BTC Coin-M vs USD-M Funding-Basis Carry — Market-Neutral Cross-Margin Structural Cash-Flow Harvest (Short the Richer-Funded Perp / Long the Cheaper-Funded Perp When the 8h Funding Differential Between BINANCE_CM BTCUSD_PERP and BINANCE USD-M BTCUSDT Hits a Trailing Extreme, Delta-Neutral, Rebalance-Banded, 3-Parameter)

A DELTA-NEUTRAL, CROSS-VENUE structural carry harvest on the SAME underlying (BTC) expressed through two co-listed perpetuals with DIFFERENT collateral regimes: BTCUSD_PERP.BINANCE_CM (inverse, BTC-margined) and BTCUSDT.BINANCE (linear, USDT-margined). This is NOT a price-prediction strategy — it does not forecast BTC direction, momentum, or reversal, and it is NOT a beta/residual/cointegration reformulation of any directional idea (explicitly avoiding L95/L96). The traded object is the FUNDING-RATE DIFFERENTIAL between the two contracts, which is a real, mechanically-paid cash flow, not a mean-reverting price spread. Coin-margined perps are held disproportionately by BTC-native, structurally-long participants (miners, treasuries, BTC-denominated books) who accept/demand different funding than the stablecoin-margined crowd; this produces a persistent, regime-dependent skew where CM funding diverges from USD-M funding, especially during directional stress. When that differential reaches a trailing extreme we SHORT the richer-funded leg and LONG the cheaper-funded leg, both sized to equal USD notional so net BTC delta ≈ 0, and simply collect the differential cash flow every 8h until it normalizes. Horizon is multi-day (median hold ~5–10 days) precisely to let the funding accrual dominate the one-time entry/exit fees. Fills three of the furthest-from-target diversity buckets at once: BINANCE_CM (2.7% vs ≥5% target), cross-venue (7.0% vs ≥15% target: two distinct venue suffixes), and long-short/market-neutral (14.7% vs the long-only-dominated 85.3%).

coding_stuck_3_timeouts