NEAR Hyperliquid Funding-Against-Price Squeeze with OKX Confirmation

Hypothesis

Derived by the Research Lead from Kraken Learn's funding-positioning discussion: test whether persistent funding opposed to an already-established price move identifies trapped positioning and subsequent directional continuation. Source: https://www.kraken.com/learn/futures-trading-funding-rate-strategy. Trade NEARUSD.HYPERLIQUID long or short using synchronized 4-HOUR closes; NEARUSDT.OKX supplies price confirmation only and is never traded. This is a positioning-conditioned directional experiment, not funding arbitrage. Market entries execute at the signal bar's close, the registry's faithful equivalent to next-bar-open execution on continuous markets. Intrabar protective stops are matched by the platform. Baselines: funding_window_hours=24, return_lookback_bars=6, return_atr_threshold=1.0, stop_atr_multiple=2.5, max_hold_bars=18. Parameters remain frozen after deployment.

Long-only research is the largest directional imbalance at 80% versus a 55% ceiling. This bidirectional proposal also adds Hyperliquid exposure, currently 10.8% versus a 20% target, and a cross-venue signal, currently 7.8% versus a 15% target. Options are further below their venue target, but this source describes perpetual positioning and does not justify an options contract. Corpus concentration was supplied empty, so NEAR's historical share cannot be established. This differs from ATOM extreme-funding reversal: it follows an existing price move against persistent funding without requiring a funding sign flip, an extreme absolute rate or a breakout. It differs from price-only momentum because persistent contrary funding is mandatory. No performance pre-study was used.

negative_expectancy: the Research Lead's NEAR 4-hour operationalization loses on a reliable sample, with average trade return -0.350% versus the +0.15% Hyperliquid floor and PF 0.656. Funding is credited and does not rescue the negative price-trading economics. No measured viable parameter region or implementation defect supports spending two hours optimizing. This rejects this operationalization, not Kraken's general funding-positioning discussion.

Implementation

Bidirectional NEAR Hyperliquid continuation conditioned on persistent contrary hourly funding. Enter long after a positive six-bar move of at least one ATR, confirmed by positive synchronized OKX returns, when at least 75% of observed funding hours are negative and their mean is negative; shorts are symmetric. Require 80% funding coverage and a latest observation no older than two hours. Risk 0.5% of primary-venue settled equity with a fixed intrabar stop 2.5 entry-time Wilder ATRs from the fill. Cap notional at 25% of equity and quantity at 0.1% of prior 20-bar average volume. Exit after 18 evaluated 4-hour bars or the protective stop, with a six-bar cooldown after closure. OKX is never traded.

Implements established price continuation against persistent funding rather than funding arbitrage or funding-sign reversal. Both confirmation endpoints must match Hyperliquid close timestamps exactly; future and stale OKX bars cannot contribute. Funding is read only within the trailing calendar-time window, with dual keys and hourly observations deduplicated. Missing funding blocks entries without a price-only fallback. The continuous signal is displacement in ATR units; funding and confirmation gates apply only in should_enter. Position anchors and counters are scalar state, and the supplied research parameter permissions are preserved. Verification is left to the pipeline. The requested outbox file could not be written because this session's filesystem is read-only; the complete task-bound JSON is returned directly.

Verification Results

The sandbox reports negative dollar expectancy and PF 0.935 despite positive equal-weighted average trade return of 0.111% of notional. Assess sizing-weighted economics and the strong short-side concentration on the full backtest; these are advisory performance concerns.

Backtest Review

Sharpe

-0.94

Total return

-1.98%

Max drawdown

2.03%

Trades

197

Win rate

44.7%

Profit factor

0.66

197 closed trades over 969 days provide a reasonable initial sample; synchronized price inputs passed their audits.

Code implements continuation against persistent funding with exact-timestamp OKX confirmation; inspected trades use only the Hyperliquid instrument.

Net average trade return is -0.350% versus the Hyperliquid fee floor of +0.15%; PF is 0.656, Sharpe -0.937 and total return -1.979%.

QA's concern is borne out: 193 shorts lost $4,017, while four longs earned $59. Short price PnL was negative even before funding, which credited $332.

All three volatility regimes lost money. The forecast diagnostic's positive IC is inconclusive (t=1.29), rather than evidence supporting a parameter search.

The verdict could not be saved to the required outbox because filesystem access is read-only.

Outcome Summary

This operationalization needs evidence of viable trading economics before parameter optimization; its failure does not reject the general funding-positioning thesis.

Backtest review recommended abandonment because negative expectancy fell below the +0.15% Hyperliquid per-trade floor, and funding credits did not offset price-trading losses. Optimization and subsequent analyst, risk-review, and deployment stages were not reached.

Trade NEAR continuation on Hyperliquid when persistent funding opposed the established price move, using synchronized OKX prices for confirmation only.

The backtest recorded 197 closed trades over 969 days, returning -1.979% with Sharpe -0.937 and profit factor 0.656. Average trade return was -0.350%; 193 shorts dominated the sample, and all three volatility regimes lost money.