Abstract
The human player is often required to interact and cooperate with synthetic characters, which also cooperate and interact with each other. However, unless the action is tightly linear and scripted, the expression of interaction is often confusing and difficult to understand by the human player. This work explores how traditional animation principles can be applied to the expression of interaction between the actors, both synthetic and human, and make the communication and cooperation more believable and the experience more immersive. To validate our work, we implemented our model in a multiplayer sports-game, where each character is an artificial player, and asked participants to evaluate videos of the interactions. The data we collected suggests that our approach not only significantly improves believability, but also makes the interactions between agents easier to understand and the action easier to interpret.