Article
files/pumpkin

Improving Game Design with Scary Ideas™

Alex J. Champandard on October 3, 2007

The best games are made of great ideas, many of which can be scary at first. Yet many developers hold back their ideas, while others aren’t very receptive to suggestions. By improving the flow of scary ideas, you have a better chance of finding great ideas among them.

This article introduces the ScaryIdea protocol from The Core. It’s a simple methodology for helping you get ideas out there, and help your co-developers become more receptive. Read on to learn about this protocol, and use it in practice by helping design a simulation game!

Dealing With Ideas

The human brain is poorly equipped for dealing with oddities and special cases. Brains are trained to deal with patterns and filter out things that don’t fit into their current belief systems, so there’s a natural defense mechanism against great ideas!

This is a problem game development for two reasons:

  • People mostly don’t share potentially great ideas. They might be scared or somehow ashamed of saying anything.

  • The other developers often don’t treat scary ideas with enough respect to find out if there’s a great idea within.

From this perspective, it’s a challenge to get great design ideas into a game, or great technological ideas into an engine. Certainly, the best teams manage this well enough given enough time to ship (i.e. “when it’s done”). But how can you do this more predictably and less subject to emotional reactions?

The ScaryIdea Protocol

The Core

Like the other Core Protocols, the purpose of ScaryIdea is to remove the drama and emotional baggage, specifically in the process of discussing potentially great ideas.

Here’s how Michele McCarthy puts it:

“People think, ‘Oh, I can’t say an idea because it would be impossible to implement.’ But who cares if it’s implemented? Just saying it might spark the perfect idea in someone else, or give you more insight into the real problem.”

There are two ways to do it.

  1. Do it yourself by stating “I have a scary idea, it’s…” and then explain your idea — no matter how scary it may seem.

  2. Engage the team by prompting “I want to do the ScaryIdea protocol.” Then go through each person and as them “Have you said all your scary ideas?”

Going through the motions of engaging the protocol has the effect of making your team members more lenient and receptive to what you have to say. In fact, if your team adopts The Core Commitments, you can expect a constructive response, especially if other protocols are used in combination with this one.

When to Use It

A. During fire-fighting meetings! If you become too focused on the little details, the big picture may escape you. So having someone raise a ScaryIdea in these cases may help understand the problem from a different perspective.

B. While planning for milestones or iterations. Here’s the ideal chance to come up with great ideas to help implement the features with better quality and in less time. Scary ideas here may help get the discussion started towards constructive solutions.

C. During traditional brainstorming meetings, whether design or technology related. In early game development stages, talking about ideas is easier than the previous two situations but it helps build a positive team attitude.

D. During coffee breaks! Don’t keep your scary ideas to yourself, and don’t wait until big meetings. Game developers are passionate people and always willing to discuss crazy ideas — particularly when high on caffeine!

Time to Practice

There’s a little simulation game in the pipeline here at AiGameDev.com. The plan is to make an innovative and educational game using artificial intelligence techniques. The design is in its early stages, but here’s your chance to help out! Post a comment below if you have scary ideas.

Dog Behaviors

Screenshot 1: Simple behaviors for a dog simulation.

Here are some of my scary ideas for the simulation game.

  1. Build a AI engine that’s scalable and easily edited to add new behaviors modularly — ideally by anyone playing the game. The dogs should be able to play games like tag, hide and seek, capture the bone, running around a track, and even play checkers or solve the towers of Hanoi puzzle. Of course, they’d also utilize their bodily functions…

  2. The player may take “control” of any dog in the simulation. The dogs still have their own autonomous behaviors: some of them override the player’s orders, but the other behaviors provide hints to the player what AI would have done in an intuitive way… This would be like a built-in tutorial for all the games listed above.

  3. The simulation would be networked, and the dogs could interact socially with each other. All the behaviors input by players would be recorded as sessions, then used by a learning algorithm to induce a suitable decision making and control system.

  4. As a spectator, when the player watches a dog in the simulation, its behavior is rewarded at the expense of all others not in the field of view. The underlying AI logic learns by reinforcement to this type of behavior again, which creates a positive feedback loop: you get more of what you watch!

An important part of the ScaryIdea protocol is responding to the ideas of others constructively, so feel free to practice by commenting on these points!

The Original Scary Podcast

Episode 77 of the McCarthy Show is where ScaryIdea is discussed. It’s a classic Jim & Michele podcast: the description of the protocol starts about 10 minutes in… :-)

Podcast: The McCarthy Show, Episode 77 — ScaryIdea Protocol

Remember, don’t hesitate to share your scary ideas: they’re potentially a seed for great ideas. Use the ScaryIdea protocol to encourage a positive response by your team members.

Next week on AiGameDev.com, you’ll learn about an efficient way to make unanimous group decisions that everyone is happy with.

Discussion 4 Comments

alexjc on October 6th, 2007

Hi Fernando, thanks for dropping by. You make a good point, discussions can very often be unproductive if not (self-)moderated... One of the protocols I plan to discuss next week helps with this. As for the simulation, the emotions are a great point, and simple to implement too! Learning from other dogs is also cool, which emphasizes the idea that all dogs should have unique behaviors... Anyway, nice ideas! :-)

mihaic on October 6th, 2007

Hi Alex, The Scary Idea Protocol seems to be one of the best things I heard about this week. I often do feel the that my colleagues take some of my ideas too lightly thinking that they are hard to realize. I believe I am going to show them this protocol and perhaps we can implement it and other Core protocols in our team. As for the dog game simulation, here is an idea: the players should be able to give positive or negative feedback to the dogs in order to make them learn or drop some behavior they already learned (not just by watching them). Perhaps some ownership concepts should be added of players owning one or more dogs.

alexjc on October 12th, 2007

Mihai, Thanks for your comment! Please let us know how it goes in your team, I'm very curious about adoption. It worked very well for us in parts of the team, but not everywhere... There's a Core Protocol [URL=http://tech.groups.yahoo.com/group/TheCoreProtocols/]mailing list[/URL] if you want feedback or want to share your experiences. Alex

mihaic on October 12th, 2007

Sure, as soon as I'll get them interested in the idea ;)

If you'd like to add a comment or question on this page, simply log-in to the site. You can create an account from the sign-up page if necessary... It takes less than a minute!