Game Engagement Prediction Using Real-World Resource Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users tend to disengage from video games despite efforts by game developers to create engaging experiences, leading to a need for systems and methods that enhance user engagement.
Innovation Solution
A video game environment engagement simulation system that includes a processor, communication interface, and memory device, which trains a prediction model using data from both online video games and real-world resources to predict user engagement events, and deploys this model to suggest strategies to maintain user interest.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If game developers design games to create engaging experiences, then user engagement is improved, but users still stop playing certain video games
Solution Approach 1:
The system performs preliminary actions by training prediction models in advance using historical gameplay data and user information from partner applications. The model predicts future disengagement events before they occur, allowing the game to proactively adjust parameters to prevent user churn rather than reacting after users stop playing.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring user gameplay behavior and using the trained prediction model to assess engagement risk. The model's predictions feed back into the game system, which then adjusts game parameters dynamically to maintain user engagement, creating a closed-loop control system that adapts to user needs.
2Measurement precision
If the prediction model uses data from multiple sources including partner applications, then prediction accuracy is improved, but data processing complexity increases
Solution Approach 1:
The system segments data processing by separating the prediction model training from the gameplay analysis. The model is trained offline using comprehensive data from multiple sources including partner applications, then deployed as a standalone predictive component. This segmentation allows complex multi-source data integration without complicating the real-time gameplay processing system.
Solution Approach 2:
The trained prediction model acts as an intermediary between raw multi-source data and game engagement decisions. The model consolidates and processes information from diverse sources including gameplay data and partner application data, transforming it into actionable engagement predictions that the game system can use without directly handling the complexity of multiple data sources.
Data Source
AI summary
Systems and methods train, using training data, a prediction model by iteratively predicting a target variable value of an engagement event associated with an online video game application, identifying an error between a prediction and the target variable value, and modifying weights of the prediction model for multiple iterations. The training data includes information obtained from the online video game application and a partner computer application that stores resource data of real-world resources to a real-world location and also includes specific data such as a duration of gameplay, a quantity of instances of gameplay, and a quantity of resource transactions of the virtual resource. The trained prediction model is deployed and applied to user data of users to predict attributes of the engagement event that are most likely to cause users to engage with the online video game application within a predefined number of days.