PAPER · APRIL 2026 HYPERLIQUID VERSION 3.3.2

Pick the fastest horse.
Hold the structural ones.

A long-only multi-asset strategy combining two decoupled return sources: a weekly-rebalanced equity rotation engine that ranks eleven US mega-caps by long-horizon upside conviction, plus a permanent crypto core in BTC, SOL, and HYPE with liquidation-aware dollar-cost averaging. One portfolio, two sleeves, one execution venue.

+8.8%
26-day paper return
82%
Paper win rate · 41W / 9L
14
Universe (11 stocks + 3 crypto)
10%
Vault profit share at launch
View Paper Dashboard → Read the Whitepaper
New to Hyperliquid? Sign up via our referral ↗ — Vault deployment target: end of paper validation window.
Status: paper trading on Railway since April 17, 2026 · Vault TBD

Two decoupled engines.
One discipline.

Most multi-asset strategies suffer from one of two problems: either they pick one return source and live or die by its regime, or they spread across so many that no individual edge survives the cost of diversification. Rotation takes a third path. It runs a momentum-and-conviction equity engine (the rotation, 70% allocation) alongside a structurally different mean-reversion crypto engine (the core, 30%). The equity side captures trending winners through hysteresis-gated weekly rebalancing. The crypto side captures dips through eight-level liquidation-anchored DCA grids. Both sleeves run on the same Hyperliquid venue, the same execution adapter, the same kill switches — but their entry and exit logic are independent. The result: when one is in drawdown, the other is statistically likely to be performing.

Two sleeves, decoupled

Strategy A — Fastest Horse Equity Rotation · 70%

25% permanent TSLA core reflects a long-horizon thesis not subject to weekly rotation. 45% rotation budget divides into five slots of 9% each, filled weekly from an eleven-name universe: NVDA, MSFT, GOOGL, AMZN, META, MSTR, PLTR, AMD, AVGO, ORCL, ASML. The Fastest Horse composite score combines visionary upside conviction (50 points), Wall Street consensus targets (25 points), and technical momentum (25 points). An 8-point hysteresis band prevents weekly churn. Five-level ATR-based DCA grid per position. Take-profits at +8%, +15%, +25%. Post-TP3 trail with 3% / 6% / 10% pullback rungs.

Strategy B — Crypto Core · 30%

BTC 11%, SOL 9%, HYPE 10%. BTC anchors with institutional-grade exposure. SOL adds high-beta L1 exposure. HYPE captures the exchange-token thesis on the venue Soomario uses. ETH intentionally excluded. Each core runs an eight-level liquidation-anchored DCA grid with a 3.5% minimum-spacing floor enforced between rungs — eliminating the v3.2 failure mode where all levels could compress into a single dip. Take-profits at +12%, +25%, +50%. Same post-TP3 trail logic. Hard stop 8% below L8 — typically ~−54% from entry, intentionally wide.

Shared infrastructure

Both sleeves use the same Hyperliquid integration (USDC perpetuals for crypto, HIP-3 xyz prefix perpetuals for equities), the same paper-or-live execution adapter, the same SQLite mirror of position state, and the same Flask-based dashboard. Operational controls: two-source kill switch, five-minute position reconciliation against Hyperliquid, Telegram alerts in three severity tiers, idempotent client order IDs that prevent retry-after-network-blip double-fills, per-coin size-drift sanity check on every reduce or close.

Why the short sleeve was removed

The data was clear.

An earlier version of Rotation included a third sleeve — an unlock-driven crypto short engine modelled on the same liquidation-zone thesis that drives Soomario's mean-reversion vaults. In live paper trading over a six-day window in early May 2026, that sleeve lost approximately $186 against the equity sleeve's contemporaneous gains. The pattern was not a single bad week of variance; the short thesis was structurally impaired in the current regime, where unlock-driven cascades had been clean follow-through rather than mean-reversion.

The fix was structural.

v3.3 removed the short sleeve entirely and redirected its allocation toward the long-only sleeves. In the 26 paper-trading days since the redesign, the strategy has produced +8.8% on a $5,000 starting capital. The 24 paper days that preceded the redesign (running v3.2 with shorts) produced approximately +1.1% in the same kind of market. The redesign accounts for the bulk of the improvement.

The crypto DCA was also redesigned.

v3.2 used a 5-level liquidation-derived DCA grid for crypto cores. On May 7, 2026, HYPE filled all five levels on a single intraday move from approximately $41 to $37, exhausting the deployment budget within hours. v3.3 redesigned this to 8 levels with a 3.5% minimum-spacing floor — guaranteeing the grid spans at least ~22% from L1 to L8 regardless of input clustering. Monte Carlo on the deep-crash regime across 50 seeds showed +71.95 percentage-point improvement in mean return versus v3.2.

26 days on Railway

Continuous paper trading from April 17 through May 12, 2026. The data below is paper, not live — slippage, funding rates, fees, and execution-quality differences between paper and live will reduce live returns when the vault is deployed.

RETURN
+8.8%
+$439.81 on $5,000 starting capital
CLOSED TRADES
41 W / 9 L
82% win rate · 9 open positions
CAPITAL DEPLOYED
23.6%
~$1,285 working · rest in DCA reserve
Why capital deployment stays low by design

The 23.6% deployment is structural, not opportunistic. Both sleeves' DCA grids seed only the L1 position at open and reserve the remainder for fills on dips. In benign markets, the strategy is intentionally under-deployed; in correction markets, it deploys into the dip; in crash markets it reaches near-full deployment at the trough. The reserve is the strategy's edge — it's what makes the DCA spacing-floor work. A fully-deployed range-farmer would not have meaningful dry powder to deploy when the deeper levels actually become reachable.

What Rotation isn't

Not yet live

Rotation is in paper-trading validation. The vault has not been deployed. The +8.8% return is calendar return on paper data — it annualises to a number that would be impressive if literal, but live trading involves transaction costs, slippage, funding rates, exchange downtime, and order rejections that paper trading does not. Live returns will be lower, often materially so.

Backtests use synthetic price paths

The deep-crash Monte Carlo that motivated the v3.3 zone redesign used synthetic price paths parameterised to match observed asset characteristics. Real Hyperliquid historical OHLCV data is not retrievable from the build environment. The strategy mechanics tested are identical to production, but the inputs are synthetic. The +71.95pp v3.2-vs-v3.3 improvement is robust to seed but tested on synthetic geometry, not literal historical replay.

Conviction is dominant in the equity rotation

The Fastest Horse composite overweights long-horizon visionary upside (50 of 100 points) over technical momentum (25) and consensus (25). This is deliberate but it means the strategy's equity sleeve is structurally biased toward names that visionary-analyst sources favour. When those analysts are right, the rotation works. When they are wrong, the rotation captures the wrong narrative. Visionary scores refresh quarterly — the strategy is not nimble to regime shifts within a quarter.

Leveraged perpetuals carry liquidation risk

Stock sleeves run at 5× leverage; crypto cores at 3×. The wide stop-loss bands (~−13% from entry on stocks, ~−54% on crypto cores) reduce the probability of hitting hard stops in normal conditions, but a coordinated decline across the basket — particularly in a regime of high correlation — could still produce material drawdowns. Full deposit is at risk of loss.

Paper now.
Vault when validated.

Paper-mode validation continues until the strategy has completed at least one full month of clean continuous operation. The dashboard refreshes every 15 seconds with current equity, open positions, recent trades, and Fastest Horse rankings. The Hyperliquid vault will be created once the milestone is reached.

View Paper Dashboard → Read the Whitepaper
Get ready for vault launch: Sign up for Hyperliquid via our referral ↗
Rotation will trade leveraged perpetual futures, which carries substantial risk of loss including the possibility of liquidation. Paper-trading results are not representative of live performance. Past performance does not guarantee future results.