Spot-Led Alt Selloff Rebound, BTC-Hedged: buy a Binance USD-M alt perp vs short BTC perp after a >=1.5-sigma 3-day BTC-relative drop executed mainly on BINANCE_SPOT (spot volume share z>1), hold 3 days (11 alts, 1-DAY)

Hypothesis

Fills the most under-represented buckets: cross-venue (7.3% vs a >=15% target) and long-short (16.7%). The mechanism is new to the corpus: the venue composition of volume (spot vs perp) as the conditioning variable, rather than funding, OI, taker flow or premium gates (L111), with trades sized for >2% gross moves. It avoids BTC as a traded thesis: BTC is only the hedge leg, which the beta-neutral design genuinely requires. The trigger is a market observation, not a calendar date (L179). It is not an options structure (L102). Every leg's exact bar series was verified in the local catalog with overlapping history (L186/L181). It differs from the abandoned 'Perp-Led Selloff Absorption Panel', which was long-only, perp-premium-gated and unhedged: this trades the OPPOSITE (spot-led) condition, which the pre-study shows is where the rebound lives. The spot-led RALLY continuation was tested and rejected (it worked only in 2021) and is deliberately excluded. Leverage 2.0 is required because two equal-dollar legs across up to 4 events reach ~96% gross per side.

verification_loop: sandbox_timeout: Verification failed (Layer 3 — sandbox backtest): smoke test exceeded the 300s wall-clock limit on a ~20000-bar window (2400 days of BTCUSDT.BINANCE-1-DAY-LAST-EXTERNAL). This almost always means per-bar work that scales with history — e.g. rescanning the full funding/supplementary series, or rebuilding a list and calling min()/sorted() inside calculate_signal()/on_bar() on every bar. Precompute sorted timestamp arrays ONCE in __init__ and use bisect, or cache lookups keyed by timestamp, so per-bar cost is O(log n) not O(n).

Implementation

Buys spot-led three-day altcoin relative selloffs on Binance USD-M and shorts an equal-dollar BTC hedge, holding each event for three days or until its pair stop triggers.

Addresses the Layer 1.5 failure by declaring leverage=2.0 in both the engine config and parameters checked against the fixed research plan. The full optimization plan is preserved. The only source change replaces the redundant leverage fallback with direct access to self.config.leverage; existing sizing already consumes that leverage. Signal, entry, exit and hedge logic remain unchanged. At four events, exposure reaches 96% of equity per side, 192% gross. All eleven spot signal legs are included as specified in the hypothesis description, although its market_context instrument list omits the UNI, AAVE and INJ spot legs. Equal-dollar hedging does not estimate individual altcoin betas. No verification was run and no final-exam data was read. The bound outbox file could not be written because this session permits filesystem reads only; the complete JSON is returned here.

Verification Results

sandbox_timeout: Verification failed (Layer 3 — sandbox backtest): smoke test exceeded the 300s wall-clock limit on a ~20000-bar window (2400 days of BTCUSDT.BINANCE-1-DAY-LAST-EXTERNAL). This almost always means per-bar work that scales with history — e.g. rescanning the full funding/supplementary series, or rebuilding a list and calling min()/sorted() inside calculate_signal()/on_bar() on every bar. Precompute sorted timestamp arrays ONCE in __init__ and use bisect, or cache lookups keyed by timestamp, so per-bar cost is O(log n) not O(n).

Iteration History

Verification failed (Layer 1.5 — hypothesis/config consistency) [class=hypothesis_mismatch]: - optimization_plan fixes leverage at 2.0

Outcome Summary

A 23-instrument multi-leg strategy has to fit the sandbox's 300s smoke-test budget from the first iteration, so profile its per-bar cost (including extra-bar handling and the size of the multi-instrument data load) before resubmitting, not just its signal logic.

Layer 3 verification (the sandbox backtest) failed after 3 iterations. The smoke test ran past the 300s wall-clock limit on a ~20,000-bar window (2400 days of BTCUSDT.BINANCE 1-DAY), so the run was abandoned in a verification loop. Backtest review, analyst and risk stages were never reached.

Buy a Binance USD-M alt perp and short an equal dollar amount of BTC perp for 3 days after a >=1.5-sigma 3-day BTC-relative selloff in one of 11 alts that ran mainly through the spot book (spot volume share z>1). The thesis was that spot-led selling is non-informational inventory dumping that later rebounds.

No pipeline backtest or optimization was produced. The only figures come from the hypothesis's own gross, pre-cost pre-study: 142 events with a mean 3-day BTC-relative return of +2.51% (overlap-adjusted t 2.08), positive in every year from 2021 to 2026.