 |
|
|
|
|
|
||
|
|
||
|
|
||
|
|
|
|
|
|
||
|
|
||
|
|
||
|
|
|
|
|
|||||||||||||||||
Add Forum to Favorites | Send Topic To a Friend | View Forum FAQ | Track this topic |
Last Thread Next Thread |
| Evolutionary Design |
|
 |
Anonymous Poster |
|||
 Posted - 1/19/2002 1:42:18 AM |
||||
| Imagine 'designing' a game without writing a 100 page document. How many companies do this sort of thing? -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
|
GBGames Member since: 1/25/2001 From: Chicago, IL, United States |
|||
 Posted - 1/19/2002 10:28:04 AM |
||||
| I found this article to be an easy read while still being incredibly helpful. I am making a Pong/Warlords clone, and I was intending to put in some kind of power up system. I was gonna just hack it until it worked, but after reading this article, I realize that I am gonna have to tweak it until there is a good game balance. I already had the reward system idea, but I got stuck trying to build the engine...it's in QBasic, and I haven't had much game programming experience before, but at least when I wrote the code I had a basic design to work with. I can easily drop in the code to move the paddles and tweak it without ruining anything else, such as hit detection. I might as well finish that part. |
||||
|
||||
 |
DavidRM Staff Member since: 7/16/1999 From: Tulsa, OK, United States |
|||
 Posted - 1/19/2002 12:38:58 PM |
||||
| Damn fine article...and not just because I find designing and playing board games Really Fun. :-) DavidRM Samu Games |
||||
|
||||
 |
Khawk Staff Member since: The dawn of time
From: Port Orange, FL, United States |
|||
 Posted - 1/19/2002 9:06:56 PM |
||||
| I suppose one could call this article a step in the right direction as far as taking systems engineering disciplines into the abstract arena of game design. Process is of utmost importance when it comes to developing quality products, and games are no exception. Kevin "Khawk" Hawkins CEO and News Director, GameDev.net Author, OpenGL Game Programming Developer Diary |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 1/19/2002 10:30:37 PM |
||||
| One could easily write a book or three describing a game design process and the useful patterns any designer should have at their disposal. Until now we've only seen a few philosophical dissertations and discussions of 'random things that seem to work'. I'll be disapointed if there isn't a synthesis of the decades of successful projects into a unified set of practices sometime in the next 5 years. It should be noted that a discussion of 'game design' cannot exist in a vacuum. You still need: - an organization with a shared goal - quality people - a solid programming process - a solid content development process. - a solid QA process A system that integrates all these items...now there is a book that would be worthwhile reading. -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
 |
Kylotan Moderator - Scripting Languages and Game Mods Member since: 3/8/2000 From: Nottingham, United Kingdom |
|||
|
Posted - 1/20/2002 11:46:48 AM |
||||
Daniel, this is the best article on game design that I've seen for quite some time. It sums up most of my approaches to design very succinctly with real world examples to back them up. Maybe there's hope for me yet, then  Well done. [ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost ] |
||||
|
||||
|
Anonymous Poster |
|||
 Posted - 1/20/2002 7:18:26 PM |
||||
| Kylotan, Thanks! What parts did you find particularly interesting? Which parts might need a bit more information? I'm always curious to hear what experiances other folks have had with these sort of techniques. -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
|
Kylotan Moderator - Scripting Languages and Game Mods Member since: 3/8/2000 From: Nottingham, United Kingdom |
|||
|
Posted - 1/21/2002 6:50:06 AM |
||||
| I pretty much agreed with everything you said, mainly because it makes sense to me. Sometimes it seems like a losing battle, trying to introduce any kind of analytic process into game design when there are so many people who think of it as some unquantifiable art. You also share my opinion on game clones, which so many people in the Design Forum consider to be the Antichrist in binary form. Many great ideas came from clones or sequels. (Deathmatch and order queuing are two that jump to mind.) Something that might be nice for such an article is some sort of checklist of key points. You have several lines or paragraphs in there that make for good summaries: "The basic problem with 'grand vision' game design is that it is incredibly difficult predict how a complex system will behave when you make changes to it's rule set." "All games rely on emergent behavior to define their play." "By making incremental changes and then immediately testing those changes, you minimize the unexpected disruptions to the system." "After every set of rule changes, play the game." And so on. The use of economics and psychology was especially good. (Although one could argue that when you use economics as a metaphor, you're just dealing with psychology anyway.) You could flesh it out a little by elaborating on the details such as attention (examples: something as simple as the score and high score in a Space Invaders game, or a cut-scene at the end of a mission). If you wanted to go one step further, it might also be interesting to analyse games that are commonly thought to have problems, and explain that problem in terms of your ideas. This would be in the vein of your Unreal/The Circle comparison at the end, but placed in the context of whatever concept or guideline you are talking about. I'm starting to sound more like an editor than a reader now, sorry Personally I think you could get a short book out of this, if you had a few other real-world examples to draw from. [ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost ] |
||||
|
||||
|
doomhunk Member since: 7/3/2001 From: United Kingdom |
|||
|
Posted - 1/21/2002 3:26:22 PM |
||||
| "Imagine designing a game without writing a 100 page document." I wasn't sure how I was supposed to take this comment :-) I have worked at games companies that did not believe in design and would create games and never write design docs. And then I have worked at games companies where designers wrote hundred page + documents that were never used because they wrote them in complete isolation from the realities of the game engine, skill sets of the team, they never thought to do a feasibility study, did not solicit technical feedback from artists & coders whilst writing the design etc. I can only fall back on my film-making experiences and trot out the old maxim that you can make a bad film out of a good script but you can't make a good film out of a bad script. Same applies to games. If you have a bad script/game design, your best choice is to ignore it altogether rather than try to follow it. I have seen so many projects get canned because of a director-level insistence on following a game design that should have been shot in the head at birth. I have seen producers and project managers overruled when they pointed out the flaws & weaknesses in designs and offered solutions to strengthen them. Megalomaniac company owners who think they know how to make games and go round saying things like 'It's a game for xxxx's sake, how difficult can it be to make a game!' or 'Don't worry about the design, just make the game' etc. Wow, your comment really pressed my button :-) What I totally agree with is that design is an analytical process. Some kind of training is always a good idea for a game designer, as is following a procedural path, a process. Ask any researcher. Employ more game designers with PhD's, at least they'll know how to go about the iterative processes, they've got years of experience of it. Totally avoid some guy with a lot of passion for games but who has no or little idea about design processes. He will most likely bring your company crashing down. There, my experiences summed up in a few paragraphs. Edited by - doomhunk on January 21, 2002 3:28:17 PM |
||||
|
||||
|
Anonymous Poster |
|||
 Posted - 1/21/2002 3:48:58 PM |
||||
| Doomhunk, I feel your pain. Back when we were working on Tyrian, it was seat of your pants, no design allowed development. Luckily we were evolving a concept that had been design to death over the past 3000 vertically scrolling shooter clones so we couldn't get too much wrong. There is a strong difference between not believing in design and designing with a strong minimalist process. The question to answer is "What is the most flexible process with the least unnecessary artifacts that is needed to produce the best possible product?" Those who do not believe in game design are in some ways lost already. I certainly don't wish to encourage a lack of design. The other side of the spectrum are those spend copious amounts of resources on 'big upfront design'. They have their heart in the right place, but I suspect an iterative process may be more effecient and result in a more enjoyable game. I would tend to steer away from the script analogy. A script is a description of a narrative construct. It is a planned sequence that is relatively stable from one plot point to the next. Games are fundementally different creatures. Unlike a movie script, a poor design document can result in a great game. Doom and Unreal are good examples. On the other hand great design documents can result in a horrible game. The reason is that because game play is an emergent property of the complex interactions of the ruleset, gameplay tends to immediately and rapidly diverge from the 'designed system' the moment development begins. The other elements you mention are perhaps more important. If you have the 'evil company owner', no matter what your process, the game is most likely doomed before it even starts. People matter. :-) -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 1/21/2002 6:06:05 PM |
||||
| Kylotan, There's a technique in business rags where each article has a small box with 'The take away, a list of summary bullet points. Maybe something of the sort might be useful. "Although one could argue that when you use economics as a metaphor, you're just dealing with psychology anyway" Absolutely. :-) I think there's a line in there someplace that states the entire point of the economic analysis is to set psychological value on certain activities that are then worked into the reward systems. I also like the idea of looking at a 'good example' and a 'bad example' throughout. That might add quite a few pages, but folks seem to find examples useful. Perhaps this is off topic, but it seems that game developers *love* hearing about what other game developers have gone through. It is like we are starved for stories about our industry. The popularity of the Post Mortems in Game Developer point to this. There is a certain comfort in realizing that the intensity (hell? ecstasy?) of game development does not occur in a vacuum. Editorial comments much appreciate. Very concrete and quite applicable. :-) -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
|
Martoon Member since: 9/12/2000 From: USA |
|||
|
Posted - 1/24/2002 11:43:17 AM |
||||
| Daniel, This was one of those rare, beautiful articles that takes existing concepts and presents them with a new, focused clarity. Most of the ideas (and ideals) presented were things that I already believe, but you've coalesced these many hazy feelings into a few small, concrete nuggets of truth. One thing that was new to me was the "climbing a hill" metaphor. I've always had an interest in alife, genetic algorithms, etc., so I'm quite familiar with the concept of seeking maxima in a solution space through an iterative process. (Speaking of which, is Anark ever going to give us anything as cool as Galapagos again?  ) For some reason, I've never made the connection to evolutionary game design before, but now that you've pointed it out, it's an obvious fit.As you illustrated in the article, you can arrive at the local maxima, only to find yourself on top of a very small hill (or "polishing a turd" as you aptly put it). Often your best bet is to scrap the whole idea, pick an entirely new point on the landscape, and start climbing from there. However, I think there are many such situations where the idea is not entirely bad, and the local maxima you're hung up on is just a small ripple on the incline of a larger hill. (Forgive the awful ASCII art)
You are here /
| /
| /
V /
_ /
/ \_/
_____/
In these cases, you may need a fairly substantial change to some fundamental part of your basic gameplay mechanic to get you over that little dip, but you don't need to scrap the whole concept. Of course, there are still the many other cases where scrapping the whole concept is the best solution. And one last point, regarding "clone" games. As you said, the great, original games that started new genres (the first FPS's and RTS's, etc.) basicly discovered new mountains. It's only natural that publishers are going to want to get on these same mountains, and that developers will inevitably find their way toward the same peak. I think, however, that these mountains don't just come to a simple peak. Rather, the top of a mountain is covered with many craggy peaks and valleys, and all kinds of interesting land features. The clones that I'm tired of our the many "me too" titles that are almost exactly like some established, well-known game, only not quite as good. It's like they were trying to get to exactly the same peak as the established game, but stopped somewhere on the upslope. If you're going to squat on top of the same mountain, at least find your own little knoll. As you can tell, I'm excited about this whole metaphor. Thanks again for sharing that with us. And thanks for the great article! Marty Rabens marty-at-rabens-dot-com Edited by - Martoon on January 24, 2002 11:48:18 AM |
||||
|
||||
|
Random Terrain Member since: 1/25/2002 From: USA |
|||
 Posted - 1/25/2002 2:50:02 AM |
||||
| Great article. I'm going to link to it from my "Quotes on Game Design" links page. Thanks for writing it. Random Terrain Edited by - Random Terrain on January 25, 2002 2:51:31 AM |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 1/26/2002 8:57:45 PM |
||||
| Marty, Aye, the "fairly substantial change to some fundamental part of your basic gameplay mechanic" occurs occasionally. Mario 64 compared to Super Mario brothers is a great example. I think the Homeworld guys have a philosophy (and I'm mangling this horribly) that amounts to "keep the game basically the same as a previous game, but add one big innovative element." The result is usually either a new genre or sub-genre. Just looking at the various game genres, most developers end up taking a very tiny jump. Age of Empires was Warcraft placed in a historical setting. Certain rule changes were added to make the game play better, but there really wasn't a 'substancial change' I'd argue that core mechanics change rarely, while the fluff mechanics related to setting show quite a bit of variety and innovation. Each FPS is basically the same game, but the exact pacing of the maps, the enemy placement, the accuracy of the sniper rifle etc vary dramatically from game to game. This actually points to why the evolutionary design process is in some ways incomplete. Though it provides a broad strokes describing how to design a playable game, it lacks a detailed process for facilitating the design of derivative games. We'd need to answer the following issues: - How do you choose a distinguishing factor related to setting - And then evolve that feature - while maintaining an almost exact clone of all other game mechanics. This paints professional game design in an unpleasant light. 95% of current game designers are simply people who mutate existing designs. In effect, their job is to clone a game completely, and then tweak it in some minor way. Naturally, there are exceptions. You can create a spectrum to describe a design and note how current games would rate: Original -----------Major Mutation---------- Minor Mutation 0.5% 4% 95% --- "It's like they were trying to get to exactly the same peak as the established game, but stopped somewhere on the upslope. " This points out something that is often overlooked. Though balancing an original game is a hideous amount of work, cloning a game has its own pitfalls. When an original game is created in an iterative fashion, each iteration builds upon the past iterations. The rules begin to support eachother in subtle unexpected ways. It's almost like you are building a pyramid, with each additional level supported intimately the rules below. When you clone a game, you look at the obvious rules of the game and implement them. However, the subtle interactions of the rules are not immediately obvious and are therefore not implemented. These interactions are lost, and the emergent gameplay is destroyed. It's as if you made a plaster cast of a digital watch, painted it exactly the same, and then wondered why it didn't tell time. The result is that game developers need to put the clone through another iterative process of balancing, where they slowly fill back in the lost iteractions between the game rules. If someone who doesn't actually play the game (a marketing person for example) looks at the bulleted features it looks like it is complete. '7th Legion' (a horrible RTS game from ages past) had almost every single feature that Warcraft had. Yet the iterative balancing and rule generation never was allowed to happen. The result: An empty shell of a game that has almost no gameplay value. -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
|
Random Terrain Member since: 1/25/2002 From: USA |
|||
 Posted - 1/26/2002 10:03:13 PM |
||||
| Daniel Cook: Is it OK if I use your quote on my web site? This is what I'm talking about: "Though balancing an original game is a hideous amount of work, cloning a game has its own pitfalls. When an original game is created in an iterative fashion, each iteration builds upon the past iterations. The rules begin to support eachother in subtle unexpected ways. It's almost like you are building a pyramid, with each additional level supported intimately the rules below. When you clone a game, you look at the obvious rules of the game and implement them. However, the subtle interactions of the rules are not immediately obvious and are therefore not implemented. These interactions are lost, and the emergent gameplay is destroyed. It's as if you made a plaster cast of a digital watch, painted it exactly the same, and then wondered why it didn't tell time." --Daniel Cook Thanks, Random Terrain |
||||
|
||||
|
Anonymous Poster |
|||
Posted - 1/27/2002 2:53:41 AM |
||||
| Random Terrain, Sure. :-) Nice site, btw. I enjoyed reading through it. -Daniel Cook www.lostgarden.com www.anark.com |
||||
|
||||
|
Random Terrain Member since: 1/25/2002 From: USA |
|||
|
Posted - 1/27/2002 9:16:31 AM |
||||
| Thanks! And thanks for visiting my web site. Random Terrain |
||||
|
||||
|
Kylotan Moderator - Scripting Languages and Game Mods Member since: 3/8/2000 From: Nottingham, United Kingdom |
|||
|
Posted - 1/28/2002 9:27:58 PM |
||||
quote: Someone else (someone at Westwood Studios, I think) said that you should add 2 new features to an existing game to make a new one. Any less, and it's not original enough. Any more, and it's too confusing for a new player. I would add that more than 2 new concepts might be too difficult to balance and test properly in a single project life-cycle. [ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost ] |
||||
|
||||
|
Danc Member since: 1/27/2002 |
|||
|
Posted - 1/29/2002 3:31:31 PM |
||||
| When I say things like "Most developers end up taking a very tiny jump", I am merely stating the current situation, not endorsing it. With a more robust understanding of game development dynamics, it is very possible for developers to create a high percentage of original games. It's easy to become cynical and say: - new games are too confusing for new players - new games are too difficult to test and balance properly The truth of the matter is that it depends on the game. There exist types or classes of games that are easy to develop iteratively and easy to balance iteratively. There are others which are incredibly expensive. If you are going to 'reinvent' the RPG or adventure genre, expect to spend a huge amount of effort. If you are going to create the next Lemmings or a fantasy sports game, expect to spend a much smaller amount of effort. The defining factor is simply the number of rules. With a puzzle game like Lemmings there are a small number of generalized game mechanics that cause the emergent gameplay. Changing these rules is easy. In a game like Final Fantasy, there are an enormous quantity of special case game rules that result in the emergent gameplay. For each incremental improvement in gameplay, you need a huge investment in design. Here's a fun question. Is the emergent gameplay of Lemmings more or less complex than that of Final Fantasy? One could argue that since Lemmings results in a larger number of interesting player choices, it actually has more complex gameplay. This seems like a good place to introduce the concept of game mechanic 'elegance.' Elegance is a measure of the richness of behavior caused by a minimal set of rules. If you have a dozen rules that result in hours of intriguing gameplay situations and discoveries, you have a relatively elegant game. Chess fits this category. On the other hand if you have hundreds (if not thousands) of rules that result in relatively fixed gameplay, you have a rather inelegant game. King's Quest and other adventure games fit into this category. So, it would seem: - Original Elegant games are inexpensive to create - Orginal Inelegant games are more expensive to create - Cloned Inelegant games are less expensive than Original Inelegant games. - Original Elegant games can be less expensive than Cloned Inelegant games. (Good old Sid proves this over and over again.) -Danc. Edited by - Danc on January 29, 2002 3:33:43 PM |
||||
|
||||
|
Random Terrain Member since: 1/25/2002 From: USA |
|||
|
Posted - 2/14/2002 1:08:51 PM |
||||
| If you haven't written a book yet, I wish you would. If a book is too much trouble, how about a whole game design web site with your thoughts on everything you can think of (that you can slowly build on over time)? If you don't care that much about making money from your thoughts on game design because you just want to get the information out to game designers and future game designers, the web site could be free for all to read (no subscriptions or anything). Whatever you choose, I would buy your book or visit your web site if it would have the same kind of stuff from your article and what you have been posting here (I love this stuff). (I tried to visit the web site address that you gave here, but I get "You are not authorized to view this page.") Random Terrain |
||||
|
||||
| All times are ET (US) |
Last Thread Next Thread |
|
