A Gamma-Based Regression Model for Market Price Prediction

Summary

The article proposes a price forecasting model built from transient functions related to the Gamma and Erlang distributions. It divides a price move following a disturbance into three components: a remaining external force, the current rate of change, and the cumulative change already realized. The model requires these components to satisfy a normalization or balance condition, with single-cell and multiple-cell forms described.

For calibration, the author derives model parameters from observed prices by estimating interval changes and their timing, then applying a semi-logarithmic regression to the resulting series. The text also describes corrections to the initial price and total move using areas under actual and theoretical curves. It claims the balance identities hold algebraically and refers to practical testing, including an EUR/USD illustration, but the available excerpt gives no performance metrics or rigorous out-of-sample evaluation. The approach depends on identifying a disturbance and assumes the move follows the proposed transient shape; the article itself notes market behavior can shift between structured and apparently random regimes.

Key ideas

• The model represents a price move with force, instantaneous change, and cumulative change components.
• Gamma- and Erlang-related functions describe the components in the multiple-cell formulation.
• Parameter estimates are derived from observed price changes using numerical differentiation and regression.
• The component values are normalized to sum to the total disturbance.
• The article offers limited empirical evidence and does not establish robust predictive performance.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.