Strategies
OKX Perpetual–Quarterly Basis Reversion
Outcome: Abandoned
OkxPerpetualQuarterlyBasisReversion
Outcome Summary
This hypothesis tried to capture mean reversion in the annualized basis between the OKX BTC perpetual and the quarterly future, trading both legs in equal notional and only on volatility-gated extremes. The strategy was coded over 2 iterations. It was abandoned before any backtest could run because no bar data could be collected for the quarterly leg (BTC-USDT-261225.OKX, 4-hour) after 5 retries. Backtest review, optimization, analyst review and risk review were never reached, and the pipeline noted the hypothesis should be re-evaluated once that symbol is backfilled.
Hypothesis
A low-frequency calendar-spread strategy adapted from the source’s near-versus-distant futures hedge. Trade equal-notional opposite positions in the OKX BTC perpetual and currently listed quarterly contract. Enter only on large, volatility-gated annualized basis dislocations; no averaging or martingale adds.
Updated both legs to the exact instruments declared by the hypothesis: BTC-USDT-SWAP.OKX and BTC-USDT-261225.OKX. The expiry parser now supports the hyphenated OKX quarterly identifier; all signal, sizing, and pair-management logic is preserved.
data_unavailable: backtest data for BTC-USDT-SWAP.OKX could not be collected after 5 retries: - bars (bar_type=BTC-USDT-261225.OKX-4-HOUR-LAST-EXTERNAL): got 0, needed 100, or 30+ days span The data layer (not the strategy) is the blocker — re-evaluate this hypothesis once collection is fixed / the symbol is backfilled.
Implementation
Equal-notional BTC OKX perpetual and quarterly calendar-spread basis reversion, entering only on large volatility-gated annualized basis dislocations and exiting on convergence, expiry proximity, or a holding-time limit.
Iteration History
Verification failed (Layer 1.5 — hypothesis/config consistency) [class=hypothesis_mismatch]: - config.instrument_id is BTCUSDT.OKX but the hypothesis declares BTC-USDT-SWAP.OKX, BTC-USDT-261225.OKX — a different asset. Implement the hypothesis as written (or the Research Lead must revise it).
Outcome Summary
Before coding a multi-leg OKX dated-futures strategy, confirm that the catalog holds enough history for the specific quarterly contract (at least 100 bars or 30+ days at the chosen timeframe).
The run was abandoned for data_unavailable. After 5 retries, 4-hour bars for the quarterly leg BTC-USDT-261225.OKX returned 0 bars, short of the 100 bars or 30+ day span required, so the backtest never ran; the pipeline flagged the data layer, not the strategy, as the blocker.
An equal-notional calendar spread between the OKX BTC perpetual and the listed quarterly contract, opened only on large, volatility-gated (z-score) dislocations in the annualized basis and closed when the basis reverted, with no averaging or martingale adds.
No performance was measured: the backtest report is empty and there is no optimization report, so no return, Sharpe or trade count exists.
Analysis
带你学习硬核策略 之 「C++ 版 OKEX合约对冲策略」
Backtest and paper results are hypothetical. Trading involves risk of loss.