By name: salamander-strategy description: >- SALAMANDER v1.0.0 — Pullback catcher in established 4h trends. LONG when 4h trend is BULLISH AND 1h price pulled back 3-7% from recent high AND Smart Money is > 55% long. SHORT when 4h trend is BEARISH AND 1h price rallied 3-7% from recent low AND SM is > 55% short. Universe: BTC, ETH, SOL. Wide DSL throughout — Phase 1 10% max_loss (pullbacks need room to develop) AND Phase 2 "let winners run" ladder (T0 +10% / lock 0); the resumed trend can run far. license: MIT metadata: author: jason-goldberg version: "1.0.0" platform: senpi exchange: hyperliquid requires: - senpi-trading-runtime>=1.1.0 - senpi_runtime_helpers
🦎 SALAMANDER v1.0.0 — Pullback Catcher
Buy dips in uptrends, short rallies in downtrends. Salamander only acts when the 4h trend is already established AND there's been a 3-7% counter-move on 1h timeframe AND Smart Money agrees with the original trend direction.
Why this strategy exists
Pullbacks in established trends are one of the highest-edge setups in trading — the trend has already proved itself, the counter-move provides a better entry, and Smart Money confirmation filters out trend-break risk.
The 3-7% band is the sweet spot:
- < 3% pullback: noise, not a real entry
- 3-7% pullback: tradeable counter-move within trend
7% pullback: the trend may be breaking; don't catch a falling knife
Salamander requires ALL three:
- 4h trend BULLISH (or BEARISH)
- 1h pullback in the 3-7% band
- Smart Money agreeing with the 4h trend direction (> 55% tilt)
If any one fails, no entry.
CRITICAL RULES
RULE 1: 4h trend gate, not score
The 4h trend must be NON-NEUTRAL. Salamander does not trade chop. If higher-lows-vs-lower-highs is too close to call, the producer outputs WAITING.
RULE 2: Pullback band is hard
The pullback must fall in [pullbackMinPct, pullbackMaxPct] (default 3-7%). Outside that band, no entry. Configurable per asset volatility (smaller bands for stable majors, wider for HYPE-tier).
RULE 3: Wide DSL throughout
- Phase 1 max_loss 10% (wider than Hawk's 8%) — pullbacks within a trend need a few percent of room to develop
- Phase 2 wide "let winners run" ladder (T0 +10% / lock 0 → T5 +100% / lock 85) — when the pullback resolves, the resumed trend can run far. The old tight T0 (+5% / lock 30) locked a floor after ~1-2% price and chopped the continuation out on the first wiggle (HYPE 40→60 post-mortem, 2026-05-21).
If you tighten Phase 1 below 10%, you'll cut legitimate pullback continuations as noise — and never re-tighten Phase 2, or you cap the very continuation Salamander exists to ride.
RULE 4: Producer enters. DSL exits.
No close_position call site. DSL Phase 1 + Phase 2 + hard_timeout 48h own all exits.
How Salamander scores a trade
Gates (all required):
- 4h trend non-neutral
- 1h pullback (vs prior 24h max for LONG, prior 24h min for SHORT) in [3%, 7%] band
- SM direction agrees with 4h trend
- SM tilt >= 55%
Score components (max ~9):
| Signal | Points |
|---|---|
| 4h trend aligned (gate-confirmed) | +3 |
| Pullback in 3-7% band (gate-confirmed) | +2 |
| Pullback in 4-6% midpoint sweet spot | +1 |
| SM aligned (gate-confirmed) | +2 |
| SM strongly tilted (>= 70%) | +1 |
Floor: minScore: 5. Typical entry clears 5-7.
DSL preset (wide throughout — Phase 1 roomy, Phase 2 let-winners-run)
| Phase | Component | Setting |
|---|---|---|
| Phase 1 | max_loss_pct | 10% (wider — pullbacks need room) |
| Phase 1 | retrace_threshold | 6 |
| Phase 1 | consecutive_breaches | 1 |
| Time cuts | hard_timeout | 48h (pullback resolution can take time) |
| Time cuts | weak_peak_cut | 90min / 3% min |
| Time cuts | dead_weight_cut | DISABLED |
| Phase 2 | T0 | +10% / lock 0 (let the continuation breathe) |
| Phase 2 | T1 | +20% / lock 25% |
| Phase 2 | T2 | +30% / lock 40% |
| Phase 2 | T3 | +50% / lock 60% |
| Phase 2 | T4 | +75% / lock 75% |
| Phase 2 | T5 | +100% / lock 85% |
Scanner pattern
This strategy uses the Multi-asset whitelist scanner with pullback-detection scoring — see senpi-trading-runtime/references/producer-patterns.md. Primary MCP calls: market_get_asset_data, leaderboard_get_markets.
Operator install
See README.md.
Changelog
v1.0.0 (2026-05-20) — initial release
Second in the technical-pattern pair (Hawk = breakouts, Salamander = pullbacks). Asymmetric DSL distinguishes from Hawk's tight breakout profile — pullback theses need wider Phase 1 floor.
License
MIT — Copyright 2026 Senpi (https://senpi.ai).
Skill Attribution
When creating a strategy, include skill_name and skill_version. See references/skill-attribution.md.