Making Better Designers

About a week ago, I was in Paris talking with other Game/Level Designers about how to do our jobs better. We covered a lot of interesting topics, but one evening over dinner, we hit upon an interesting truth: How good a designer you are depends mostly on how much design you get to do.

In other words, how fast you can iterate, and how much trust you have with your team to try new ideas, will dictate almost everything about how good a game designer you become. And this means: a game dev team with super-specialized roles and tools will create employees that have less value.

Author’s Note: I use the terms “Game Designer” and “Level Designer” kind of interchangeably, because I don’t see a useful distinction between these skill sets. A Level Designer is a game designer who can work with gameplay spaces as well as game systems and mechanics.

Time to Not-Suck Isn't About Time

Newbie designers make rookie mistakes and produce low quality work. It’s a learning phase that we all have to go through. With good leadership and a positive attitude, every iteration that a designer does brings them closer to producing better design.

So the question is, how long does it take to make a newbie designer into a good designer, or a good designer into a great designer?

I’ll pick a round number for convenience: Let’s say “It takes a hundred iterations to make a newbie into a good Level Designer”. Now, at your studio, how long does it take to reach 100 iterations?

I’m not going to name any “overtime crunch factory” studios in this post, but: we all hire people from crunch factories, and we mostly hire them because we know that with 2 years at [crunch studio X] a given designer has already done hundreds of iterations.

Designers need time to become good, but it’s iteration not time that triggers improvement. 100 hours at my desk, headphones on, doesn't automatically make me a good Level Designer. I become a better designer each time I iterate on my work and try to create something better.

Iteration-Friendly Tools Enable Learning

I’m a huge fan of making level designers more efficient with systems and flexible data structures. I want to work faster, but this isn't just about saving time (and money).

Tools allow designers to iterate faster, or slower. That means your design (and your designer) evolves at a speed which is influenced by your tools pipeline, and by the extent to which your designers are working on improving player experience, not ‘manipulating the many bits of data required to test or reach an iteration’.

Keeping my not-scientific number of 100 iterations to reach the level of “good designer”, we can assume that for every 100 iterations, a designer could become a “better designer”. The faster a designer can iterate and evaluate results, the faster he/she can improve and grow.

In Large Teams, Not Everybody Will Improve

As games grew in scope for the 360/PS3 generation, teams often became very large. The teams that became large frequently also became specialized so that they could control production quality — once a game project has an established core vision, bring in many people, and execute the Grand Plan.

Sadly, specialization and lack of designer autonomy reduces the effectiveness of each designer over time. Designers that do less design and touch fewer systems don’t improve, and may stagnate.

Teams that have a large hierarchy may (inadvertently) block designers from feedback on their design. For example, a top-level review could trigger a task which reads “we are changing mission X so enemies will attack from the west” — without insight into why this change will be better for players. Designers will technically complete an iteration, but will not learn.

On large teams, only the highest-autonomy designers will be improving at a normal speed, and higher headcounts will reduce iteration time drastically. Even with great tools, fewer iterations will occur because of the need to co-ordinate people.

Do you even script, bro?

Disclaimer: I’m not a good programmer. I never studied code; I just picked up some VB script and C+ along the way because it helped me make cool things, and because I kind of had to that one time.

That said, I believe every designer should be capable of doing simple code, and that our tools should support this. I’m a fan of designers being able to try ideas without the help of a trained programmer. I think this is part of the appeal of using Unreal 4, or prototyping in Unity.

I want to design systems, not just content… and to do that, I need my tools to allow me as a designer to create systems. Designers that can work this way and try crazy ideas without support from dedicated programmers, will be iterating in a truly powerful way.

100 iterations in which I can design and test systemic content is exponentially more powerful than 100 iterations in which I move objects around inside a level. To get better, I have to do more.