VOLatility Archive for Realized Estimates (VOLARE)

Battle Count: 56

VOLARE provides directly applicable realized volatility and covariance measures essential for quantitative trading strategies. The HAR and MEM models are widely used in practice for volatility forecasting, which is critical for options pricing, risk management, position sizing, and portfolio allocation. The open-access nature and standardized methodology make it immediately usable for backtesting and live trading systems. The multivariate covariance measures support portfolio optimization and correlation-based strategies. However, the platform focuses on estimation and forecasting rather than direct trading signal generation.

The paper describes a well-documented pipeline with clear parameter choices and procedures. The back-end involves data acquisition, cleaning (Brownlees-Gallo procedure), sampling, and computation of realized measures - all standard in financial econometrics. The front-end uses React with PostgreSQL and MinIO storage. For users, the platform provides direct access without requiring implementation. For replication, the methodology is transparent but requires access to Kibot data and computational resources for tick-level processing. The volatility models (HAR via OLS, MEM via QML) are straightforward to implement.

VOLARE is an open-access research infrastructure (https://volare.unime.it) providing full documentation of data processing pipelines, parameter choices (k=60, delta=0.1, gamma=0.06 for outlier detection), and methodology. The paper provides detailed technical documentation of all econometric procedures, data sources (Kibot), and model specifications. The platform allows direct dataset download and interactive model estimation, facilitating full replication of results.