Mental Modeling: Character Attitudes and Quest Intent
One thing that frustrates me when playing games is the inability to have the character feel the same way as I do about other characters or situations in the game. Some games use conversation trees to express yourself and how you want your character to respond to situations. The usual internal model of the character that results from this usually a slider at best, representing some flavor of good/evil continuum.
One of the first things too look at is character interactions. Games usually don't store the player character's attitudes toward characters, though some games store characters' attitudes towards the player. The result of this is you get the full spectrum of conversation options when speaking to someone, allowing you to be incredibly fickle with your approach to them, praising them in one breath while cursing them in the next.
What I'd like to see is some sort of character list you can access and assign what you think (or what you think the player character thinks) about important characters. This could even extend to side characters if you prioritized them and made the interface easily navigable. I'd envision this as being a grid of portraits. When you highlight one you either get a like/dislike slider or something more involved, like a radial wheel of adjectives to select from. For instance, a overly friendly character could have options like: Bob is "a kiss-ass", "cheerful", "deluded", etc. Whenever the player has interactions with this guy, that opinion will inform the conversation choices. This sort of system could even obviate a tree system altogether, letting the player "front-load" all the conversation choices by choosing what the protagonist thinks. This potentially removes interaction though and could relegate the player to watching instead of participating.
One potential pitfall lies in the frequency in which players can alter their opinions, flip flopping between loving and hating someone. This could be reduced by giving the player "points" any time he interacts with, or hears news about the character in question. These can then be spend to slide the slider or move the adjective pointer a certain distance. A more obtrusive way could be popping up a question every time the character in question does or says something significant that asks what the player thinks about what just happened. "Bob stabbed your brother. What do you think? [he had it coming] [holy crap!] [that such Bob is such a kidder]"
All of these settings would start in some sort of neutral position, so a player that didn't want to deal with it could let it sit in the background and focus on the action elements of the game (unless the game was centered around relationships, in which case this would be a core mechanic instead of a cumbersome aside). I think this system may have the unfortunate burden of a pretty involved UI, but I can't think of a more subtle way to allow the player to accurately model the main character's outlook on the world. Clever UI design could mitigate the unpleasantness of the system, and a low frequency of opinion 'tuning' might make it less disruptive to immersion.
Another facet of modeling the thought processes of the character is to address intent with regards to in-game goals. One thing the above system could allow is a deceptive relationship with a character. You could absolutely despise someone, but become buddies with them to learn how to bring them down. In this case, you could accept quests from them with the sole intent of undermining them. Most games currently have succeed/fail scenarios for these sort of situations. Suppose you did take a quest from a character you didn't particularly like with the intention of botching it or somehow bending the situation to benefit you instead. The player would have to know that these are valid options in a quest through some non-systemic means. There would have to be some inner monologue saying "or I could just do things my way" or some sort of popup for the player to know that completing this quest in the non-standard way is a legitimate course of action that has its own reward path.
In my proposed system you'd have a journal to track quests like most other games, but there would be different success options (probably opened up depending on your attitude towards characters involved). So "retrieve the gem for Lord Evilface" would have the options: "Complete as is" and others like "Steal the gem for myself" or "Find a fake gem to give Evilface instead of risking my neck". This might seem heavy handed, but hopefully the options presented will seem like natural extensions of the character's wishes based off of the input from the character attitude matrix described above.
Again, I'm not certain if these are steps forwards or backwards in relationship to the more streamlined console era that's taking place. I think I'll mull over interface for a while and come up with a followup article to this exploring some ideas of ways to take this system and not make it a bloated spreadsheet management chore for the player.
One of the first things too look at is character interactions. Games usually don't store the player character's attitudes toward characters, though some games store characters' attitudes towards the player. The result of this is you get the full spectrum of conversation options when speaking to someone, allowing you to be incredibly fickle with your approach to them, praising them in one breath while cursing them in the next.
What I'd like to see is some sort of character list you can access and assign what you think (or what you think the player character thinks) about important characters. This could even extend to side characters if you prioritized them and made the interface easily navigable. I'd envision this as being a grid of portraits. When you highlight one you either get a like/dislike slider or something more involved, like a radial wheel of adjectives to select from. For instance, a overly friendly character could have options like: Bob is "a kiss-ass", "cheerful", "deluded", etc. Whenever the player has interactions with this guy, that opinion will inform the conversation choices. This sort of system could even obviate a tree system altogether, letting the player "front-load" all the conversation choices by choosing what the protagonist thinks. This potentially removes interaction though and could relegate the player to watching instead of participating.
One potential pitfall lies in the frequency in which players can alter their opinions, flip flopping between loving and hating someone. This could be reduced by giving the player "points" any time he interacts with, or hears news about the character in question. These can then be spend to slide the slider or move the adjective pointer a certain distance. A more obtrusive way could be popping up a question every time the character in question does or says something significant that asks what the player thinks about what just happened. "Bob stabbed your brother. What do you think? [he had it coming] [holy crap!] [that such Bob is such a kidder]"
All of these settings would start in some sort of neutral position, so a player that didn't want to deal with it could let it sit in the background and focus on the action elements of the game (unless the game was centered around relationships, in which case this would be a core mechanic instead of a cumbersome aside). I think this system may have the unfortunate burden of a pretty involved UI, but I can't think of a more subtle way to allow the player to accurately model the main character's outlook on the world. Clever UI design could mitigate the unpleasantness of the system, and a low frequency of opinion 'tuning' might make it less disruptive to immersion.
Another facet of modeling the thought processes of the character is to address intent with regards to in-game goals. One thing the above system could allow is a deceptive relationship with a character. You could absolutely despise someone, but become buddies with them to learn how to bring them down. In this case, you could accept quests from them with the sole intent of undermining them. Most games currently have succeed/fail scenarios for these sort of situations. Suppose you did take a quest from a character you didn't particularly like with the intention of botching it or somehow bending the situation to benefit you instead. The player would have to know that these are valid options in a quest through some non-systemic means. There would have to be some inner monologue saying "or I could just do things my way" or some sort of popup for the player to know that completing this quest in the non-standard way is a legitimate course of action that has its own reward path.
In my proposed system you'd have a journal to track quests like most other games, but there would be different success options (probably opened up depending on your attitude towards characters involved). So "retrieve the gem for Lord Evilface" would have the options: "Complete as is" and others like "Steal the gem for myself" or "Find a fake gem to give Evilface instead of risking my neck". This might seem heavy handed, but hopefully the options presented will seem like natural extensions of the character's wishes based off of the input from the character attitude matrix described above.
Again, I'm not certain if these are steps forwards or backwards in relationship to the more streamlined console era that's taking place. I think I'll mull over interface for a while and come up with a followup article to this exploring some ideas of ways to take this system and not make it a bloated spreadsheet management chore for the player.
Labels: game design
posted by Rich Wilson at 12:36 PM
2 Comments:
What game am I thinking of.. Planescape? Or maybe it's Kotor. Many dialogue options would be appended with a parenthetical description of your intent in making your response. So one set of responses might include "You can trust me," and well as "[LIE] You can trust me." It was clever, and now that I think of it it must have been Kotor, because I think the difference between the two above responses would simply be light side or dark side points.
I think an effective approach would be to maintain the player's ability to make a standardized set of responses in any given situation, and simply introduce a number of characters that persist in the game for a very long time while tracking the player's expressed attitude towards them. So as the player chooses the responses over hours of play that communicate to the NPC "this person is my good friend," their own behavior toward the player will tend to reflect this. And then after the player has invested in gaining that character's trust, and suddenly chooses the dialogue choice that is truly hurtful, that NPC would react with 'genuine' surprise, as it goes against the internal graph they've been building over the course of dozens of micro-interactions earlier in the game. It seems like attempting to directly query and record the player's intent might be less effective than presenting NPCs that convincingly reflect the impact your dialogue and action choices are having on a simulated person over time.
The system in KOTOR is something I kept in mind while thinking about this system. I liked the [LIE] tag to let the player know the character's intent.
This was also cool because it tied into a player's skills, so you could actually fail the lie and have a different conversation outcome.
Post a Comment
<< Home