rhino-strategy

🦏 RHINO v1.0 — Tail-Risk / Crisis-Alpha Hedge Fund

Rhino carries cheap convexity: it bleeds a little in calm and pays big in shocks — geopolitics, macro stress, liquidation cascades. The mirror lesson of the 2026 oil war: while everything else bled, the winners were long oil and gold. Rhino productizes that as an always-on insurance book plus a dormant convex add. One producer script (rhino-producer.py) serves both books; the RHINO_LEG env var selects which.

Book	Style	Gated by	Wallet env	Runtime	Scanner
hedge	Small always-on long defensives	nothing (always on)	RHINO_HEDGE_WALLET	runtime-hedge.yaml	rhino_hedge_signals
escalation	Long crisis + short risk, leveraged	the STRESS detector	RHINO_ESCALATION_WALLET	runtime-escalation.yaml	rhino_escalation_signals

Together: the hedge book means you already own the crisis trade when the shock hits; the escalation book adds leveraged convexity the moment stress confirms, then banks it before the violent reversal.

What makes Rhino different

• vs. Elephant (global macro): Elephant trades the macro trend in both directions for its own sake. Rhino is asymmetric by design — it is built to lose small in calm and win big in stress, not to track macro.
• vs. a Thesis Fund ("war escalation"): a Thesis Fund is a fixed manual bet you turn on. Rhino is always-on insurance that self-activates its convex leg on a measured stress signal, with no view required.
• It is the portfolio hedge of the fund lineup — the thing that's green on the days everything else is red.

The shared stress detector (the brain)

Each tick, Rhino tallies cross-asset stress probes — each fires on a 4h trend confirmation OR a 1h range break + ATR surge in the stress direction:

Probe	Fires (stress) when
Oil (xyz:BRENTOIL)	spiking up (4h BULLISH or breakout-up + surge)
Equities (xyz:XYZ100 → SP500)	breaking down (4h BEARISH or breakdown)
Gold (xyz:GOLD)	bid (4h BULLISH or breakout-up)
BTC	rolling over (4h BEARISH or breakdown)
Vol (BTC ATR)	recent ATR / baseline ATR ≥ volSurge (1.5)

STRESS is declared when the count clears stressThreshold (default 2). The escalation book trades only under STRESS; the hedge book ignores the gate (but reports the read for telemetry).

HEDGE book — the standing insurance (LONG only, always on)

Universe = the defensive whitelist (defensives: gold/oil/dollar/yen). Scores each for a clean uptrend (4h BULLISH backbone + 1h confirm + 24h momentum + RSI room; minScore 5) and carries it LONG — no falling-knife hedges (a defensive must be trending up to be added). Sized small (margin_pct 10%): the position size is the cost control, so calm-time bleed is bounded while a crisis winner is free to run on a wide DSL.

ESCALATION book — the convex add (both directions, stress-gated)

Dormant in calm (cash = dry powder). When STRESS confirms, it goes LONG the spiking crisis complex (crisisLongs: gold/silver/oil/CL/natgas/dollar/yen) and SHORT the cratering risk complex (riskAssets: BTC/ETH/SOL/HYPE/SUI + growth indices), each scored in its mandated direction (minScore 5). Sized larger (margin_pct 22%) — this is the payoff leg.

Execution & exit

• both books: slots 3, strict 5x clamp, tick 300s
• HEDGE margin_pct 10% — DSL wide let-it-run (hold the insurance): phase1 max_loss 16%, all time-cuts OFF, hard_timeout 10d; phase2 12%→0 / 25%→45 / 50%→65 / 90%→80 / 140%→90
• ESCALATION margin_pct 22% — DSL moderate-tight (crises reverse violently, bank the spike): phase1 max_loss 12%, weak_peak_cut ON (3h @ 2.0), dead_weight_cut ON (6h), hard_timeout 2d; phase2 8%→35 / 18%→60 / 35%→75 / 60%→88

XYZ handling

Every defensive + most stress probes are XYZ — candle fetches route with dex="xyz" (the producer keys off the xyz: prefix). XYZ trades 24/7 (no market-hours gating), so the hedge carries and the stress detector watches through weekends. The main/xyz clearinghouse sections are two VIEWS of ONE cross-margined wallet, so get_positions() takes accountValue ONCE via max() — never sums.

Risk gates (risk.guard_rails)

Gate	hedge	escalation
daily_loss_limit_pct	8	14
max_entries_per_day	3	8
max_consecutive_losses	4	5
cooldown_minutes	120	45
drawdown_halt_pct	20	22
per_asset_cooldown_minutes	360	120
data_retention_hours	168	96
drawdown_reset_on_day_rollover	true	true

Entries and exits both use FEE_OPTIMIZED_LIMIT (ensure_execution_as_taker true).

Files

File	Purpose
runtime-hedge.yaml	Hedge-book runtime spec (small, wide let-it-run DSL)
runtime-escalation.yaml	Escalation-book runtime spec (larger, moderate-tight DSL)
scripts/rhino-producer.py	Book-aware producer daemon (shared stress detector + both books)
scripts/rhino_config.py	Leg resolution + SenpiClient wrapper + helpers
config/rhino-hedge-config.json	Hedge-book tunables (defensive whitelist)
config/rhino-escalation-config.json	Escalation-book tunables (crisis longs / risk shorts, stress params)

Operator install

See README.md — the two books are two daemons (RHINO_LEG=hedge and RHINO_LEG=escalation) on two wallets (default 50/50 funding — the hedge runs small, the escalation holds dry powder), each with its own runtime YAML.

Hard rule for user-conversation Claude sessions

User-conversation Claude sessions MUST NOT call any of: create_position, close_position, edit_position, ratchet_stop_add, ratchet_stop_edit, ratchet_stop_delete, cancel_order, strategy_close, strategy_close_positions.

These tools are reserved for the producer daemon (entry path) and the DSL ratchet engine (exit path). User-conversation sessions are read-only. Each producer daemon handles real signals on its next tick.

License

Apache-2.0 — Copyright 2026 Senpi (https://senpi.ai)