Crypto Funding Reversal
Mean-reversion strategy trading perpetual futures funding rate extremes across major crypto exchanges (Binance, Bybit, Deribit).
Edge concentrated in 2021-22 funding spikes. No persistence post-2023. Out-of-sample Sharpe 0.21.
How It Works
Fade extreme funding rates by taking opposite positions when funding reaches historical percentiles, expecting reversion to neutral.
Mechanics
- 1.Monitor perpetual funding rates for BTC, ETH across 3 exchanges
- 2.Enter short perp when funding rate > 95th percentile (e.g., +0.1% 8hr)
- 3.Enter long perp when funding rate < 5th percentile
- 4.Hold position until funding reverts to 50th percentile or 7-day timeout
- 5.Hedge spot exposure with inverse perp on backup exchange
Signals
Performance Results
Implementation
BTC-PERP, ETH-PERP on Binance, Bybit, Deribit
Aggressive limit orders, 5-minute TWAP for larger sizes
Event-driven (triggered by funding extremes)
Max 2% portfolio allocation, -5% stop loss per trade
WebSocket feeds, Python execution, CCXT library for exchange APIs
Risk Analysis
Funding Persistence
High Impact7-day timeout prevents indefinite holding
Exchange Risk
High ImpactDiversify across venues, but insufficient in 2023 blowups
Regime Shift
Critical ImpactNone effective - post-2022 funding volatility collapsed
Execution
Medium ImpactLimit orders, avoid market orders in thin books
Backtest Results
Stress Period Analysis
Funding spikes frequent, strong edge
Negative funding extremes profitable
Funding volatility collapsed, edge disappeared
Deployment Status
Rejected - not deployed
Rejection Analysis
Post-2022, crypto perpetual funding markets normalized. Funding rate extremes became rare and short-lived. Out-of-sample testing (2023-2026) showed Sharpe of 0.21, well below deployment threshold. Edge appears structural to 2021-22 mania period and not persistent. Risk Committee rejected deployment due to regime dependency and lack of recent performance.