TRB Binance–OKX Price-Gap Convergence After Initial Contraction

Hypothesis

Derived by the Research Lead from FMZ’s TRB arbitrage account: test equal-base-quantity long-cheap/short-rich perpetual positions after a large inter-exchange price gap first contracts. Source: https://www.fmz.com/digest-topic/10341. The local full article was read; linked strategy code was not read. Use synchronized 15-MINUTE bars and all available overlapping history. Execute market orders at signal-bar closes, the capability registry’s equivalent of next-bar-open execution on continuous markets. Funding is included in measured economics rather than assumed to supply the edge. The unmeasured frequency prediction is approximately 10 paired episodes annually; scarcity must be reported without weakening thresholds to manufacture trades.

Adds long/short and cross-venue exposure against the portfolio’s largest excess, long-only strategies, while increasing underrepresented OKX exposure. Binance remains overweight but is retained because the source specifically identifies its TRB divergence from OKX. Corpus concentration is unavailable. Snapshot prior art includes failed TRB funding-carry entries and a long-tail convergence experiment stopped by unavailable data. This tests price convergence with substantial fee headroom and initial contraction confirmation, rather than funding income or tightly coupled BTC spreads.

hold_expired: data_review:input_unavailable held 216 h, past its 216 h limit, unresolved (owner: operator). Not a verdict on the strategy; revivable once resolved. Recommended: approve when the venue's own funding history skips the settlement for every symbol checked; otherwise leave it and close the run

Implementation

Trades equal-base-quantity TRB Binance/OKX perpetual legs after one synchronized 15-minute contraction bar with the absolute midpoint-normalized price gap still at least 1.5%. Buys the cheap venue and shorts the rich venue. Closes both when the direction-adjusted gap reaches 0.2%, at 48 hours, or when either 10% intrabar leg stop fires. Market orders at signal-bar closes implement the registered continuous-market next-open equivalent. Uses all overlapping history supplied by the runner; funding cash flows and costs are measured by the engine.

Implements the Research Lead operationalization without importing the unread FMZ strategy. Gap percentages use 100*(OKX-Binance)/mean(OKX,Binance); positive gaps buy Binance and short OKX. The preceding gap must exceed the entry threshold, the next gap must contract without changing sign and still meet that threshold. Equal quantities are floored to the common venue quantity increment and constrained by 5% of each separately funded venue account, so a 10% stop risks at most 0.5% of either account before gaps and costs. Stops are platform-managed intrabar orders; closing one leg immediately requests closure of its hedge. A one-bar post-flat wait prevents same-bar reentry. Persistent episode state is scalar, pair timestamps must match exactly, and per-bar work is constant. No funding signal, frequency-driven threshold weakening, or additional filter is added. Roughly ten annual episodes remains an unmeasured prediction to report alongside observed scarcity.