---
license: mit
language: en
multilinguality: monolingual
size_categories: 10K<n<100K
language_creators: expert-curated
task_categories: time-series-forecasting
task_ids: binary-classification
---

# MLB Game Prediction Dataset

## Dataset Description

### Dataset Summary

This dataset contains historical MLB game data from 2000 to 2024, including team information, game results, team statistics, game statistics, and Vegas odds. It is designed for predicting MLB game outcomes.

### Dataset Structure

The dataset consists of several CSV files, each representing a table in the database:

- **teams.csv**: Information about MLB teams
- **games.csv**: Historical game data including scores and outcomes
- **team_stats.csv**: Team statistics for each game date
- **game_stats.csv**: Detailed statistics for each game
- **vegas_odds.csv**: Vegas odds for each game
- **predictions.csv**: Model predictions for game outcomes
- **model_performance.csv**: Performance metrics for prediction models

### Data Fields

Please refer to the individual CSV files for the specific fields in each table.

### Data Splits

The data can be split by season for training and testing purposes.

## Additional Information

### Source Data

The data is sourced from the MLB Stats API and Vegas odds sources.

### Maintenance

This dataset is automatically updated when the live database is updated.