AvaxCrossVenueFundingDifferentialCarryNeutral

Cross-Venue Funding-Differential Carry on Mid-Cap Alts — Market-Neutral Binance USD-M vs Hyperliquid, Collect the Persistent Funding Spread (AVAXUSDT.HYPERLIQUID + AVAXUSDT.BINANCE, 3-Parameter, Delta-Neutral)

A DELTA-NEUTRAL, CROSS-VENUE, market-neutral carry strategy that is a deliberately DIFFERENT FAMILY from the trend/momentum/breakout collision that keeps dying. It holds a long perp on one venue and an equal-notional short perp on the other venue for the SAME mid-cap alt (AVAX), and the edge is the ACCUMULATED FUNDING DIFFERENTIAL between the two venues — a mechanical cash flow — NOT price convergence. This is the key distinction from the spot-perp/inter-exchange BASIS-REVERSION family that recently died six times (those bet on a price dislocation closing; this one is price-neutral and simply harvests the funding spread while it persists). It targets the two most under-represented buckets in the portfolio simultaneously: HYPERLIQUID (6.4% vs 20% target) and cross-venue (7.1% vs 15% target), and it is market-neutral, which pushes against the over-weight long-only direction (86.5%). Crucially it is deployed on a MID-CAP alt rather than BTC/ETH: cross-venue funding dispersion on majors is arbitraged down to near-zero (the exact 'funding compressed on liquid Binance names' lesson that killed the major-pair carries), whereas on thinner alts like AVAX the two venues' distinct participant bases and thinner cross-venue arb capital leave WIDER, more persistent funding differentials. Only 3 parameters: entry differential threshold, same-sign persistence count, exit-compression threshold. Because it is delta-neutral it inherits neither crypto's bull-market beta (avoiding the L48 long-only-drawdown risk_reject) nor the outlier-day dependence that flags OHLCV trend followers as overfit.

verifying_stuck_3_timeouts

Verification failed (Layer 4 — QA review) [class=code_defect]: - [edge_concern] Separate from the blocking sizing defect, the carry edge itself looks thin and the analyst should scrutinize it once re-sized: avg_trade_return_pct 0.0169% (≈zero net) over 28 trades, total_return −25.3%, PF 0.98, stark leg asymmetry (long_win_rate 0.214 vs short 0.786). The delta hedge works directionally (beta −0.14, benchmark_correlation −0.16), so the drawdown is residual cross-venue basis noise + costs, not market beta — but a USD-vs-USDT quote mismatch between the legs means the hedge isn't perfectly price-neutral and adds divergence PnL. Whether the AVAX cross-venue funding spread net of the two-leg round trip and realistic impact is positive is the analyst's full-backtest call — judged at a capacity-respecting size, not the tested one. - position_size sizes the HL leg at equity*per_leg_frac*leverage = equity*0.5*3.0 = 1.5x equity notional, and _submit_entry mirrors an equal-notional Binance leg → ~3x gross on a THIN mid-cap alt traded CROSS-VENUE. The sandbox shows this blows venue capacity: impact_cost_pct 153.9%, capacity_usd only $42,217, avg_position_pct 178%, per-trade PnL swings of ±$151k. Per the fail_class guidance, an impact/capacity limit blown by position_size() is a sizing bug (more extreme than the 60%-of-equity example: 150% per leg / 300% gross). At this size the result is dominated by unrealizable impact, so the backtest can't be trusted as a measure of the carry. Not an edge statement — the funding spread is collected on notional regardless of leverage, so the amplification only multiplies impact and divergence risk. - max_funding_lag_hours is 12h for BOTH legs, but Hyperliquid settles HOURLY. A 12h tolerance lets an HL rate up to 12 settlements old count as fresh — in a data gap the strategy could trade on a materially stale HL rate (the higher-frequency leg that drives most of the differential). 12h is fine for Binance's 8h cadence but too loose for HL.

AvaxCrossVenueFundingDifferentialCarryNeutral tried to escape the six prior basis-reversion deaths by harvesting a pure funding-differential cash flow rather than a price convergence, and by choosing mid-cap AVAX on the thesis that thinner cross-venue arb leaves wider, more persistent spreads than on arbitraged majors. With funding now credited by the engine, the backtest cleanly falsified that premise: the annualized differential exceeded the entry threshold only 15 times in 911 days, those trades were collectively net-losing (avg trade 0.0102%, PF 0.98, -23.2% total), and market impact of 159% exceeded the gross PnL against a ~$39k capacity. The analyst abandoned it at backtest review, noting the sufficient funding data ruled out a data bug and that loosening the threshold would only add thinner, costlier legs. It never reached optimization, analysis, or risk review.

Moving a cross-venue funding carry to a mid-cap alt to find wider spreads trades one dead end for another — the differential is neither wide nor frequent enough to clear costs, and the thin liquidity that supposedly leaves the spread open also makes market impact (159%) dwarf the tiny carry, so the edge is unrealizable at any meaningful scale.

The analyst abandoned it at backtest review: the premise of wide, persistent mid-cap funding differentials was falsified (only 15 entries in 911 days, collectively net-losing), and even where a differential existed the thin AVAX-Hyperliquid liquidity made impact cost (159%) exceed the gross PnL. Funding data was present and sufficient (12,106 events), so this was not a data bug — no parameter region makes a differential this small and infrequent positive net of costs.

A delta-neutral, cross-venue funding-differential carry strategy (3 parameters) that held a long AVAX perp on one venue and an equal-notional short on the other (Hyperliquid vs Binance USD-M), harvesting the accumulated funding spread while it persisted — deployed on a mid-cap alt on the thesis that thinner cross-venue arb leaves wider, more persistent funding differentials than on majors.

The credited-funding backtest (AVAX on Hyperliquid vs Binance, 4H, 911 data days, 2024-02 to 2026-08) was a net loser with almost no edge: the differential cleared the entry threshold only 15 times, giving avg_trade_return_pct 0.0102% (vs the ~0.15% floor), profit factor 0.98, expectancy -487, total return -23.2%, and negative every year. Capacity was catastrophic — impact_cost_pct 159%, ~$39k capacity, avg_position_pct 165% — so impact cost exceeded gross PnL.

Verification failed (Layer 2 — synthetic scenarios): Parameters used: ['smooth_bars', 'min_notional', 'per_leg_frac', 'persist_bars', 'basis_stop_pct', 'exit_diff_annual', 'entry_diff_annual', 'max_funding_lag_hours'] Check that __init__ sets all attributes from self.parameters.get(). - steady_uptrend: TypeError: AvaxCrossVenueFundingDifferentialCarryNeutral._ts_of() takes 1 positional argument but 2 were given (bar timestamp: 1735689600000) - steady_downtrend: TypeError: AvaxCrossVenueFundingDifferentialCarryNeutral._ts_of() takes 1 positional argument but 2 were given (bar timestamp: 1735689600000) - flat_ranging: TypeError: AvaxCrossVenueFundingDifferentialCarryNeutral._ts_of() takes 1 positional argument but 2 were given (bar timestamp: 1735689600000) - volatility_spike: TypeError: AvaxCrossVenueFundingDifferentialCarryNeutral._ts_of() takes 1 positional argument but 2 were given (bar timestamp: 1735689600000) - zero_volume: TypeError: AvaxCrossVenueFundingDifferentialCarryNeutral._ts_of() takes 1 positional argument but 2 were given (bar timestamp: 1735689600000) - price_gap: TypeError: AvaxCrossVenueFundingDifferentialCarryNeutral._ts_of() takes 1 positional argument but 2 were given (bar timestamp: 1735689600000)