Video Game Deconstruction Techniques?

An email got me to thinkin…

There isn’t much good deconstruction of games out there. That is, someone who really knows how to make games trying to break down a game in a way that’s useful for themselves and for those looking to make better games. I’m pretty green on the subject, but I’m always interested when I come across a new way of looking at games that might be useful. What are your angles of attack for deconstruction? Some stuff I’ve seen that I think has merit:

Granularity (heard it first from Will W. – not really sure where it came from originally) Breaking the game down by various recurring time periods (heard this idea or a variation called loops as well.) What does the player do, what are they thinking about, and what are their goals at various increments – every 1 second, every 5 seconds, 10, 30, minute, 5 minutes, 20 minutes, hour, and so on. One example is from an FPS like Half-Life (via Matthew’s original IGDA talk:)

Fire the rocket launcher
|-|

Predictive aim the rocket launcher at helicopter
|—|

Take cover between shots as helicopter circles
|——–|

Destroy the helicopter to cross the bridge
|————————|

Find control room
|—————————————–|

Deactivate air defensives so friendly transport can land
|————————————————————–|

Escape the compound
|—————————————————————————–|

The Valve folks actually mentioned having a process like this in their Game Developer postmortem…they said they’d played through the game over and over again to make sure that the player would never go more than a minute without finding some puzzle, enemy, or item. Anyhow, I’ve found this useful for taking a game that’s mostly completed and looking for the area that could most benefit from extra love. If a game sucks at the lowest granule (bad feel, usability issues etc…) that’s really easy to see, and so is suckage at the largest granule (poorly written story, long-term balance issues.) But in the middle there are some great opportunities to pull stuff out and examine it, find out what’s going right and wrong and maybe change it for the better.

Reward Schedules. This is a sort of interesting idea, and one that gets applied a LOT when designing MMOs and other mass-market games. I feel a bit weird about it, like dirty or something, but it can’t be denied that when you understand the way that people respond to a given stimulus you can manipulate their perceptions and actions. Gamasutra article that goes into various reward schedules, if you haven’t read it: http://www.gamasutra.com/features/20010427/hopson_01.htm. I’ve found this useful for examining why people do what they do at particular moments in a game, especially with respect to when they stop playing entirely.

Flow Breakers. Frustration or boredom, yeah? I think it’s super interesting to look at a game with an eye for flow. The only way you can measure flow is by noticing when it’s broken. By definition, if you’re thinking to yourself, about yourself, or about anything, you’re not in the flow state. To examine a game for flow you have to let yourself zone out and try to write down instances when you start thinking things about yourself and how you’re doing, snapping back into self consciousness. In as much as the goal of lots of games is to get you into the flow state and keep you there, this can be an interesting prism through which to examine a game. I’ve found this useful as an experiential sanity check. As in, it’s a catchall for anything that’s detrimental to the experience. It could be the feel of the controller, bad level design, a weird art asset, a physics or motion glitch, or anything else that just seems a bit off. Anything which detracts from the overall experience of playing the game tends to snap you out of flow. Lots of stuff that isn’t a “show stopper” bug shows up under this kind of scrutiny. I haven’t tried it on a real project, but it seems like it might be interesting to rate bugs in severity not only by whether they break the game, but by whether they break flow experience (the flow.)

Emotional/Visceral Reaction. Genres are a pet peeve of mine. Like, Soldat is more closely related to Counterstrike or Tribes than it is to Worms. I feel the same way about playing Soldat as I do about playing Counterstrike. I have the same kinds of memories. They occupy a similar space to me experientially. But no one would ever classify them in the same genre. Why is that? It seems like there’s a huge opportunity here to look at the types of experiences a game gives you and try to relate them not only to other games but to other stuff in general. I think this is why I’m so struck by the idea of starting a design with an emotion or a experience – it seems like the only way a game can actually come full circle from idea to reality. In the end, that’s what we’re building, no? Experience?

If anyone reads this, I’m interested: how do you deconstruct a game? Does it help you make better games? How?

Catmilk.

7 Responses to “Video Game Deconstruction Techniques?”

  1. October 18th, 2007 | 12:09 pm

    I find it useful to deconstruct games by the cognitive skills that are brought to bear upon it. How much memory is involved? How many ways is your attention divided? How do you learn implicit memory movements while consciously keeping a nested goal structure in mind? Do you have to find patterns? Can things be represented symbolically?
    Admittedly, this is low-level, and doesn’t explicitly deal with deep emotional experiences, but cognitive workload is a great way to study flow. Comparing skill level to task demands while taking learning via repetition and reinforcement into account can yield insights into pacing difficulty properly, and training in general.

  2. October 19th, 2007 | 9:37 am

    That’s a big Swinkie,

    I find that your last example combined with the “granularity” concept is how I deconstruct most of my stuff. I think I’ve had some pretty successful levels that I put together based off of Valve’s concept of “vistas”. When Half-Life came out, I read some kind of design doc that turned me onto the idea. It takes a lot of time to implement, though.

    Basically, the idea is that the player should be rewarded by some kind of “scene” that shows them they are making progress. Each “scene” consists of some kind of landmark (such as the helicopter or the dam in Half-Life) that is followed with an experience that occurs with that landmark. Therefore, there’s not just one long memorable experience of playing the game where it all gets lost. It’s breadcrumbs to a series of events. Whereas the player’s immediate motivation might be to find a way out of the command bunker, they’ll know when they’ve found the way out because of the fight with the helicopter. That “experience” signals the end of their search.

    I find that using this method, I can tease the player into seeing areas where they know “something” is going to happen and then leading them past the area only to return for some kind of event. This way, you’re using the granularity concept to make sure that the player is always feeling certain things at certain times to affirm that they’re doing the right thing. It also makes the accomplishment factor higher because the player is proud of getting to that point. These vistas really make the individual experiences come together. It’s like an interactive progress meter for the player.

    That’s my 2 cents, anyways.

  3. October 19th, 2007 | 9:44 am

    btw. Portal = for the GOOOOOOOOOAAAAAAAAAAAALLLL!! BRAAAAZIIIIL!!! BRAZZZIIIIILLLLL!!

  4. October 19th, 2007 | 4:53 pm

    My slides with that HL2 example are available from http://www.igda-phx.org/resources.php?id=5 (it’s an old IGDA Phoenix talk). Credits plz :P

  5. October 19th, 2007 | 9:35 pm

    Oh snap, didn’t know that was on the tubes. Fixed!

  6. October 21st, 2007 | 9:40 am

    @Marc: Yeah, that’s a great approach. Chris Crawford on Game Design has my favorite breakdown of cognitive skills (with examples.) As he would say, it appeals to my simple mind :). I would be super interested in further reading on the subject, though. Is there a good starting point for understanding all the different cognitive skills at a deeper level, or have then been categorized in some smart way by someone smart?

  7. October 21st, 2007 | 12:47 pm

    Deconstrucción y análisis de videojuegos

    El autor de este artículo reflex´ona sobre los métodos que hay para analizar un juego y sacar sus puntos fuertes y débiles en cuanto a usabilidad, capacidad de mantener el interés, jugabilidad, premios, etc.

Leave a reply