Capping Long-Only Grid Exposure with Fractional Kelly Sizing

Summary

This research prototype separates a long-only grid’s entry and exit rules from its total capital budget. A trend filter permits new buys only in a bullish regime; ATR and transaction-cost estimates set grid spacing, and each filled batch keeps a target based on its actual entry price and original spacing. Grid levels receive declining incremental weights rather than larger buys deeper into a decline. A Kelly layer sets the maximum aggregate exposure, while execution and risk controls handle fills, reconciliation, and position reductions.

Instead of estimating Kelly from completed grid trades, which can omit unrealized losses, the design samples mark-to-market equity returns and normalizes them by exposure. It searches empirical samples for the allocation that maximizes average log growth, then discounts the estimate for sample confidence and applies fractional Kelly. The article includes illustrative calculations but reports no validated performance comparison. It explicitly warns that grid inventory can build during extended declines, Kelly estimates are noisy, and fees, slippage, funding, gaps, and out-of-sample testing remain important limitations.

Key ideas

• The grid determines trading levels while the Kelly layer caps the total capital the strategy may expose.
• Mark-to-market equity sampling includes unrealized losses that completed-trade win rates can conceal.
• Fractional Kelly sizing is discounted for limited sample confidence and can set the permitted allocation to zero.
• Spacing incorporates minimum limits, ATR-based volatility, and an estimated transaction-cost floor.
• A lower estimated budget can require batch-by-batch position reductions, which may add turnover and slippage.

Tags

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