Part 1: Pre-Production vs Production
80% of mistakes in game development are the direct result of things done - or not done - in pre-production. These mistakes are the result of a myth:
MYTH #1: PLANNING
"It is possible to plan and schedule the creation of your game"
You Cannot Plan and Schedule Pre-Production. Why?
- Pre-production is a chaotic process.
- We are looking for fresh answers to the eternal question, "What is gameplay?"
- You cannot plan when inspiration will strike.
- You cannot schedule the date when you will have worked out these design problems.
Consequences of Planning and Scheduling in Pre-Production
Those who do plan and schedule Pre-Production end up in one of two boats:
- Get so frustrated and disillusioned with sticking to a plan that you cancel the project.
- Build a project which is not yet ready, and therefore carries all of its deficiencies into production to keep on schedule.
So how do you manage chaos?
Pre-Production: Managing Chaos
- Assemble core team
- Create successive prototypes
- Prototypes become like game levels
Assemble the Core Team
Use Your Best and Highest-Paid Staff
- Get the best people you can possibly find, and get them early.
- This core team will determine everything that's important about your game, and most likely go on to become your team leaders during Production.
- Start writing checks, because Pre-Production is not going to be cheap. You should expect to spend a million dollars before you know what your game is.
- By aiming to be cost-inefficient during pre-production, you are actually being cost-efficient for the actual production of your game. That's because if the game is not going to work out, you only blew a million dollars.
Do NOT Use Juniors
- Unfortunately, it is a common habit to kick off the Pre-Production phase using juniors.
- Juniors might not have the experience to determine everything that's important about the game.
- Juniors may not be prepared to lead the team when the game transitions to Production.
Prototyping
What are they doing with a million bucks? Prototyping.
- The team is creating lots of successive prototypes.
- It is important not to wait before you start making a prototype. Take the tech and art that you have, and build the best you can.
- Prototypes are where you learn and validate game design
At some point, the team shifts over to "real-level" prototypes
“Real-Level” Prototypes
- Prototypes that bring together artwork, game mechanics and technology to show an entire level of a "game that could be", were you just to finish it.
- These prototypes are targeted in that they become indistinguishable from game levels.
The Cycle
- Make a level.
- You (and others) play it.
- Review the game and its design. Finding the strengths and weaknesses.
- Change your design goals or your approach.
- Then you put it aside and start making the next level.
During the prototyping cycle is where you run head-to-head with another myth of development:
MYTH #2: PRODUCTIVITY
"Working productively means throwing out nothing"
You Will Throw Away your Prototypes
I'm going to give you a number. That number is 5.
- The fifth real-level prototype is where I predict you have defined your game.
- If you're really good, you might be able to do it in four.
- And if you're a team that's been together for a while, working in a familiar genre on familiar hardware, maybe, just maybe, three.
- Five real-level prototypes means that you are going to discard four completed levels of your game.
- By the way, games in this category tend to have about 20 levels when completed so we're talking about 20% of a game here!
You Will Throw Away your Art Assets
- If you care at all about efficient asset management, this is the time for you to kick that habit and throw out some good work.
- Throughout pre-production you will not create any material designed to be played by the public.
- If you are very lucky, you might find a use for the best of your Pre-Production work.
- Do not schedule for the creation of art assets.
If you've successfully kicked those bad habits, planning, scheduling, saving money, and asset-management, you'll be fine.
Pre-Production: You Are Not Making a Game
You Are Making a Game Design
Through real-level prototypes, you're building your game design, which includes the following things:
- 3 C's: character, camera, control
- Game look
- Key technology
- Holistic game design (if appropriate)
Three C’s: Character, Camera, and Control
You can't create any meaningful gameplay prototypes until these things are complete.
Character
- What does the character look like?
- How does it move?
- How does the animation support the moveset?
Camera
Camera is always a tough problem, and requires not only a fair amount of work, but also a sort of philosophy of camera movement.
- Will it be more or less responsive?
- Close or far?
- Biased toward certain directions?
- How does it support your gameplay?
Control
Control is how the first two integrate with the game controller into something that is fun.
Analogs of the Three C’s in Other Genres
These 'three C's' are somewhat specific to a character-action game, but there are analogs in every genre.
Examples:
- In a real-time strategy game, replace the word "control" with "interface".
- In a driving game, replace "character" with "car".
Game’s Look
- In order to be successful, a game must have a unique and compelling look.
- You need to have this look nailed in the game before exiting Pre-Production.
Completed key technology.
- If there is any technology that is vital to how your game will feel to the player, it must be completed during Pre-Production.
Examples:
- Spyro and Jak & Daxter, this was an engine that could do long views exceptionally well on the hardware.
- Luigi's mansion, all the realtime lighting effects needed to be complete before gameplay could really be called ready.
- Command & Conquer, you'd better know how your pathfinding works for all those units.
Holistic Game Design
Finally in the case of a 'holistic' game, you also need to know the bulk of your individual game design elements, and probably your story as well.
Holistic vs Sequential Design
Holistic game
- A game in which the various "levels" or areas are tightly interrelated by inventory, story, or other similar aspects.
- The whole game represents an interrelated puzzle to be solved by the designer. This puzzle must of course be resolved prior to entering Production.
- Example: Zelda
Sequential' game
- A game in which each level or area more or less stands on its own.
- Example: Jak and Daxter
Pre-Production: Take Chances
In order to get a game design worth building, you have to take chances. “Leap before you look”. You have to make best-guesses at how your game will work, and just try it out.
Taking chances means that you might find yourself in these situations: Building levels without complete knowledge of your character's moveset. Building levels without knowing the real limits of your technology. Not knowing the global context of the piece of the game you're working on. Etc.
All these are absolutely terrible things to be doing during Production, and are considered signs that your game is in serious trouble. But in Pre-production, these are exactly the kinds of things you want to be doing.
Vital Chaos
This is that "vital chaos" I keep talking about. When your prototype ideas crash and burn, who cares? You were throwing it out anyway! You can immediately, painlessly figure out what about your idea didn't work, and try the next great idea.
From a financial standpoint, you are spending money now to save money later. Ask yourself, have you ever cancelled a project near completion? Wouldn't you rather have done five prototype levels before fully funding it?
That's it
Pre-Production. The most important phase of your game development.
On to the next myth: Technology.
MYTH #3: TECHNOLOGY
"Cutting edge technology is very important, so build your technology first"
Technology is one of the three major components of a game, along with design and art, and that by getting the best technology you can "stack the deck" in your favor.
Example:
- Crash Bandicoot’s streaming technology let us have a very dense organic "look" that instantly sets us apart from competing products. It also allowed for very long, load-free levels as well.
The Conventional Method: Sequential
STANDARD PRACTICE
- Designers Write the design. Create a Design Document.
- Engineering extracts the technology requirements. Create "Required Technology" Document.
- Engineering builds the technology.
- Designers build the game.
Limitations of the Standard Approach
- It presupposes that we know the game’s design when we start out.
- It eliminates any chance of prototyping gameplay. Because before any of the game is playable, you're already in full-blown production of levels.
Solution: The Double Track Method
DOUBLE-TRACK IT!
- Build or borrow "Quick and Dirty" Tech for rapid prototyping.
- Build "Cutting Edge" Tech for actual game.
Example:
- Jak and Daxter - We want to do large, detailed worlds with infinite views,
- Part of the engineering team is working on level-of-detail technology.
- At the same time, designers have a "Quick and Dirty" engine that does not use level-of-detail and restricts view distance, but we have enough of an engine to start working on Jak's moveset, enemy behavior, even the look of the levels!
Later on, but still while we are in Pre-Production, we can take what we found out about the level requirements, such as their size and polygon count, bring it to the engineering team and finetune the engine specifications.
On to the next myth: Milestones.
MYTH #4: MILESTONES
"Frequent project review is essential to good management"
Milestones are a wonderful thing during Production; they allow you to break down the game into manageable chunks and set deadlines to track your progress.
However, traditional milestones should not exist during Pre-Production. Pre-Production is unpredictable, and therefore can’t produce viewable, playable milestones for review. Milestone deliverables require special effort, effort that takes away from the Pre-Production process. Any milestones produced in Pre-Production will have little resemblance to the real product and will do significant damage to the Pre-Production process.
Pre-Production and Milestones
- "Offline" status for team during Pre-Production
- No formal milestones.
- Pre-Negotiated terms for Pre-Production
- A limit on how long it can continue, since clearly if no progress is being made, continued experimentation doesn't benefit the team, the game, or the publisher.
- Strict deliverables for Pre-Production
- First playable
- Macro design
- Results of experiments are visible externally
- A brand-new experiment will be assembled and demonstrable every month or two.
- The timing of these builds of the game is determined by the natural flow of the work, not by external factors. But the team needs to invite management in, whenever practical.
- No "stench of failure"
- The lack of understanding that these are experiments and will all be flawed to one degree or another.
- So, it's crucial to have a corporate culture that understands how different a Pre-Production experiment is from a formal Production milestone.