Adaptive Futures Grid Using Recursive Trend and Mean-Reversion Estimates

Summary

This perpetual futures grid strategy adapts its range, direction, and exits using recursive estimates rather than a fixed lookback window. A Kalman filter estimates local trend and volatility, while recursive AR(1) estimation on a cointegration residual estimates mean-reversion half-life and equilibrium variation. Optional reference contracts help distinguish broad market movement from target-specific moves, and the system can fall back to target-only trend estimates when reference data becomes stale.

The design includes automatic, long-only, short-only, and two-sided modes. Before opening positions, it checks trend, net exposure, inventory, and drawdown limits; an aging position may be exited by moving its take-profit level toward the market. Grid spacing and range shifts respond to volatility, with additional constraints for fees and minimum order size. The published settings show a short one-minute test on an OKX perpetual contract, but provide no performance results. The strategy description itself cautions that leveraged grids remain exposed to extreme one-way markets; the code excerpt is incomplete, so implementation details cannot be fully assessed.

Key ideas

• A Kalman filter estimates local slope and volatility incrementally to guide grid direction and width.
• Recursive AR(1) estimation on a cross-asset residual supplies a mean-reversion half-life and equilibrium scale.
• Optional reference contracts help separate shared market trends from target-specific movement.
• Exposure, inventory, trend, and drawdown checks restrict new grid positions.
• The short published test settings contain no reported results, and leveraged grids remain vulnerable to sustained one-way moves.

Tags

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