Game design theory essay...
I did a lot of thinking about level design over the weekend. For most of it, I stuck to shmup levels - my thoughts centered around Machine City - but I kept coming back to level design "in general", which to me means applying the same principals to a variety of genres.
That's the problem with being a theorist. You always want to generalize.
Anyhow, one of the things that I wanted to clarify is play types. I'll also clarify player vision some more at some point, in another essay.
Most games play either fast or slow. Most of the time, a fast game will eat levels for breakfast, pound through acres of map. Most of the time, a slow game will use the same map and criss-cross over it a half-dozen times. There are exceptions, and this simplified level design philosophy is really only true of one-player and co-op games.
Either way, one of the things you need to do is to keep the player from getting bored of the pattern of gameplay. Some games do this well, some games bore the player to tears. This is more than just offering them play they don't much like: it's offering them too much of a given kind of play.
An example of each can be found in the game No One Lives Forever 2. If you haven't played this game, go buy it. It is my second favorite first-person game, period, and I haven't seen any game which matches it for flawless quality on every level. Save level design, which was occasionally quite good and occasionally quite bad.
Much of this game is slow play. You're playing a spy, after all. A good example of slow play being used well is the undersea base. You play through the whole base. It gives a sense of large size and scale, despite the fact that it is a rather small level. It goes on a little long, actually, but not too badly thanks to the ever-increasing variety of encounters.
Then, after a boss fight, you get to revisit the station, seeing it from a radically different viewpoint trying to escape it as it floods.
This is something similar to what Doom III tried to do: show you the base "before" and "after". "After" is inherently more interesting for various reasons, so it is natural to try to stretch it out a bit. The undersea base doesn't, but I wouldn't have minded a bit more level on the escape.
Anyhow, on the other side of the spectrum is India, the city. A large, open level in which you can't kill anyone. All you do is run back and forth on errands, dodging cops. On the third playthrough, you can blitz all the India city levels, but it's a protracted and rather painful experience the first time through.
Later on, you revisit India city. To do what? Run errands and not kill anyone. Wow, and this time the errands are even more inane, featuring running back and forth for water while being shot at. Across the same paths. Over and over. To save people who are too stupid to climb through the giant, open window two feet from them. There are a few moments in which HARM's doomed Indian soldiers attack you, but these are bizarre episodes which aren't really explained: why are they attacking you instead of trying to escape their demise? Although a nice attempt to keep the level from being "turn off the computer" bad, they break suspension.
On the surface, these two levels appear very similar. You run through a lot of corridors in each, can approach them both by stealth, have an intermission (boss fight in one, several levels in another), then work through the "ruined" version.
But a play type analysis shows a very different set of levels. Both levels initially play "slowly", although one is a "conquest"-style play and the other is a "foxhunt"-style play. One is voluntary slow play, the other is involuntary. As you go through the undersea base, you choose to go slowly, killing the threats and making that part of the base safe for you. In India, you are forced to go slow because you need to time your movements to keep you alive. I believe this kind of "foxhunt" play to be distinctly inferior, since it is nonadaptive and submits the player to the same challenges repeatedly.
On the last visit to these levels, things are distinctly different. The undersea base speeds up to a moderately fast game where India slows down even further. India remains "foxhunt" play, but the undersea base becomes "shockwave" play: you blast through each part of the level and leave nothing you ever bother to revisit.
My dislike of "hunted" play aside, using the same gameplay every time you hit a level, you might as well be playing the same level. Shake it up!
As a basic rule, the human brain only stays interested in something for about 20 minutes. Some people can go for vastly longer, but few go for shorter. This means that you need to change modes of play ("patterns") no less often than once every twenty minutes. You don't have to change it for long, but you need to change it. And in order to get maximum use out of a level design, you can get an entirely new experience out of a map by hitting it with a different style of play.
"Twenty minutes?" you say, "that's a long time to be sneaking around without a gunfight. It seems to me that most games interrupt with a new style of play every two minutes!" Yes, for two reasons:
1) Resetting the clock insures maximum attention.
2) Many players have an exceedingly slow first play-through. You can beat XIII in 130 minutes? Sure, but the first time it took you thirteen hours.
So level designers do - and should - err on the side of short attention spans.
To give you an idea of some of the basic kinds of play, here is an incomplete list. Feel free to suggest things I've missed:
Foxhunt: moderate to slow play speed
Explanation: Level is littered with invulnerable threats that have few successful approach vectors; player speed is therefore slowed to allow for higher relative vision.
In English: Effectively a puzzle level, foxhunt emphasizes careful watching and considered movements on pain of death. The slower the play style, the more careful you have to be not to exceed your 20-minute maximum play length.
Conquest: slow play speed
Explanation: Player earns ownership of level.
In English: The player destroys enemy presence in a level so that he owns it. Although this is, in itself, probably medium-speed play, the ownership of the level then encourages the player to carefully search every room. Depending on the player's level-affecting capabilities, they may slow the game even further by destroying parts of the level or by building new structures.
As with foxhunt, care must be used to insure that a player doesn't go more than 20 minutes without a play style change.
Hunted: fast play speed
Explanation: Player vision limits are enforced by forcing player movement. Play speed is therefore accelerated.
In English: Effectively a timed level. Rising water, a rampaging demon, and other active forces push the player to keep moving or lose the game. This play style can also result from moving levels. For example, conveyor belts.
Please note, "hunted" and "hunter" levels are functionally the same. Whether the player is forced to move along by a wall of fire or whether he has to constantly chase a monkey, the core idea is that there is a time-sensitive thing pushing the player to constantly move forward.
Shockwave: moderate to fast play speed
Explanation: Player's only goal is to get through the level
In English: This can be very similar to conquest, but the design is such that there is no way to slow down and appreciate your hard-earned territory. Perhaps there is a constant stream of soldiers, or perhaps there is simply nothing worth appreciating.
Often, "shockwave" is almost indistinguishable from "hunted". The difference is that when you are hunted, something is actively pushing you. Shockwave may send streams of soldiers, but your skill could be enough to hold them off forever. Shockwave may slam doors shut, but only after you use them - not as a possible loss condition.
---
These four types arise from basic elements of play, which I imagine I'll cover sometime. I don't think it's a full list, but each of those has a very different approach, both for play and design.
Monday, November 28, 2005
Wednesday, November 23, 2005
A Weekend Most Fowl
I'm having a Thanksgiving dinner tomorrow!
This is rather unusual for me. My usual Thanksgiving dinner consists of trying to find some place that's open, failing, and wishing I had remembered to shop for groceries.
Just thought I'd break the parade of pent-up pedagogy with a post proffering paltry poultry.
This is rather unusual for me. My usual Thanksgiving dinner consists of trying to find some place that's open, failing, and wishing I had remembered to shop for groceries.
Just thought I'd break the parade of pent-up pedagogy with a post proffering paltry poultry.
Tuesday, November 22, 2005
Gravitic Memetics
I thought up a huge theoretical framework for considering memetics as celestial bodies, with gravity and spin and light-speed and all that jazz. I even worked out a little bit of half-advanced waves/particles, but I decided that really wasn't going to help me any.
In the end, however, there were few things that really popped up as a useful:
Some things are big - black holes and stars with considerable "memetic gravity". For example, you might consider BoingBoing, health care, or your office as having a truly impressive gravity field (each person builds their own "star map"). Some things are more moderately sized. This site, for example, has noticeable gravity - perhaps comparable to one of Saturn's smaller moons. However, similar to a comet, few visitors get close enough to get slowed by the gravity, let alone pulled into an orbit.
Other things - the most common type of things - are really tiny. Dust, asteroids. Ads, for example. Their purpose is to tug on a user until his vector is headed in a direction more to your liking. They utilize high proximity and careful spin to maximize their gravitic pull. Actually HITTING a user dead-on is nearly impossible, so it's best to think in terms of gravity, rather than physical impacts.
Of course, celestial bodies both big and small don't pull in one direction. So if the user comes in from a direction you didn't expect, you can easily end up slingshotting them the wrong way. Fortunately, most users are orbiting in the plane of a larger celestial body when they encounter your little celestial bodies, so you know roughly their vector.
But if you guess wrong, you can do worse than have no effect: you can push them away. Throw them somewhere further from you. Like if you advertised Windows on SourceForge. Like if you put up any Republican-type posters anywhere in Seattle. These things do pass very close to the passerby - but the vector of the passerby is wrong, and the slingshot hurls them away from you.
Anyhow, the key to this kind of thinking is to think in terms of orbits. Your ads will be placed in orbit around the same things the users are orbiting, and your job is to pull them away. You need to have a very similar orbit, a lot of spin, and a carefully placed storm of ads.
You need to be delicate. Place it wrong and you'll end up flinging the audience deeper into the heart of the system, or out into the depths of space - certainly not at you.
Of course, without the conceptual framework identifying memetic orbits, orbital patterns, and measurement of the audience's relative distances, you can't really usefully use this. Plus, each such "orbital framework" is built in the mind of each user, so you're operating at one step removed...
But the idea that you can think of ads as tiny gravity wells placed in the orbital path of users might be helpful when you think of marketing.
In the end, however, there were few things that really popped up as a useful:
Some things are big - black holes and stars with considerable "memetic gravity". For example, you might consider BoingBoing, health care, or your office as having a truly impressive gravity field (each person builds their own "star map"). Some things are more moderately sized. This site, for example, has noticeable gravity - perhaps comparable to one of Saturn's smaller moons. However, similar to a comet, few visitors get close enough to get slowed by the gravity, let alone pulled into an orbit.
Other things - the most common type of things - are really tiny. Dust, asteroids. Ads, for example. Their purpose is to tug on a user until his vector is headed in a direction more to your liking. They utilize high proximity and careful spin to maximize their gravitic pull. Actually HITTING a user dead-on is nearly impossible, so it's best to think in terms of gravity, rather than physical impacts.
Of course, celestial bodies both big and small don't pull in one direction. So if the user comes in from a direction you didn't expect, you can easily end up slingshotting them the wrong way. Fortunately, most users are orbiting in the plane of a larger celestial body when they encounter your little celestial bodies, so you know roughly their vector.
But if you guess wrong, you can do worse than have no effect: you can push them away. Throw them somewhere further from you. Like if you advertised Windows on SourceForge. Like if you put up any Republican-type posters anywhere in Seattle. These things do pass very close to the passerby - but the vector of the passerby is wrong, and the slingshot hurls them away from you.
Anyhow, the key to this kind of thinking is to think in terms of orbits. Your ads will be placed in orbit around the same things the users are orbiting, and your job is to pull them away. You need to have a very similar orbit, a lot of spin, and a carefully placed storm of ads.
You need to be delicate. Place it wrong and you'll end up flinging the audience deeper into the heart of the system, or out into the depths of space - certainly not at you.
Of course, without the conceptual framework identifying memetic orbits, orbital patterns, and measurement of the audience's relative distances, you can't really usefully use this. Plus, each such "orbital framework" is built in the mind of each user, so you're operating at one step removed...
But the idea that you can think of ads as tiny gravity wells placed in the orbital path of users might be helpful when you think of marketing.
Chapter Four
I've posted Chapter Four of my tutorial on vector math. You can find the earlier chapters here.
The latest chapter has more math, but I've tried to go slow and clearly explain. Let me know what you think, if you think. :)
The latest chapter has more math, but I've tried to go slow and clearly explain. Let me know what you think, if you think. :)
Google Analytics
ROCKS!
It finally kicked in for me. It is awesome, even though it highlights the fact that I'm three inches tall.
It also shows me that I have the most hits over the weekend, even though my weekends contain no postings, and fewest hits at the beginning of the week, even though the beginning of the week is when I have the most postings.
Hm. Weekends are probably just really popular. Once I'm on my own, I'll start weighing in on weekends.
Oh, and most of my hits come from google searches, but apparently no particular Google searches, since there's no popular ones leading to me. Next in line are people from grand text auto, presumably having read Patrick's advertisement of my Movies review.
And MOST of my visitors use FireFox. Huzzah for you! On the other hand, NONE of my visitors use Linux. 95.84% of them use Windows. I guess that's to be expected: Linux isn't hugely game-friendly.
My god! This is a wealth of information. Most screen resolutions are 1024x768 or 1280x1024, with all others a very distant last place. And I got a small but significant number of visitors using 16-bit color instead of 32-bit color. Most of my audience is registered as English-American, has Java enabled, and has Flash. The majority have Cable/DSL, with a significant chunk still using dialup.
On the negative side, it doesn't seem to count feeds. Hm.... I bet if I embedded the javascript in the feed... but that doesn't work, does it? I'm not in the header. Oh well, it doesn't matter, blogspot doesn't allow "script" tags.
On the positive side, my average view time is over seven minutes! This is a brutally weighted length, however. The vast majority of my visits are over in less than ten seconds - sob! - but a significant chunk stayed more than ten minutes. Did they just forget, and leave the page open?
Anyhow, go sign up for Google Analytics. At the very least, it will let you feel the depth of your unpopularity!
EDIT: People found me through some very strange searches. "Smegging" is my favorite, for sheer classiness. But things like "Cringely Google", "Anti-stealth 2005", "anakin armpit hair clone wars", "i will survive sang by aliens" (ha, that one actually makes sense!), and "pictures of deer taking a crap".
Bizarre... what could these people have been looking for? Whatever it was, now I'm TWICE as likely to get pinged by people searching for pictures of deer taking craps.
EDIT the SECOND: My god, I'm the second result for 'anakin "armpit hair" "clone wars"'. This because one of my month-archives has a post talking about armpit hair, and another post about clone wars. I would have wondered about my sanity if they'd been in one post.
I don't think I'll do the Googling for "pictures of deer taking a crap"...
It finally kicked in for me. It is awesome, even though it highlights the fact that I'm three inches tall.
It also shows me that I have the most hits over the weekend, even though my weekends contain no postings, and fewest hits at the beginning of the week, even though the beginning of the week is when I have the most postings.
Hm. Weekends are probably just really popular. Once I'm on my own, I'll start weighing in on weekends.
Oh, and most of my hits come from google searches, but apparently no particular Google searches, since there's no popular ones leading to me. Next in line are people from grand text auto, presumably having read Patrick's advertisement of my Movies review.
And MOST of my visitors use FireFox. Huzzah for you! On the other hand, NONE of my visitors use Linux. 95.84% of them use Windows. I guess that's to be expected: Linux isn't hugely game-friendly.
My god! This is a wealth of information. Most screen resolutions are 1024x768 or 1280x1024, with all others a very distant last place. And I got a small but significant number of visitors using 16-bit color instead of 32-bit color. Most of my audience is registered as English-American, has Java enabled, and has Flash. The majority have Cable/DSL, with a significant chunk still using dialup.
On the negative side, it doesn't seem to count feeds. Hm.... I bet if I embedded the javascript in the feed... but that doesn't work, does it? I'm not in the header. Oh well, it doesn't matter, blogspot doesn't allow "script" tags.
On the positive side, my average view time is over seven minutes! This is a brutally weighted length, however. The vast majority of my visits are over in less than ten seconds - sob! - but a significant chunk stayed more than ten minutes. Did they just forget, and leave the page open?
Anyhow, go sign up for Google Analytics. At the very least, it will let you feel the depth of your unpopularity!
EDIT: People found me through some very strange searches. "Smegging" is my favorite, for sheer classiness. But things like "Cringely Google", "Anti-stealth 2005", "anakin armpit hair clone wars", "i will survive sang by aliens" (ha, that one actually makes sense!), and "pictures of deer taking a crap".
Bizarre... what could these people have been looking for? Whatever it was, now I'm TWICE as likely to get pinged by people searching for pictures of deer taking craps.
EDIT the SECOND: My god, I'm the second result for 'anakin "armpit hair" "clone wars"'. This because one of my month-archives has a post talking about armpit hair, and another post about clone wars. I would have wondered about my sanity if they'd been in one post.
I don't think I'll do the Googling for "pictures of deer taking a crap"...
Monday, November 21, 2005
Music!
Music geekery follows...
As I mentioned a long while back, I'm looking into music composition. My equipment is severely limited, but my studies are not.
One thing I've found in my studies is that I have picked up an unnatural ability to hear synthetic instruments. I suppose this probably came from my sonogram studies. They've started to really grate on my ear.
I don't mean "bweep baweep bweep" instruments. I don't even mean midi- or tracking-fed instrumentation. I mean I can hear filters, compressions, limiters, and the very best synthetic orchestras. I've recently deleted songs I've always had on my playlist, just because they've started grating on my ear.
Synthetic orchestration - especially simplistic versions like midi or tracking, are the most grating. I've turned myself into an elitist!
This is really bad news, since the only capabilities I have are midi. I don't own any physical instruments other than a (primitive) midi keyboard and similar computer software (it should be noted that I canot actually connect the keyboard to the computer...). I'm writing primitive songs, now, but I know that I will never be happy with anything I write, because it'll be in midi. This is especially bad because I'll eventually need music for Machine City and consecutive games.
On the other hand, I know what the problem is with these fake instruments. It's their attacks and sustains. They're unnaturally sharp and flat - not the notes, the envelopes. The computer program doesn't distinguish between a G placed in the middle of a rising sequence or at the beginning of a triumphant reveal. It hits both with the exact same authority.
The author can try to deal with this. It's possible, with great effort, to deal with some of it. You can put in volume changes, accents, and so forth. But it takes time, and there are no fine controls. If there were, it would take even more time!
That got me thinking. Why not make a program with "musicians" in it. You can assign a musician to an instrument, and teach them how you want them to play. You can carry these musicians from composition to composition, even sharing them over the internet. You can train them to play the horn with a certain level of attack and vibrato depending on what kind of sequence is being played at that moment.
Your controls would be largely limited to marking the music with emotional notation that guides the musicians. There's already denotation for sheet music, but this is not enough: we need more explicit terminology, such as "fiery".
If your pattern recognition was good enough, you could leave off with the labels and just instruct the musicians directly. They would work out what kinds of situations called for which kinds of performance, backed and instructed by you.
Or - here's an awesome idea - use the Nintendo Revolution controller to play as a conductor. Ha! That would make my YEAR.
I think there's already a program with this kind of capabilities out there. I've heard a couple of orchestral pieces which must be synthesized, but they're really good, instrumentally. They must simulate the actual instrument, which would probably do the trick nicely. They're still flawed, but they may actually be just badly mastered, mediocre-played real instruments. However, their sustains are still unnatural, so I doubt it...
Anyhow, just musing on the matter.
As I mentioned a long while back, I'm looking into music composition. My equipment is severely limited, but my studies are not.
One thing I've found in my studies is that I have picked up an unnatural ability to hear synthetic instruments. I suppose this probably came from my sonogram studies. They've started to really grate on my ear.
I don't mean "bweep baweep bweep" instruments. I don't even mean midi- or tracking-fed instrumentation. I mean I can hear filters, compressions, limiters, and the very best synthetic orchestras. I've recently deleted songs I've always had on my playlist, just because they've started grating on my ear.
Synthetic orchestration - especially simplistic versions like midi or tracking, are the most grating. I've turned myself into an elitist!
This is really bad news, since the only capabilities I have are midi. I don't own any physical instruments other than a (primitive) midi keyboard and similar computer software (it should be noted that I canot actually connect the keyboard to the computer...). I'm writing primitive songs, now, but I know that I will never be happy with anything I write, because it'll be in midi. This is especially bad because I'll eventually need music for Machine City and consecutive games.
On the other hand, I know what the problem is with these fake instruments. It's their attacks and sustains. They're unnaturally sharp and flat - not the notes, the envelopes. The computer program doesn't distinguish between a G placed in the middle of a rising sequence or at the beginning of a triumphant reveal. It hits both with the exact same authority.
The author can try to deal with this. It's possible, with great effort, to deal with some of it. You can put in volume changes, accents, and so forth. But it takes time, and there are no fine controls. If there were, it would take even more time!
That got me thinking. Why not make a program with "musicians" in it. You can assign a musician to an instrument, and teach them how you want them to play. You can carry these musicians from composition to composition, even sharing them over the internet. You can train them to play the horn with a certain level of attack and vibrato depending on what kind of sequence is being played at that moment.
Your controls would be largely limited to marking the music with emotional notation that guides the musicians. There's already denotation for sheet music, but this is not enough: we need more explicit terminology, such as "fiery".
If your pattern recognition was good enough, you could leave off with the labels and just instruct the musicians directly. They would work out what kinds of situations called for which kinds of performance, backed and instructed by you.
Or - here's an awesome idea - use the Nintendo Revolution controller to play as a conductor. Ha! That would make my YEAR.
I think there's already a program with this kind of capabilities out there. I've heard a couple of orchestral pieces which must be synthesized, but they're really good, instrumentally. They must simulate the actual instrument, which would probably do the trick nicely. They're still flawed, but they may actually be just badly mastered, mediocre-played real instruments. However, their sustains are still unnatural, so I doubt it...
Anyhow, just musing on the matter.
Subscribe to:
Posts (Atom)