The Burden of Choice
Games are interactive systems. The fidelity of a player's choices determines the nuance with which he can affect and express himself within that system. This goes broad as well as deep. More verbs = more potential choices. More 'adverbs' means greater control over the specifics of any given interaction (e.g. maim vs. kill, flirt vs. marry).
While you could argue that more of either of these axes would make for a richer gameplay experience, I think there are ramifications to consider when constructing a possibility space and the verb set the player can use to poke, prod and explore that space.
Feedback and Consistency
Can the player form a model of the system in his head and use that to anticipate the consequences of his actions?
This is a fairly prevalent point in Norman's The Design of Everyday Things (1). If the player doesn't have a clear idea of how the game world's components interact, he can't make an informed choice when acting on that world. For all the design work that might have gone into building the system, if that model is too opaque, the player might as well be using brute force trial and error to achieve his goals.
Harvey Smith addressed this at his GDC talk in 2002 on Systemic Level Design (2), and it was a real eye opener when I was getting my start as a designer. The idea is that you build a system with components that are granular enough to interact in interesting emergent ways and then maintain a strict consistency with the way these things behave. The more you frame your environments and scenarios around these consistent building blocks, the more a player can observe a scenario, form a plan and act on it, either getting the thrill of seeing his plan come to fruition or learning from his mistakes in a way that doesn't feel like the system cheated him.
Does the system acknowledge when the player is making a choice that impacts the state of the game world?
Games have differing levels of agency, the idea that the game is acknowledging that the player is acting within its world and is having some sort of impact. At the low level, this can be as simple as particle effects erupting from a surface as the player sprays it with gunfire. At higher levels, the player could impress a character in a dialogue exchange, resulting in an entire mission tree appearing at a later point in the game.
The issue that arises is how the player knows that his actions will have repercussions and at what agency level of feedback. Take GTA IV, for instance. The player can obliterate literally hundreds of people with no lasting effect on the game world. The game says "hey, you wanted to run over a guy, watch him crumple and bounce off of the hood of your car". During certain scripted sequences however, a large prompt appears on the screen giving the player the option to kill or spare specific individuals. In this case the game is trying to let you know your choice here will likely impact the narrative or mission structure of the game at a later date.
The thing to note here is that the game asks the player to step outside of system space to make the 'important' decision. The player doesn't really get to make the decision with his native tools so the whole arrangement feels artificial. The game implies that what the player can do with his standard interaction set doesn't really matter in the grand scheme, as major events will play out the same way regardless.
High impact choices like these are more sparse in narrative heavy games like the recent GTA games. This is because content is expensive, and generating branches of content that nobody will see generally gets a thumbs down from management types. More player driven games like Civilization or (yet again) Dwarf Fortress have a much more granular choice structure. Almost everything the player can do in these games has lasting implications that will sculpt the emerging narrative and interaction space down the line. Crawford touches on this a bit when he talks about Process Intensity (3). By leveraging the ability for the machine to generate content, you can often allow the player to explore a wider range of options as situations and locations you couldn't anticipate or wouldn't have time to build can be created on the fly by the computer. This is a bit hand-wavy, as there are many more details to building good procedural content, but it'll do for my purposes for now.
It's also worth mentioning that supporting inaction as a player option when he's using his native verb set is tricky. You're basically asking the player to walk away in light of instructions possibly to the contrary. An example would be person X telling you to kill person Y if you want his help. The player would probably assume this was the only valid option when the alternative is probably something like walking through a door or waiting around for a certain number of seconds. In this case the UI pop-up is alluring if you can't think of a clever way to contextualize the scene or provide some dialogue that doesn't seem too contrived.
Reflection
When the player is presented with the consequences of his choices, does he recognize the causal relationship?
Reflection is akin to feedback, but it's more of a retrospective thing. It's the ability for a player to observe the changes he's affected in the game world and recognize that he had an affect, as well as possibly gauge what might have happened otherwise. For these correspondences to be apparent to the player, the chain of events between cause and effect needs to be well presented. This means the event sequences should be fairly logical, easy to parse and that the time between cause and effect fairly short. I haven't gotten to play it yet, but I've heard The Witcher does a good job of showing the player the trail of events from A to B whenever he's confronted with the repercussions of his choices from earlier in the game.
If choice effects aren't made apparent, the player might as well be uncovering a prescribed sequence of events. Branching paths with subtle outcomes could just feel like a magician's choice (4) where the plot is predetermined but the player is given the illusion of a narrative rudder.
There are a couple of other ways that a player's choices can be presented to him in the context of the path not taken.
One option is a short but deep game cycle. Emphasis in this case is placed on replaying the game to explore all of the possibilities. The brevity of the game is to help make the development more manageable as well as allow the player time to check out multiple scenarios. I've posted a couple of times on this in the past(5)(6), and I think it's a model that should be explored more. Thus far, the most notable examples of this kind of idea come from Daniel Benmergui(7) and Gregory Weir(8).
Another idea is to facilitate the comparison of worlds and characters between players. Spore is a decent example of this, although the differences between user choices are primarily cosmetic and have little to no impact on the systems of the game. The Sims would probably be closer to a show and tell system of people demonstrating the range of possibilities the game has to offer to one another. Another good example is Animal Crossing. More than merely seeing how your friend's world diverges from your own, you can explore it with him and experience the differences firsthand, even bringing back parts of it to use in your own game world.
Paralysis
Does the player have enough information to make a choice from the available options? Are there so many options that the player can't make a clear distinction between them?
There's nothing more intimidating than a blank sheet of paper (looking at my post frequency confirms this). If a player is thrown into the thick of important decision making with little context, or simply given too many options at once, he can freeze up. Each option can either seem equally valid or vary in such minute ways that the 'correct' path is unclear. There's a good TED talk by a guy named Barry Schwartz that addresses this very issue. (9)
There are a few ways to mitigate negative side effects when presenting players with a vast array of options. One is to introduce verbs and player tools piecemeal, allowing the player to get comfortable with a smaller set of choices available at any given moment before adding more. Once the player gets used to the broader palette of options, he can begin to look at scenarios in terms of the available tools and proceed from there.
Another is to rely on stereotypes. If you have a set of commonly accepted attributes that the player is likely to be familiar with, you have a jumping off point for a broader range of options. If you give the player a choice of archetypes to start a game with, for instance, you could provide the standard fare as a base level. Thieves are sneaky, warriors are tough, etc. As the game progresses, you could offer more nuanced sub-classes of these characters that could refine the set of verbs the player is using as he plays. Your initial 5 or 6 basic classes that rely on general assumptions could blossom into 15 or 20 without rattling the player if you pace it right. The Elder Scrolls games do this pretty well by offering different ways to start. You can dive into the pool of options straight away, or you can let them ease you into it by choosing pre-made packages with pretty clear explanations on what to expect, in addition to the standard fare fantasy classes most people recognize.
I think a decent illustration of overwhelming choice is the Oblivion character creation system. At first blush, all the sliders that control every aspect of the character's face can be overwhelming. If the player has a plan up front however, usually something like 'I want to make my own face', then these tools become useful as a means to an end. From a bottom-up approach, however, the player can tell little difference between a 56% and 57% forehead-sellion-nose ratio.
Wrap Up
I'd intended on adding a section about user interface an accessibility, but it ended up leaning more on the side of usability design and not just about choices. The main take away from that was that the UI is not only the window the player uses to look in on the world within the game, but it's also his hands and feet. What is the interface (controller input, UI) telling the player he can or can't do? What information is being presented and what is being hidden behind layers of menus? All of these feed into the player's internal model of the system and how he thinks he can interact with it and what options are valued above others.
And that's about it. Just food for thought when you're constructing choice systems and verb sets in your games, and some of the implications different decisions can have on the way your game is experienced and understood.
References
While you could argue that more of either of these axes would make for a richer gameplay experience, I think there are ramifications to consider when constructing a possibility space and the verb set the player can use to poke, prod and explore that space.
Feedback and Consistency
Can the player form a model of the system in his head and use that to anticipate the consequences of his actions?
This is a fairly prevalent point in Norman's The Design of Everyday Things (1). If the player doesn't have a clear idea of how the game world's components interact, he can't make an informed choice when acting on that world. For all the design work that might have gone into building the system, if that model is too opaque, the player might as well be using brute force trial and error to achieve his goals.
Harvey Smith addressed this at his GDC talk in 2002 on Systemic Level Design (2), and it was a real eye opener when I was getting my start as a designer. The idea is that you build a system with components that are granular enough to interact in interesting emergent ways and then maintain a strict consistency with the way these things behave. The more you frame your environments and scenarios around these consistent building blocks, the more a player can observe a scenario, form a plan and act on it, either getting the thrill of seeing his plan come to fruition or learning from his mistakes in a way that doesn't feel like the system cheated him.
Does the system acknowledge when the player is making a choice that impacts the state of the game world?
Games have differing levels of agency, the idea that the game is acknowledging that the player is acting within its world and is having some sort of impact. At the low level, this can be as simple as particle effects erupting from a surface as the player sprays it with gunfire. At higher levels, the player could impress a character in a dialogue exchange, resulting in an entire mission tree appearing at a later point in the game.
The issue that arises is how the player knows that his actions will have repercussions and at what agency level of feedback. Take GTA IV, for instance. The player can obliterate literally hundreds of people with no lasting effect on the game world. The game says "hey, you wanted to run over a guy, watch him crumple and bounce off of the hood of your car". During certain scripted sequences however, a large prompt appears on the screen giving the player the option to kill or spare specific individuals. In this case the game is trying to let you know your choice here will likely impact the narrative or mission structure of the game at a later date.
The thing to note here is that the game asks the player to step outside of system space to make the 'important' decision. The player doesn't really get to make the decision with his native tools so the whole arrangement feels artificial. The game implies that what the player can do with his standard interaction set doesn't really matter in the grand scheme, as major events will play out the same way regardless.
High impact choices like these are more sparse in narrative heavy games like the recent GTA games. This is because content is expensive, and generating branches of content that nobody will see generally gets a thumbs down from management types. More player driven games like Civilization or (yet again) Dwarf Fortress have a much more granular choice structure. Almost everything the player can do in these games has lasting implications that will sculpt the emerging narrative and interaction space down the line. Crawford touches on this a bit when he talks about Process Intensity (3). By leveraging the ability for the machine to generate content, you can often allow the player to explore a wider range of options as situations and locations you couldn't anticipate or wouldn't have time to build can be created on the fly by the computer. This is a bit hand-wavy, as there are many more details to building good procedural content, but it'll do for my purposes for now.
It's also worth mentioning that supporting inaction as a player option when he's using his native verb set is tricky. You're basically asking the player to walk away in light of instructions possibly to the contrary. An example would be person X telling you to kill person Y if you want his help. The player would probably assume this was the only valid option when the alternative is probably something like walking through a door or waiting around for a certain number of seconds. In this case the UI pop-up is alluring if you can't think of a clever way to contextualize the scene or provide some dialogue that doesn't seem too contrived.
Reflection
When the player is presented with the consequences of his choices, does he recognize the causal relationship?
Reflection is akin to feedback, but it's more of a retrospective thing. It's the ability for a player to observe the changes he's affected in the game world and recognize that he had an affect, as well as possibly gauge what might have happened otherwise. For these correspondences to be apparent to the player, the chain of events between cause and effect needs to be well presented. This means the event sequences should be fairly logical, easy to parse and that the time between cause and effect fairly short. I haven't gotten to play it yet, but I've heard The Witcher does a good job of showing the player the trail of events from A to B whenever he's confronted with the repercussions of his choices from earlier in the game.
If choice effects aren't made apparent, the player might as well be uncovering a prescribed sequence of events. Branching paths with subtle outcomes could just feel like a magician's choice (4) where the plot is predetermined but the player is given the illusion of a narrative rudder.
There are a couple of other ways that a player's choices can be presented to him in the context of the path not taken.
One option is a short but deep game cycle. Emphasis in this case is placed on replaying the game to explore all of the possibilities. The brevity of the game is to help make the development more manageable as well as allow the player time to check out multiple scenarios. I've posted a couple of times on this in the past(5)(6), and I think it's a model that should be explored more. Thus far, the most notable examples of this kind of idea come from Daniel Benmergui(7) and Gregory Weir(8).
Another idea is to facilitate the comparison of worlds and characters between players. Spore is a decent example of this, although the differences between user choices are primarily cosmetic and have little to no impact on the systems of the game. The Sims would probably be closer to a show and tell system of people demonstrating the range of possibilities the game has to offer to one another. Another good example is Animal Crossing. More than merely seeing how your friend's world diverges from your own, you can explore it with him and experience the differences firsthand, even bringing back parts of it to use in your own game world.
Paralysis
Does the player have enough information to make a choice from the available options? Are there so many options that the player can't make a clear distinction between them?
There's nothing more intimidating than a blank sheet of paper (looking at my post frequency confirms this). If a player is thrown into the thick of important decision making with little context, or simply given too many options at once, he can freeze up. Each option can either seem equally valid or vary in such minute ways that the 'correct' path is unclear. There's a good TED talk by a guy named Barry Schwartz that addresses this very issue. (9)
There are a few ways to mitigate negative side effects when presenting players with a vast array of options. One is to introduce verbs and player tools piecemeal, allowing the player to get comfortable with a smaller set of choices available at any given moment before adding more. Once the player gets used to the broader palette of options, he can begin to look at scenarios in terms of the available tools and proceed from there.
Another is to rely on stereotypes. If you have a set of commonly accepted attributes that the player is likely to be familiar with, you have a jumping off point for a broader range of options. If you give the player a choice of archetypes to start a game with, for instance, you could provide the standard fare as a base level. Thieves are sneaky, warriors are tough, etc. As the game progresses, you could offer more nuanced sub-classes of these characters that could refine the set of verbs the player is using as he plays. Your initial 5 or 6 basic classes that rely on general assumptions could blossom into 15 or 20 without rattling the player if you pace it right. The Elder Scrolls games do this pretty well by offering different ways to start. You can dive into the pool of options straight away, or you can let them ease you into it by choosing pre-made packages with pretty clear explanations on what to expect, in addition to the standard fare fantasy classes most people recognize.
I think a decent illustration of overwhelming choice is the Oblivion character creation system. At first blush, all the sliders that control every aspect of the character's face can be overwhelming. If the player has a plan up front however, usually something like 'I want to make my own face', then these tools become useful as a means to an end. From a bottom-up approach, however, the player can tell little difference between a 56% and 57% forehead-sellion-nose ratio.
Wrap Up
I'd intended on adding a section about user interface an accessibility, but it ended up leaning more on the side of usability design and not just about choices. The main take away from that was that the UI is not only the window the player uses to look in on the world within the game, but it's also his hands and feet. What is the interface (controller input, UI) telling the player he can or can't do? What information is being presented and what is being hidden behind layers of menus? All of these feed into the player's internal model of the system and how he thinks he can interact with it and what options are valued above others.
And that's about it. Just food for thought when you're constructing choice systems and verb sets in your games, and some of the implications different decisions can have on the way your game is experienced and understood.
References
- The Design of Everyday Things
- Systemic Level Design slides:
Harvey's site: http://www.witchboy.net/2002/04/
My own pilfered copy to spare you a trip to Fileplanet - http://www.erasmatazz.com/library/JCGD_Volume_1/Process_Intensity.html
- http://en.wikipedia.org/wiki/Card_force#Magician.27s_Choice
- http://www.kongregate.com/accounts/danielben
- http://ludusnovus.net/my-games/the-majesty-of-colors/
- Reflection: Why the Indigo Prophecy Demo Was Better Than the Full Game
- The Ferry: Short but Deep Webs of Causality
- http://www.ted.com/talks/barry_schwartz_on_the_paradox_of_choice.html
Labels: choice, game design
posted by Rich Wilson at 12:09 AM
2 Comments:
Terrific post! I'm not a designer myself, but as a student of games and gaming, I wonder how much import we should place on the mechanics of a clear "reflection" (that is, making it all too clear what the feedback is and why it took place) rather than surprising the player with particularly high quality or strong feedback. You say that if the reflection is not made clear enough, the player might as well be playing a prescribed series of events, but since everything in the design is inevitably scripted (unless you're handing over to the player an unusually large lexicon of imagery or tools to use, ala LittleBigPlanet or Scribblenauts), I'm not sure that this would necessarily *devalue* the experience.
I'm thinking, for example, of Maniac Mansion. The game had several endings and routes based on initial character selection and your responses to certain puzzles. What made the experience refreshing was not necessarily the clear 1:1 correlation of action to response but that a certain action COULD even result in a specific response, some significant enough to conclude the game in a pleasurable (albeit abrupt) manner.
As you note, GTA makes such connections all too apparent and they have little bearing on the gameplay and narrative. Oblivion provides an immense amount of choice but not much in the way of distinct feedback for those choices.
Maniac Mansion, on the other hand, was a very simple game that was filled to the brim with feedback both clear cut (stereotypical adventure game puzzles) and completely unanticipated. I think finding the right balance of meeting expectations and surpassing those expectations is what goes the furthest towards creating a strong sense of player efficacy without the player getting lost in "what if's" (so many variables that my particular playthrough doesn't matter) or "what now's" (so many options that I feel overburdened before I even start getting to the meat of the game).
Great point about the mental model of the game in the mind of a player. From my earliest years as a gamer I enjoyed all kinds of games except the quest genre: I just counldn't understand how am I supposed to figure out what to do - not in a given situation, but in general.
Of course, later I discovered a few great quests, where designers managed to create something like a "mental model", where puzzle solutions are actually logical for your character and the game mood, and you were tought that kind of thinking in the first easy puzzles. Unfortunately, most games in this genre are still about guessing and reading walkthroughs.
Post a Comment
<< Home