Orbit Wars Elite-Player Trajectories: 2.2 Million Decision States

2.2 million decision states were collected from real Orbit Wars games played by high-skill players with an Elo rating of 1650 or above. The data includes both value and policy information, likely representing game states and player actions. The dataset was uploaded to Kaggle, but the author, organization, and specific collection details are unknown.

Use Cases

• Train reinforcement learning agents based on high-level player decision states.
• Analyze policy and value estimation patterns from expert gameplay.
• Benchmark imitation learning models against elite player trajectories.

Strengths

• Contains 2.2 million data points, providing substantial volume for model training.
• Data originates from high-skill players (Elo 1650+), representing expert-level decision-making.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the specific game and player pool on Kaggle.

Provenance

Source

Kaggle

Collection Method

Collected from real Orbit Wars games.