Cross-Sectional Crypto Momentum with Funding and Execution Controls
Summary
The article proposes a market-neutral strategy for volatile crypto perpetual futures. It ranks contracts using price momentum and funding rates, then holds equal-notional long and short baskets so performance depends on relative strength rather than predicting the market’s overall direction. Rolling cross-sectional regressions estimate factor weights; coefficient stability informs how much each signal contributes. The approach also normalizes funding rates across different settlement intervals, allocates risk inversely to volatility, and scales total exposure to a target volatility.
The author reports that risk-parity weighting improved three sample metrics and that four-hour rotation performed more consistently than very frequent rebalancing in the tested data. The strategy further emphasizes unified margin and execution controls: verify positions, keep long and short exposure balanced, repair incomplete fills, and close positions if balance cannot be restored. The evidence is specific to the author’s sample and implementation. It does not establish future profitability; results may depend on turnover costs, liquidity, exchange conditions, and whether the observed momentum and funding relationships persist.
Key ideas
- The strategy ranks perpetual futures by price momentum and funding rate to build opposing long and short baskets.
- Rolling regressions set factor weights and reduce the influence of signals with unstable coefficients.
- Funding rates with different settlement schedules must be normalized before comparing contracts.
- Inverse-volatility allocation and target-volatility scaling aim to manage portfolio risk.
- The author’s execution rules prioritize verified, balanced positions and require repair or closure when fills are incomplete.
- Reported performance is sample-specific and does not guarantee the factors will remain effective.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.