'll find it at