Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 22, 2014
arrowPress Releases
October 22, 2014
PR Newswire
View All
View All     Submit Event

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

Proportion in Narrative Design
by Altug Isigan on 09/15/10 09:21:00 am   Featured Blogs

The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.


The distribution of scenes among acts is an important issue in structuring stories. There needs to be a sense of proportion in order to keep the experience balanced for the reader/player. In this article I first have a look at how it is usually done in feature films. I present some bad examples to make it clear what disproportion takes away from the audience. Later on I have a look at how games use scene proportion and what can be done to improve game designs by making use of this concept.


Three Acts, Four Quarters

Structuring a story is not only a matter of creating plot and characters, or writing dialogue, but also a matter of proportion. Traditional writing asks us to keep a balance between the amount of time (or scenes) we allot to the beginning, middle and end of a story. Typically, the distribution would look like this:

Three Act Structure

In most cases we would come accross four quarters; one allotted to the beginning act, one to the final act, and the remaining two to the middle act. Such scene distribution can be seen in a variety of comics, too. Just think of Dylan Dog, Ken Parker, Lucky Luke or Asterix.


Out of Proportion

Theoretically one could tell from an analyses of scene distribution whether a story maintains suspense or not. Consider the “stories” below:

bad proportion of acts

This first example takes too long to create the conflict, hence the beginning is out of proportion. The audience will wait for something intriguing to show up, but since this will take too long, they will get bored early on and lose faith that anything interesting will happen: that’s where they would zap to another channel. The second handicap of this example is that it has a too narrow middle, the meat of the story is not enough to satisfy the audience that wants to enjoy how things develope. And just as we think that things start getting interesting, it ends abruptly. That’s the third handicap. It’s a simple fact: bad proportioning brings the spectator’s expectation out of balance and ruins the experience.

bad proportion of acts

This second example looks good for the first two acts, but the denouement (the after-word) that follows the climax is way too long; it asks us to stay there for things that do not matter anymore since the conflict has been resolved and there’s nothing more to achieve. More than that, the spectator starts to wonder whether the story is now finished or not. Being busy with such questions, the joy that came from the great finish will fade away instead of remaining as something great to tell to friends if the end would have been kept tight.


Proportioning in Games

How does proportion work for games? An analyses of many games would reveal that their beginning and end is pretty narrow, compared to the huge middle that features the bulk of the action. Have a look at this:

proportion of acts in games

Is it a bad thing to have such a huge middle act? Actually no: the middle is a very valuable part, because it “hangs” there with the support of beginning and end. Since its joints with beginning and end are (ideally) of a logical nature, the middle is in possess of telos, that is, it is headed towards a goal (the solution) and driven forwards by the strong current of a source (the conflict). Hence, as long as it is justified and exposed in the beginning, and flowing in unity towards an (anticipated) end, theoretically, the middle could be expanded endlessly.

While games’ ability to blow up the middle looks in contradiction to the structure of the rather equally proportioned commercial feature film, we don’t need to go very far to find something similar in the media world: In terms of broadcasting formats, we can compare games’  elasticity in regard to the middle, to that of the TV serial.

The plotlines in TV series are planned to last over longer periods. In the pilot program, an initial problem is set up for every character, their interests are put up against each other. After this necessary introduction, the struggle among characters is being laid out over several chapters, turning into many directions (quests, subplots etc) along the process. Only on the final day of the series, all plotlines are brought to an end. So, the bulk of time is alotted to the middle.

Still we must notice an important difference between the game’s structure as a whole (which resembles the overall structure of TV series) and the structure of the single levels that make up that whole.  A single level will most likely be structured like a feature film, that is, it will have its own three act structure divided into quarters.

So basically we could claim that a lot of games resemble in their overall structure the TV serial, whereas the proportion in the structure of particular levels is closer to that of feature films. Hence, we have this:

the proportion in overall game structure versus the proportion in particular game levels


What Can We Learn From This?

We can identify at least two types of possible mistakes in the proportioning of acts and levels in games:

1. The first type of mistake would be in regard to the overall game structure. If the design is too late in establishing the conflict (a too long beginning) or if it continues too long after the climax/resolution (a too long end), we’d either lose the audience before they even arrive at the middle, or we’d ruin their enjoyment of the climax.

2. The second type would be in regard to the structure of particular levels. Regardless of where the level is positioned within the overall structure of the game’s narrative, if the particular level takes too long to set up its problem and goal, or if it allows the tension to fall too much after reaching the solution, we would have presented to our players a rather less enjoyable experience. A few such levels in a row could see them quit.

Finally it must be said that a too long middle can break the faith of a player, too. If the player feels that it will take too long to connect the beginning to an end, she might just give up playing. This requires designers to make sure that the middle and the levels that it is constituded of always drive the story forward, and that they don’t feel like we’ve got stuck in a unnecessarily long (row of) side quest(s).



In this article I tried to share some insights on narrative design in regard to the proportioning of acts/levels. I claimed that most games’ overall structure resembles that of TV series, while particular levels seem to be closer to feature film structure in their proportioning. While it is evident that these claims need to be supported with more research in order to be verified, I still hope that the perspective I put forward here will help narrative designers in structuring their stories in a more compelling way.

Related Jobs

Petroglyph Games
Petroglyph Games — Las Vegas, Nevada, United States

Illustrator / Concept Artist
DoubleDown Interactive
DoubleDown Interactive — Seattle, Washington, United States

Game Designer
Zindagi Games
Zindagi Games — Camarillo, California, United States

MOBILE Art Director
Treyarch / Activision
Treyarch / Activision — Santa Monica, California, United States

Senior UI Artist (temporary) Treyarch


JB Vorderkunz
profile image
Altug - a good piece as always. This post raised one question for me:

what constitutes a 'scene' in a game? In other media a scene is of course a subsequence within a story, the events of which are connected by location, time, or some other factor. This doesn't map 1:1 to games, at least to me. I think that the problem lies in the contiguity of games - levels are more akin to chapters than scenes, yet internally there are often few natural boundaries separating the level into separate 'scenes'. Clearly some designers plan this way, including several 'breaks' within the flow of their levels in order to introduce 'scenes'; but games are obviously not consistent in their pacing and proportioning of scene elements. A corollary question - is there a VG equivalent of the 'shot'?

Jacob Pederson
profile image
In addition to this excellent point, even if games did have clearly delineated scenes, what would stop the player from just stopping in the middle of one and ruining the dramatic pacing of their next play session?

Altug Isigan
profile image
My answer to this is, necessity and articulation.

A good level simply doesn't let you do that. When you go through a level in Diablo, ideally, the line-of-sight of the next monster would slightly expand into the location of your latest kill. Hence, the moment you collect the reward for your last kill, you would spot new monsters already coming at you. This forces continuity on you. This type of enemy-AI-based event articulation keeps you "on track". Still, the level design would present to you some non-articulated areas in which you can pause, examine, re-arrange your inventory etc. But once you tap into a new chain of articulated events, you'd find yourself unable to drop out. No need to say that this cycle of articulated and non-articulated level areas establishes rhythm and controls pace.

Altug Isigan
profile image
For instance Good of War does such articulation partly through anticipatory camera movement. Consider how at certain points, the shot scale would enlarge, the cam would move forwards to the next event (thereby pointing into a direction), dragging you behind. Hence there is constant secondary motion that puts you in context, and keeps you on track. It's very similar to the articulation of events in Diablo, only that it is guided by camera AI rather than by enemy AI. The result is continuity in action, and it is structured and designed in a way that doesn't let you drop out easily.

Altug Isigan
profile image
By the way, many narratologists asked a similar question in regard to literay texts: What stops a character from just stopping in the middle of a scene, thereby causing the action, and maybe the whole story, coming to an end?

Again the answer here is necessity. The events are constructed in such a way that the character has pretty much no other option than to do what he has chosen to do. As readers, necessity convinces us that the way things happened are acceptable and understandable. The characters choices are justified. More than that, it feels like it is fate, and not the writer, that is the force behind the development of events. This is necessary to create verisimilitude (believability).

The French narratologist Bremond studied many novels and tales in order to locate the decision-nodes in them. He was interested to find out why the characters ultimately always decided to do what's good for the continuity of the story. Theoretically, they could simply have decided to do something different (for instance to simply not to follow up their interest into the plot any longer). In theory, every decision node is a moment where we risk losing the story, because the character could have chosen to do what brings the story to an end. Hence he called decision-nodes, the "areas of risk" in a story.

However, in the context of novels, we speak of a potential risk here. In the end, the writer makes sure that the character chooses to do the right thing for the story, and that this feels plausible to the reader.

Games on the other hand differ from novels in that the risk is no longer a potential one. It has become a real risk, because ultimately, the author can no longer make sure that the character (now, the player) choses what is good for the story. But the above methods I mentioned, show us that the game designer is not completely defenseless against this type of risk.

Altug Isigan
profile image
Hi Bronaugh!

Thanks a lot for reading!

You pose some nice questions - as always :))

I had to think about it for a while, but now I'd say this is a matter of scope of the 'dramatic' material. I'd like to emphasize that I use the word 'scene' first and foremost in the dramatic sense, and not only in the audio-visual sense (as a purposeful compilation of shots). In that context, I would define a game as a hierarchical order of problem-solution arcs. You could break this hierarchy of problem-solution arcs down into acts, sequences, scenes etc.

Take for instance the Trondheim/Norway level in MOH:AA. I'd label it a sequence rather than a scene. The goal of the sequence is to destroy the overall U-Boot project of the Germans; this covers more than one task. Hence, the designers break it down into several scenes: (1) Infiltration of the base; (2) Stealing the project documentation; (3) Reaching the U-Boot; (4) Destroying the U-Boot; (5) Reach the train and escape the place. Each of the scenes could be broken down into smaller tasks: Get to the end of the corridor, eliminate that guy that waits at the corner etc. I would refrain from speaking of these smaller tasks as shots since I believe that the shot is not really a helpful concept in analytical terms.

One reason why I think so is that most games are visually presented to us as one uncut long shot. In most VGs, there isn't as much shot variation as can be observed in cinema. Hence, the shot as raw material to construct a scene is not used in VGs as it is the case in most feature films. Rarely ever will you see a cut from one shot to another when you rush through a level in MOH:AA or get to race in Need for Speed. In other words, games almost never make use of tertiary motion; they mostly stick with secondary motion (camera movement) or primary motion (object movement). Hence, 'shot' is not really the best of analytical categories to break down dramatic events in VGs.

However, that doesn't mean that there is no punctuation. VGs need to make use of it to give those long uncut shots that they are based on a feel of rhythm and pace. In other words, you can have one very long uncut shot, yet change in factors like angle, scale, lighting, color, setting, or increase and decrease in the pace of the camera movement can be used to suggest transition to a new scene, although it isn't done with a cut. Now this is the point where I believe we can learn a lot from architecture and interior design since we basically deal here with issues like circulation and spatial transition. Architecture can provide us with many methods in audio-visual (and yes, tactile) transition during one long uncut movement (such as walking through a park or building).

Level designer would maybe like to speak of such scene changes as transitions from one AI territory to another; that is, as you move through the level in one long uncut shot, one event is triggered after another, and in total these events make up the dramatic unit we call scene, and those scenes form the whole sequence of events that the level is.

Uhm, yes :)

Jacob Pederson
profile image
You touch on something here that has always bugged me. Even though editing in the film sense is very difficult to pull off in real time, there are some things you can do. The God of War series is very good at sneaking in cuts to other angels. Even though the player has to give up camera control to achieve this, I think the trade off is worth it in some situations. Also, I think you can get away with an actual edit at certain points. Traveling seems the best place to do this, especially back-tracking. However, flashbacks and inter-cut story-lines are becoming more feasible with today's tech.

Altug Isigan
profile image
I wrote about this in a previous post: "How to build tertiary motion into gameplay". I can't say it's one of my best articles, yet at the core it could insipre a few new thoughts. You might like to check it out. Thanks a lot for reading and sharing your thoughts, they're highly appreciated, Jacob!

Altug Isigan
profile image
By the way, I've ignored another very important method of transition between scenes: the use of sound. A different set of sound elements may foster a feel of change in location. And switching between "loud" and "silent" areas can be used for punctuation.

Jake Akemann
profile image
Interesting article, Altug.

I really like your comparison of video game levels/quests as episodes of TV shows. There's not a lot of story-based entertainment that can keep an audience's attention for 20-30 hours, but both television and video games seem to do it nicely.

And I definitely agree with you on the importance of a long middle act in video games. We can often gauge how long a video game will be based on the first act/"tutorial" section in the beginning. When games break that mold and have a very long first act compared to the rest of the game, the player can feel a bit cheated when they are confronted with an unexpected climax. If a game just drops the player in a heated conflict with no "real world" established, they can feel removed and uninvolved with the game. This is, of course, just from my experience.

Altug Isigan
profile image
Hi Jake!

Thanks for reading, I'm glad you found this article useful. The example you give fits in very well, it's exactly one of the problems I tried to address in this article. Thanks again for this nice comment...

*And yes, after giving it a though, your statement about how games and TV serials are capable of keeping an audience' attention for 20-30 hours is very important indeed. That's a good point to support the familiarity between these format structures.

Piotr Zygadlo
profile image

Nice and interesting article (as are the most of yours).

I have one question however. It was stated few times before, that it is important to "not allow" player to stop, to keep him/her engaged with the story pace. Great. But what with Adventure Games (like recently more and more popular BigFishGames Puzzle Adventure games)? Do you have any idea, how to control pacing in such game? In my opinion it is impossible to avoid such situation in non-action (logic, adventure etc.) games. But maybe someone there outsmarts me ;)

Altug Isigan
profile image
Hi Piotr!

Thanks for reading (and the nice comment). :)

First let me say sorry. I guess I've never played a Big Fish game. Yet I'll try to answer your question.

As long as the pause (or non-action, as you name name it here) possesses an element of motivation (in other words, as long as the player is still engaged into the game despite not carrying out an action), the pause can be regarded as an active element and will be functional in pacing and creation of rhythm. For example in Chess, I face a game state (non-action), and make my calculations for my next move (non-action). Finally I make the move (action) and my turn (or sequence of events) is over. You could look at this as a problem-solution arc that shares structural similarities with a scene or act in drama. The game state I face at the beginning is the problem (introduction), the calculations in my head fill the middle, and the move I finally carry out is the resolution (ending). Turn by turn (or scene by scene) the game session reaches it's ultimate finish: One of the players calls "Schach Matt!"

It is common that some games put limits on non-active yet motivated gameplay sections. A clock would tick for example to indicate that the pause doesn't last forever. But even in games without such time constraints, it is ultimately the presence of motivation that keeps a non-active (pause) section as a lively and "active" experience. Say in Civilization I am faced with a decision. The only real action is to click on what I've decided. But most of the action takes place in my head, when I calculate the possible outcomes of the course of action that I'm going to take. So, the pause here is never empty, it is always motivated.

In creating rhythm, what really matters is that both action and pause are motivated sequences, that is, sequences in which the player is engaged with the game, whether he acts or not. This makes it possible that I feel rhythm, because I go through a constant change of high-paced and low-paced sequences.

So, have you been outsmarted now?? :D:D