Strategies

TrendDayLeveragedEtfUseq ported QuantRocket's afternoon trend-day continuation rule to three leveraged/volatility ETFs on USEQ hourly bars. It was adapted to a 14:30 ET checkpoint because the Yahoo hourly grid has no 14:00 close. Both iterations failed the Layer 3 sandbox backtest by exceeding the 300s limit on an 833-day TQQQ window. The developer attributed the timeout to catalog file discovery under heavy host I/O load, not to strategy logic. The run was abandoned in a verification loop without reaching backtest, optimization, analyst or risk review, so its edge was never measured.

SkippedMonthSectorMomentum tested whether skipped-month formation momentum across eight US sector ETFs predicts the next 21 sessions, using a top-2 long / bottom-2 short book. Across 3 iterations the code never cleared Layer 4 QA review. The rotation logic sized new legs while the old cohort's closes were still in flight and depended on exact-timestamp re-entry, so cohorts were partial and unbalanced. The one sandbox measurement (Sharpe -0.02, PF 0.98, -0.125% per position over 178 positions) was judged not representative of the hypothesis. The run was abandoned before backtest review, optimization, analyst or risk stages.

This strategy tested whether an FMZ-style one-sided recycling grid on LTC spot could clear fees, using a 1.8% take-profit on a 0.9% grid against a 0.20% spot round trip. Over 3 iterations it never got past verification. QA found that stop_requested and the cycle bookkeeping were reset only when a new cycle started, so after any losing liquidation the 24h halt kept moving forward and the grid shut down for good. The sandbox matched this, with only 10 fills and 4 equity-moving days in 834 days. QA also flagged that the cycle trailing loss would trigger before the node stops and that per-node accounting conflicted with the single netted spot position. The strategy was abandoned without a full backtest, optimization or analyst review.

HyperliquidLtcBchCompressionBreakout was built to trade continuation in the LTC/BCH ratio after volatility compression, buying the winner and shorting the loser in equal dollar amounts on Hyperliquid perpetuals. Across 3 iterations, it never passed the Layer 3 sandbox backtest: the smoke test kept exceeding the 300-second limit on about 20,000 4-hour bars. The verification loop then abandoned the run with a sandbox_timeout, so the idea's economics were never tested and no performance metrics exist.

DogeHlBinanceFundingGapCarry tried to profit from a persistent gap between DOGE perpetual funding on Hyperliquid and on Binance, holding opposite positions on the two venues so that price moves cancel out. The pre-study looked promising, but the code never got through verification. Over 3 iterations it kept failing the Layer 1.5 hypothesis/config consistency check, which flagged that the optimization plan fixes leverage at 2.0. It was abandoned in a verification loop with no backtest, optimization or analyst verdict, so the funding-gap idea itself was never tested.

The hypothesis proposed shorting turnover-qualified Bybit perp launches for 21 days with a BTC hedge, citing a gross pre-study edge of about +13% per event across 76 listings. The developer coded BybitNewListingSupplyOverhangShortHedged and revised it three times for performance: funding prefix-sum caches with bisect, keeping only two bars per instrument, and skipping completed listings early. Every version still went over the 300s Layer 3 sandbox limit, so the run was abandoned as a verification-loop sandbox timeout. No backtest, optimization, analyst or risk-review stage was reached, so the premise is still untested in the pipeline.

The hypothesis proposed a new mechanism family for the corpus: a long-short price-volume correlation cross-section across 14 OKX perps, with 1.0x gross exposure, weekly rebalances and 25% per-leg catastrophe stops. The developer used rolling deques and indexed order queries to keep the work bounded, but across 3 iterations the code never got past Layer 3 sandbox verification, where it went over the 300s limit on a 2,400-day daily window. The run was abandoned for a sandbox timeout before any backtest, backtest review, optimization, analyst or risk review stage ran, so the factor's edge was never measured.

The hypothesis came from price-and-volume anomaly detection research: altcoin surges on Hyperliquid that are followed by fading volume and a pullback should reverse, traded as an independent short basket on INJ, NEAR, OP and SEI. The strategy was coded with per-leg rolling buffers for returns, volume baselines and ATR, plus gap-reset logic. Over 3 iterations it never passed Layer 3 verification, because the sandbox smoke test kept going over the 300s limit on a 2400-day, ~20,000-bar window. It was abandoned with no backtest metrics, so whether the surge-exhaustion edge exists on these perpetuals is still untested.

The hypothesis applied Amihud's illiquidity measure to four Hyperliquid DeFi perpetuals, trading one equal-dollar pair at a time: long the token whose liquidity was recovering, short the one whose liquidity was deteriorating. Over 3 iterations the developer moved the code to incremental rolling sums and bounded sorted windows. The sandbox smoke test still went over the 300s limit on a ~20,000-bar window. The run was abandoned at verification, so no backtest metrics were produced and none of the later stages ran.

OkxFiftyTwoWeekHighAnchoringSpread tested whether anchoring to the 52-week high produces a cross-sectional edge. Each Monday it ranked 15 OKX perpetual swaps by their proximity to the trailing high and held equal-dollar long and short legs. The hypothesis itself flagged sandbox-timeout risk, and the code used bounded deques with monotonic-queue maxima. Even so, after 6 iterations the Layer 3 sandbox smoke test still went over the 300s limit on a 2400-day daily window. The run was abandoned in verification, before any backtest, optimization, analyst or risk review took place.

SpotLedAltSelloffReboundBtcHedged tested whether spot-led relative selloffs in 11 liquid alts rebound against BTC, using an equal-dollar BTC hedge and a 3-day hold. The gross pre-study looked promising (+2.51% mean over 142 events). The coded strategy never got through verification: across 3 iterations the Layer 3 sandbox backtest kept exceeding the 300s wall-clock limit. Even the last version, written with incremental O(coins) rolling moments, hit the limit, and the run was abandoned before any backtest, optimization or analyst review.

The hypothesis was to fade temporary return dispersion between AAVE and UNI on Hyperliquid with an equal-dollar long-laggard/short-leader pair on synchronized 4-hour bars, and catalog inspection found 5,879 shared timestamps. The strategy went through 7 iterations, and the latest code notes that construction was blocked by conflicting source funding observations, with the trading logic kept as-is pending a data/loader repair. The backtest then failed with a FundingObservationError for hyperliquid/AAVEUSD and the run was abandoned. No backtest metrics were produced, and the review, optimization, analysis and risk stages were not reached.

This hypothesis tested whether a three-day LTC-versus-BCH price divergence on low volume reverses. It traded synchronized 4-hour Hyperliquid perpetual bars, with catalog coverage from 2024-01-25 to 2026-10-02. The developer coded HyperliquidLtcBchLowParticipationReversal, a long-laggard/short-leader strategy with equal dollar notionals, but over 3 iterations it never got past verification. Each time the Layer 1.5 consistency check flagged a hypothesis mismatch because the optimization_plan fixes leverage at 1.0. The run was abandoned before any backtest, optimization, analyst or risk review.

The strategy tried to capture partial reversal of gold-miner moves that gold does not explain, trading GDX against a beta-hedged GLD leg on daily US equity bars. Over 4 iterations the code never passed verification: Layer 1.5 kept flagging a hypothesis/config mismatch because the optimization_plan's exit_z baseline was missing or outside its domain. The run was abandoned in that verification loop, so backtesting, optimization, analyst review and risk review were never reached and no performance metrics exist.

HlXlmAlgoFailedRangeBreak tested whether failed range breakouts on Hyperliquid XLM and ALGO 4-hour perpetuals revert toward the range interior. Over 3 iterations the code never passed verification. The final iteration mainly reworked range tracking for speed after a sandbox timeout. The Layer 4 QA review then found two critical logic defects: confirmation ignored the opposite range boundary, and it accepted reclaims outside the fail window. QA also noted that the sandbox's 31 trades averaged only 0.0435% net. The run was abandoned as a code defect before backtest review, optimization, analyst review or risk review.

AtomPersistentCrossVenueFundingCarry tested whether a lasting funding-rate gap between Hyperliquid and Bybit ATOM perpetuals could be earned with a dollar-neutral long/short pair, using a basis veto to block bad entries. The code indexed settled funding events with prefix sums and bisect lookups. Even so, after 3 iterations the strategy still failed Layer 3 sandbox verification by going over the 300s time limit on a 2,400-day 4-hour window. It was abandoned with no backtest metrics, so the funding-carry hypothesis was never measured.

The hypothesis proposed a beta-hedged, sector-relative residual reversal on 20 US large caps across four SPDR sector ETFs, and its own pre-study reported a modest positive edge per event. The developer built an O(1) rolling-statistics implementation and revised it over 3 iterations. Iteration 3 was meant to set the config to the plan's fixed 2.0 leverage, but verification still failed the Layer 1.5 hypothesis/config consistency check on that leverage requirement. The run was abandoned as a verification loop, so backtesting, optimization, analyst and risk stages were never reached.

SandNftSectorLeadLagCatchUp tested whether SAND moves late after GALA and APE make strong moves together, using synchronized 4-hour Hyperliquid perpetual bars with GALA and APE as signal sources only. It went through 3 iterations without producing a backtest. One iteration timed out in the sandbox while refilling 401 missing GALA bars, before any strategy bars were evaluated, and the third changed only the documentation. The run was abandoned in the verification loop with a Layer 1.5 hypothesis/config mismatch (the optimization_plan fixes leverage at 1.0), so backtest review, optimization, analysis and risk review were never reached.

HyperliquidDefiStressDispersion was built to test whether relative resilience during a joint DeFi selloff predicts further AAVE-vs-UNI divergence, trading the pair long/short on 4-hour Hyperliquid bars. Across 3 iterations the code never passed verification: each time, the Layer 3 sandbox backtest exceeded its 300s limit on a ~20,000-bar window. The run was abandoned in the verification loop, so the backtest review, optimization, analyst and risk stages were never reached. No performance metrics exist, and the hypothesis is still untested.

This strategy faded confirmed LTC/BCH relative shocks on Hyperliquid 4-hour bars with equal-dollar, opposite perpetual positions. The code was written, but over 3 iterations it never got through verification. Each time, the Layer 1.5 hypothesis/config consistency check flagged that the optimization_plan fixes leverage at 1.0. The run was abandoned in the verification loop, so the backtest, optimization, analyst and risk stages never ran and no performance metrics exist.

This strategy aimed to short ETH perpetuals on Hyperliquid while funding stayed positive, using open-interest growth and a non-negative premium as confirmation. Over 3 iterations it never passed code verification: the Layer 3 sandbox backtest ran past the 300s limit on 833 days of hourly ETH data. Backtest review, optimization, analyst review and risk review were never reached, so the edge was never tested. The run ended as abandoned because of a performance problem in the code, with no verdict on the trading idea itself.

DogeBybitOkxOuDislocation tested a QuantConnect-style Ornstein-Uhlenbeck pairs mechanism on DOGE perpetuals quoted on Bybit and OKX. After 5 iterations it produced clean, fully aligned data and accounting, but over about five years only 15 pairs fired, for a 0.23 Sharpe whose confidence interval spans zero and a 0.0277% average per-trade return. Impact and commissions used up almost all of the gross profit. The analyst abandoned it at backtest review because no parameter tuning could make spreads this small cover the cost of the trades, so optimization, analysis and risk review never ran.

This strategy tested whether 12-month relative momentum between large caps (SPY) and small caps (IWM) gives a tradable monthly spread. The six style ETFs named in the source were not in the USEQ universe, and the hypothesis noted that earlier two-leg USEQ runs had timed out. All three iterations failed at Layer 3 verification by exceeding the 300s sandbox limit, so the run was abandoned before any backtest, optimization or analyst verdict. The strategy's docstring says a 2026-10-02 measurement put almost all of the runtime in ParquetDataCatalog file discovery rather than in the O(1) per-bar strategy loop, which suggests the problem is in the infrastructure, not the strategy code.

The hypothesis claimed that volume-confirmed breakouts on liquid OKX alt perps keep moving in the breakout direction after an initial absorption retrace, and that the gain peaks around 18h. The initial backtest showed real per-trade economics: 0.394% per trade on 1,880 trades and a +227% total return. However, the edge was fragile: Sharpe was 0.79, max drawdown 37.7% against a 25% pre-registered limit, PnL came from the right tail and DOGE longs, and the calm and stressed volatility regimes lost money. The analyst sent it to optimization anyway, where nested walk-forward validation found no parameter candidate that passed in window 5, and the strategy was abandoned.