Game Design Meets Machine Learning

Note: This documentation is still under construction. There will be links that go more in depth.

Measurement Of Fun And Its Relation To Engagement

Measurement Of Fun	Relationship To Engagement
Session Length	The more the player is engaged, the longer the player stays.
Map Coverage	The more the player is engaged, the more the player explores.
Variety Of Items Collected	The more the player is engaged, the more the player collect different items.
Amount Of Resources Spent	The more the player is engaged, the more the player spends on resources.
Quest Completion	The more the player is engaged, the more the player spends time on completing quests.
Number Of Online Players In A Server	It is related To players’ individual session length, where more players overlap means very high session length per player.

What’s Your Goal?

• Goal Maximization -> Use “measurement of fun” metrics as rewards and combine it with reinforcement learning models.

• Prediction -> Use regression and classification models.

• Best Middle Values -> Use clustering models.

Model Calculation Speed Vs The Game Engine

• Per Frame (Physics / Render) -> Model must be fast. Ideally use single datapoints or online models here.

• Per Interval -> Model calculation time must not exceed the interval. Ideally use mini-batch training here.

• Per Session End -> Batch training is allowed.

Game Environment Data Is Far More Cleaner Than Real World Data

• Noise usually comes from overlapping interactions.

• Your model’s global optimum might be a real global optimum.

• Game environment states are just a series of physics calculations. Your model may accidentally associate certain things with certain states!

Intepreting Local And Global Optima In Game Design

• Local Optima -> The best solution for anything related to the current game session.

• Global Optima -> The best solution for all game sessions.