The art of feature-lead game design

The art of feature-lead game design

By Mike Simpson

September 23rd 2011 at 9:00AM

Creative Assembly creative director Mike Simpson shares advice from Total War's design process

The Creative Assembly’s approach to design is a little unusual. The Total War series of games is ‘feature based’ rather than ‘level based’.

As players progress through the game, it’s not just new content they explore; they get to unlock new features with new game mechanics. We often have a huge feature set compared to an average game. That means we don’t have as much time to spend on iterating and balancing each feature.

The vast majority of them have to work first time. At CA our design efforts are guided by principles that help to achieve this.

The first is: ‘The design is in the detail’. High-level design ideas are easy to generate and almost entirely without value.

It still amazes me that in our industry many decisions get made primarily based on the quality of a high-level design pitch. There is little correlation between the quality of a concept and the quality of the outcome.

STYLE AND SUBSTANCE

Almost all of the quality comes from the fine-grained detail of exactly how a concept is actually executed.

We divide design into style and substance. Without style your game is dry and mechanical; players will not get immersed in it. Without substance it’s quite simply not a game.

Complex games require both style and substance, but designers tend to focus on style or substance, being artists or engineers, but rarely both.

There are perhaps four basic approaches to designing a feature – cut and paste, gut feel, design by first principles and the newcomer – design by metrics. All are valid and they can be mixed.

The cut and paste approach takes a feature that works well in another game and transplants it to your own. It’s a concept you’re taking, not an implementation, and the ‘design is in the detail’ principle means that the feature still needs to be re-implemented for your own game.

All you get from copying is proof that the idea can be made to work, not that it will actually work in the context of your game and implementation. There are some real risks with cut and paste. It’s far too tempting, for example, to pull in ideas from whatever masterpiece you’ve just played.

‘Gut feel’ design uses your imagination as a rapid prototyping tool; take a high level concept and imagine how it might play.

Sadly, the brain is brilliant at assessing the style aspects of an idea, but tends to skip over the details where the quality is found.

It’s great for designing the style aspects of your game, but for core gameplay it has to be combined with iterating; something we don’t always have the time to do.

A MATTER OF PRINCIPLE

‘Design by first principles’ is like ‘gut feel’ and iteration, but you work out why a concept should work instead of imagining how it plays.

It allows us to get complex features more or less right first time, giving strong gameplay without iterating.

There isn’t space here to go through everything we do, but for Total War games, the most important principle is ‘counterpoint’.

It’s the simple idea, at the heart of gameplay, that for every choice we put in front the player there needs to be a reason not to take it. Gameplay emerges in the player’s exploration of right and wrong choices.

In an FPS for example, taking time to aim carefully will get you kills, but taking too long will get you killed.

Every element of every feature can be analysed for counterpoint. If it’s too strong, the choice is never worth taking. If it’s too weak, the choice turns into a compulsory chore. When it’s just right, the feature is balanced.

Our game mechanics designers work in ‘counterpoint space’, pulling apart and re-assembling the factors that feed into player’s choices. This is more like engineering dynamics than a creative process, with positive and negative feedback, damping, lag and so on being considered.

‘Design by metrics’ is the first addition to the designer’s toolbox since the ‘80s, and we’re not ignoring it. Our games are instrumented, and we use the data to guide updates and future designs.

The Total War series may be unusual, but I think the basic principles of game design we follow can apply to all games, from the simplest board game to the most complex PC epic.