Speaker Q&A: Leif Walter discusses the possibilities of artificial intelligence in games

Strategy games are about evaluating strengths and weaknesses of many moving parts. These strengths and weaknesses influence how players engage with the game, and it also affects AI decision-making. As a designer, the balancing of these parts is crucial to make the game engaging and interesting.

If you are such a designer trying to improve your game’s design, you should know you should know that Creative Assembly’s Leif Walter will be giving a talk at GDC 2018 discussing the company’s new approach to several old AI problems.

To get you ready for Walter’s talk, which will be in the design and programming tracks, we’ve reached out to him for a few answers about life as a game designer that may prove illuminating for GDC attendees.

Don’t miss out! The Game Developers Conference in San Francisco next March is going to be full of interesting and informative sessions like Walter’s. For more visit the show’s official website.

Tell us about yourself and what you do in the games industry.

My name is Leif, and I work as a Game Designer on Total War: Three Kingdoms, the next historical strategy game at Creative Assembly. As a designer, I work primarily on system design for various game features and game systems. But at the same time, I am responsible for content design and creation for some of these systems. Often enough, this means I get to stretch the boundaries of the very systems I designed.

What inspired you to pursue your career?

After university, I first pursued a career mathematical research, but I always felt like games were my true passion. I loved playing games, I loved making games. I always knew that at some point I had to start working in the games industry, and an opening at Creative Assembly seemed like the perfect opportunity to make my dream job come true.

Without spoiling it too much, tell us what you’ll be talking about at GDC.

I will be talking about an important core system of Total War: A system to evaluate unit strength in specific situations, which is important for AI decision-making, for generating battle results within the game, and for balancing during the development process.

What are some of the biggest challenges you face in your work?

Working in a large and multi-disciplinary environment where features are always connected to multiple teams is a great challenge. If you work on a system, it’s never just about getting the maths right. It’s also about the narrative. It’s never just about getting the art right, it’s also about creating a functional UI.

What are the most rewarding parts of your job?

Additionally, it’s very fast-paced and requires lots of versatility – it’s great, but it can be a challenge.
Mastering the challenge above is the most rewarding part of my job. As a designer, you are at the core of game development, bringing together artists and programmers, and working alongside with them to create the best game possible. Seeing things come together during that process is the most rewarding and enjoyable part of my job.

Do you have any advice for those aspiring to join your field someday?

Don’t feel discouraged to enter the games industry just because you studied something different or because you pursued another career first. Doing a degree in game design is good – but sometimes it can be better to acquire a deeper understanding on a specific field first to build up your core skills. Never underestimate core skills. As a designer, this skill set needs to be versatile – which can be learned and trained in many other careers or university courses such as nature and human sciences: Logic, empathy, emotion. It all comes together when working as a game designer.

What has been one of the most interesting parts of watching game AI evolve over the years, especially in the strategy genre?

My first experience with game “AI” was writing build order scripts for Age of Empires 1 – since then, lots of exciting scientific methods have permeated into game AI development: machine learning, fluid dynamics, cluster analysis, nonlinear programming, to name a few. As a former StarCraft 1 player, I was very excited to see machine learning being applied to StarCraft AI. But it’s not only great for developing AI opponents, machine learning can also support understanding of the game systems and inform game design. Allowing machine learning algorithms to find optimal strategies within your game can often reveal interesting and surprising insights about mechanics, systems, and how systems interact.

Have you had any opportunity to talk to AI experts outside of the games field, and learn about how their work helps yours?

Before joining the games industry, I did work with AI researchers on Poker AI. This provided interesting insight into AI development and games, as our AI systems revealed surprising strategies in well established games that human minds had not necessarily thought of. Especially in strategy games, where finding dominant strategies is part of the enjoyment, this kind of AI analysis is an interesting way to understand and improve on your own game and game design.

GDC 2018 will take place March 19-23rd at the Moscone Center in San Francisco. For more information, visit the show’s official website, or subscribe to regular updates via FacebookTwitter, or RSS.