Game Design legend Chris Crawford explains a fundamental mistake made by many game designers:
One of my most deeply-held beliefs recently gelled for me. Throughout my career, I have devoted my most serious design efforts towards working out the algorithms that form the foundation of my games. My thinking here is so deeply held that only recently have I realized that other people don’t think that way. They think about games or interactive storytelling in terms of graphics, mechanics, user experience, gameplay, and all manner of other minor details. They seem to think that game design and interactive storytelling are just a matter of putting all the right features in place.
Frankenstein’s Mistake
To make my point, I’ll use an analogy based on the Frankenstein meme. The basic strategy used by Dr. Frankenstein was to stitch together a lot of pieces of the body (features) and then ‘animate’ them with electricity. To him, a human being is a collection of pieces: arms, legs, kidneys, lungs, heart, brain, and so forth. Just put them together, give them a jolt of electricity, and you’ve got yourself a human being.
That’s entirely wrong, and it could never work, because the human body is not a collection of pieces; it’s a system of processes. To build a human body, you start with the most elementary processes: cell metabolism. Once you’ve got the biochemistry of a cell working, then you have to design lots of specialized cells: muscle cells, cells that manufacture and secrete special hormones, nerve cells, blood cells, and on and on.
Once you’ve developed the ability to manufacture all the different kinds of cells, you need to start assembling them system by system rather than piece by piece. Perhaps you start by putting together a skeletal system, then adding a circulatory system, a nervous system, musculature… things get very complicated here. But the key point is that you build it system by system rather than piece by piece.
In exactly the same manner, games and interactive storytelling are not assemblages of features, they are systems of processes. People do build Frankenstein games by just stitching together a bunch of features, but the results are as clumsy and stupid as the Frankenstein monster was. If you want to do it right, you’ve got to stop thinking in terms of the conventional features listed above and instead think in terms of the processes that you will build into the design.
What he's talking about here includes my observations on imitative design, but he goes much deeper than that superficial categorization. What he's saying is that it is the interlocking elements that go into the various features of the game, all of which ultimately depend upon the algorithms that create them, that should be the the designer's focus. This simultaneously requires an ability to think in both holistic and highly detailed and technical terms, which is why there simply aren't very many good game designers out there.
To the extent that Crawford believes it is necessary for the designer to write the algorithms, or even comprehend the math underlying them - and I am not saying he does, I'm am only addressing the theoretical possibility - I would tend to disagree. But the designer needs, at the very least, to understand the purpose and the output of the algorithms that underlie the systems of his design.
The next important steps is making sure these subsystems interact and link together in a coherent, reproducible pattern. E.g., a routed unit doesn't spend 2x its movement capability in retreating, nor would it run forward through an attacker. Simple cases, but need is to map subsystems inputs and outputs to the same control space description.
ReplyDeleteI always imagined that a game started life as a core of code (read: algorithms) that you wrapped visuals around. I guess I'm not surprised at the end of the day - AAA games lately are as architecturally sound as a bridge with no supports - but it confounds me somewhat.
ReplyDeleteLong gone are the days of designing squeezing 2x the game into the same sized cartridge, a la Pokemon Gold/Silver, I guess.
Hes right of course, but he also made very spreadsheet like games, so that was the meat.
ReplyDeleteThe more action oriented,the more the parts matter more,and tuning rather the pre-planning becomes more important, because you cant really know the variables you actually have at your disposal until they are present and working, unless its well tred ground.
A middle ground game like RTS can be planned much more like a spreadsheet, a game like doom is more tuned once the bidy is constructed, because the question was if it would come together technically as pieces first. Then a design round happens to actually make it fun.
If you don't have that leeway its going to suck, unless again it can be planned in advance because its well tred.
Not disputing the message, just tge mirror from a different state of play.
Im building a game now and the real design must come after the pieces are built, because I'm not sure what will actually exist that can then be properly modeled into a real game.
So frankenstein monster first, then teach it to be a game after i can see how it moves.
The deeper the analysis, the more successful the synthesis.
ReplyDeleteThe people designing games, or anything else, by stitching parts together are usually completely unaware that the "processes and systems" exist. They take the components they know about and combine them.
The trouble with having a deeper understanding of more fundamental layers of abstraction is that after synthesizing them into a new design, one must do more work to rebuild the layers on top in accordance with the new underlying systems and processes.