Why I'm Using a 30 Year Old Development Method in 2018
And by far the hardest part of making games is pre-production. That statement may be confusing. We've all heard about very difficult production periods for games, and we've also observed light, easy, fun pre-production periods. So why am I claiming that pre production is harder? Because one of the things that can poison production is that pre-production is happening during it. As hard as pre-production is, it’s even harder (and way more expensive) to do while you’re also in production. Allow me to explain:
In an ideal world, you'd never go into production on a commercial game that's going to fail. If you were making a game with the intent of turning a profit, and you knew that game would not do so, you wouldn't go into production.
In spite of the obviousness of that statement, many projects, including projects of which I have been a part, do not see their look, mechanics, sound, and level design come together until the game is almost finished. The tail end of production is a terrible time to find out whether or not a game works. Frustrating or underwhelming mechanics will need to be thrown out or reworked, which will have a hugely expensive cascading effect on the technology, art, and sound.
This synthesis of all the parts (aka first playable) is also the first time you find out whether or not your team is up to the original task. This is less than ideal if they've already been working on it for years.
"But game development is inherently chaotic," the strawman in my head complains. Well he's right. It is. Innovation requires time, talent, discarded work and endless iteration. How do we accommodate that chaos and allow for innovation without risking the whole farm?
Remember when I said pre-production is harder than production? The reason so many projects have such nightmarish development is that the pre-production was never truly finished. They may have put together some concept art, a few documents, maybe even a white box prototype. But doing this is a far distant thing from completing pre-production, at least according to Mark Cerny.
In Cerny Method, by the end of pre-production, you should have one level of your game that is indistinguishable from a completed professional product. And there can be no cut corners. A demo is not a first playable. An alpha is not a first playable. A whitebox is absolutely not a first playable. The art, animation, mechanics and sound must all be working together, they must be great, and together they must represent something which will outsell competition 2 years from now.
So why bother making a first playable if it’s so much work? Because according to Cerny,
“It’s the only way to give yourself a proper chance of making a good game.”
Cerny Method isn’t just about keeping chaos out of production; it’s about keeping conservatism and arbitrary schedules out of pre-production. Games require innovation and taking chances, and you won’t want to do that once the production train is rolling.
By first deliberately setting aside time for chaotic, rapid iteration, you empower yourself and your team to come up with interesting new ideas. You are free to change things on a whim and try ideas that have never been done before. Nothing is too dangerous, too weird, too unfaithful to the genre, because you’re not counting on taking any of the things you make into production in their current form. In fact you should proceed as though you’re going to throw it all out. Then, as the game’s concept starts to congeal, you take your best ideas and give them the most polished presentation you possibly can.
Once you have endured the flames of pre-production, you know so many things about your project that most developers don’t get to learn until years into production. You know what works, what doesn't, what's fun, and what makes players hate you. You know how the art, animation, and sound mesh with and support the mechanics. You know all the things your avatar can do, and therefore what the levels, obstacles and/or enemies are like.
And perhaps best of all, because you've been iterating on so many levels, you know how long it takes to make them which means you can actually make an accurate schedule and budget. You are now free to build your game, armed with the knowledge of exactly how to do it, how long it will take, and what it will cost. But there's one giant caveat emptor:
The linchpin of Cerny Method is this: not every game will survive pre-production. If you have made 5 level prototypes, and you don't have that publishable first playable, that lightning in a bottle that will outsell the competition 2 years from now, you have kill the project.
“There’s no point in following this process if you’re not going to old the output of pre-production to a very high standard.”
It sounds harsh, but remember the entire point of Cerny Method is to ensure that you don't find yourself in the situations I described earlier, where you're in full-blown production mode on a doomed game that will never work.
Skipping pre-production is like launching a rocket and then trying to decide where it should land and figure out its trajectory and how much fuel it needs, all while it’s already in the air. When I look back on my failed projects, there's absolutely no question that I'd much rather have cancelled after 6 months of pre-production than 3 years of production.
But if you do reach the end of pre-production, you have that publishable quality first playable, your game is incredibly great, then congratulations! From here the work is easy; all you have to do is make the game. As Cerny puts it,
The publishable first playable proves that you understand precisely what your game is and what it is not, and that you know enough to produce the game. It delineates pre-production, where you know nothing, from production, where you know everything.
I'm being flippant of course, there's nothing easy about making a game. But imagine being able to make levels knowing with 100% certainty that your character's jump distance isn't going to change. Imagine being able to look at the overarching design of your game as a single-page spreadsheet. Imagine not needing to constantly re-write documentation, throw out production-quality assets, and have design arguments while a release deadline is looming.
I don't think everybody needs to use Cerny Method. It's outdated for a reason, plenty of teams are perfectly capable of using Agile or another iterative method. But I'm spearheading a game of my own for the very first time, and I know I'm not the only developer doing so
A studio is far more likely to be able to dust itself off after a failure than an indie mortgaging their home. If you're thinking of putting your own money into a game, for god's sake do pre-production. Don't wait until 90% of the production budget's gone to test a concept
There are limitations to following Cerny Method to the letter. One of them is that once you transition from pre-production to production, your mindset has to switch from boundless creative openness to staunch conservatism. And you can see this play out in Cerny games. The later levels of Crash 2 don’t have much going on that isn’t present in some form in the early levels. The difficulty increases, and the obstacles are re-combined in new ways but the fundamentals remain more or less the same.
Crash offsets this with what Cerny calls a “special mechanic”. For Crash 1, the special mechanic was riding a wild boar. But the special mechanic cannot be an after thought in Cerny Method; it must be iterated and tested during pre-production just like the base mechanics.
Crash 3 is the most varied in the series, giving Crash a new ability after each boss. Must of the uncertainty is mitigated by making these powers variations on existing abilities. The only truly new ability is a bazooka. It is also the most cumbersome mechanic in the game, and does not show the same refinement as the other mechanics. The bazooka is an example of what can happen when you aren’t conservative enough during production.
I dearly love games like Portal 2, Rayman Legends, Psychonauts, and Antichamber, where each new area seems to offer more innovation than the last. Games like these probably couldn’t be made in Cerny Method because it’s not possible to separate their production and pre-production in that way.
Games like these require constant (CONSTANT) iteration and playtesting of near-production-quality levels. Otherwise every new level will feel like an untested prototype. Ubisoft and Valve have the resources to do this, and I do not. Antichamber’s development doesn’t bear comparison because it took so long to make and because the art is so minimal (brilliant, but minimal.)
I’m making a game with a tiny team and very limited resources. I don’t want it to take ten years and I don’t want to reach the end of development only to learn that I was doomed a month in. That’s why I’m using Cerny Method.
andulkaphoto
tangentwithtanner
origamiaround
izzy806
jbbartram-illu
blakekathryn
