How Position Buffers and Dynamic Optimization Affect Trading Costs

Summary

This document investigates whether fast trading rules can contribute to a futures portfolio without causing proportionally large trading costs. It traces how rule forecasts become positions through volatility and currency scaling, contract rolling, buffering, and, in a newer system, dynamic optimization that penalizes costly trades. The author’s hypothesis is that these layers may damp small changes from fast rules while allowing unusually large forecast changes to affect positions.

The examples compare turnover at forecast, raw-position, and buffered-position stages, then examine strategy variants across instrument sets, trading speeds, and optimization constraints. The reported results say static buffering reduced turnover for a fast rule, while dynamic optimization cut costs more efficiently; excluding expensive instruments from trading helped, but excluding them from forecast construction did not. Allowing expensive instruments to trade faster rules raised costs and changed correlations with a fast strategy. The findings come from the author’s backtests and system configuration, with stated simplifications and a turnover measure that can rise in diversified portfolios. Results should not be assumed to transfer unchanged to other markets, cost models, or execution systems.

Key ideas

• Trading costs and turnover depend on both the forecast speed and later position-management steps.
• Position buffering can suppress small changes and reduce turnover before orders are placed.
• Dynamic optimization incorporates instrument-specific trade costs when selecting integer positions.
• The reported tests found dynamic optimization more efficient at reducing costs than static buffering.
• Cost penalties did not fully prevent expensive instruments from trading, so constraints affected the results.

Tags

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