Overview
Analyze the distribution of leveraged positions and their liquidation price levels to identify zones where forced selling/buying will accelerate price moves. Liquidation clusters act as "magnets" — price tends to be attracted toward large liquidation concentrations because market makers and whales profit from triggering cascading liquidations.
Core Concepts
1. Liquidation Mechanics
How liquidation works:
# Long position liquidation
long_liquidation_price = entry_price * (1 - 1/leverage + maintenance_margin)
# Short position liquidation
short_liquidation_price = entry_price * (1 + 1/leverage - maintenance_margin)
**Leverage and liquidation distance:**
| Leverage | Liquidation Distance (Long) | Liquidation Distance (Short) |
|----------|----------------------------|------------------------------|
| 2x | ~50% drop | ~50% rise |
| 5x | ~20% drop | ~20% rise |
| 10x | ~10% drop | ~10% rise |
| 20x | ~5% drop | ~5% rise |
| 50x | ~2% drop | ~2% rise |
| 100x | ~1% drop | ~1% rise |
### 2. Liquidation Heatmap Interpretation
A liquidation heatmap shows where liquidation orders are concentrated across different price levels, typically color-coded by density.
**Reading the heatmap:**
Price Level Long Liquidations Short Liquidations Interpretation $70,000 ░░░░░░░░░░ ████████████████ Heavy short liquidation zone $68,000 ░░░░ ██████████ Moderate short liquidation $66,000 ███████ ███████ Balanced (current price area) $64,000 ██████████ ░░░░ Moderate long liquidation $62,000 ████████████████ ░░░░░░░░░░ Heavy long liquidation zone
**Key principles:**
1. **Liquidation clusters are magnets**: price tends to gravitate toward large liquidation pools because the forced orders provide liquidity for whales to fill their positions
2. **Liquidation cascades**: when a cluster gets hit, the forced selling/buying pushes price further, potentially triggering the next cluster → cascade effect
3. **After liquidation wipe**: once a large cluster is liquidated, that price level often becomes support/resistance (overleveraged positions are gone)
### 3. Liquidation Level Identification
```python
def identify_liquidation_clusters(open_interest_by_price, leverage_distribution):
"""
Estimate where liquidation clusters exist based on
open interest and leverage distribution.
"""
clusters = []
for price_level in price_range:
# Long liquidations: positions opened above this level with high leverage
long_liq_volume = estimate_long_liq_at_price(
open_interest_by_price, leverage_distribution, price_level
)
# Short liquidations: positions opened below this level with high leverage
short_liq_volume = estimate_short_liq_at_price(
open_interest_by_price, leverage_distribution, price_level
)
total = long_liq_volume + short_liq_volume
if total > significance_threshold:
clusters.append({
"price": price_level,
"long_liq": long_liq_volume,
"short_liq": short_liq_volume,
"type": "long" if long_liq_volume > short_liq_volume else "short",
"magnitude": total,
})
return sorted(clusters, key=lambda x: x["magnitude"], reverse=True)
4. Liquidation-Based Trading Signals
Signal 1: Liquidation Magnet
def liquidation_magnet_signal(current_price, clusters):
"""
Price is likely to move toward the nearest large liquidation cluster.
"""
# Find nearest cluster above and below
above = [c for c in clusters if c["price"] > current_price]
below = [c for c in clusters if c["price"] < current_price]
nearest_above = min(above, key=lambda c: c["price"] - current_price) if above else None
nearest_below = min(below, key=lambda c: current_price - c["price"]) if below else None
if nearest_above and nearest_below:
above_magnitude = nearest_above["magnitude"]
below_magnitude = nearest_below["magnitude"]
if above_magnitude > below_magnitude * 2:
return "upward_magnet" # Larger cluster above → price likely moves up
elif below_magnitude > above_magnitude * 2:
return "downward_magnet" # Larger cluster below → price likely moves down
else:
return "balanced" # Both sides have similar clusters
Signal 2: Cascade Risk
def cascade_risk(current_price, clusters, direction="down"):
"""
Assess risk of liquidation cascade — multiple clusters stacked close together.
"""
if direction == "down":
relevant = sorted([c for c in clusters if c["price"] < current_price and c["type"] == "long"],
key=lambda c: c["price"], reverse=True)
else:
relevant = sorted([c for c in clusters if c["price"] > current_price and c["type"] == "short"],
key=lambda c: c["price"])
if len(relevant) < 2:
return "low_cascade_risk"
# Check if clusters are stacked within 5% of each other
gaps = []
for i in range(len(relevant) - 1):
gap = abs(relevant[i]["price"] - relevant[i+1]["price"]) / current_price * 100
gaps.append(gap)
if min(gaps) < 2:
return "high_cascade_risk" # Clusters stacked tightly → cascade likely
elif min(gaps) < 5:
return "moderate_cascade_risk"
else:
return "low_cascade_risk"
Signal 3: Post-Liquidation Support/Resistance
def post_liquidation_sr(price_history, liquidation_events):
"""
After a large liquidation event, that price level often becomes S/R.
"""
sr_levels = []
for event in liquidation_events:
if event.total_liquidated > 100_000_000: # >$100M liquidated
sr_levels.append({
"price": event.price_level,
"type": "support" if event.liquidation_type == "long" else "resistance",
"strength": event.total_liquidated,
"date": event.date,
})
return sr_levels
5. Liquidation Data Metrics
Key metrics to track:
| Metric | Description | Signal |
|---|---|---|
| 24h total liquidations | Total USD liquidated across all exchanges | > $500M = extreme, volatility spike |
| Long/Short liquidation ratio | Longs liquidated / Shorts liquidated | > 2 = longs squeezed, < 0.5 = shorts squeezed |
| Largest single liquidation | Biggest individual position liquidated | > $10M = whale liquidation |
| OI change post-liquidation | Open interest change after event | Large OI drop = leverage washed out (healthy) |
| Exchange-specific liquidation | Which exchange had most liquidations | Indicates where leverage is concentrated |
Liquidation volume interpretation:
| 24h Liquidations | Market State | Implication |
|---|---|---|
| > $1B | Extreme event | Major leverage wipeout, potential V-reversal |
| $500M - $1B | High volatility | Significant positioning reset |
| $200M - $500M | Elevated | Moderate leverage reduction |
| $50M - $200M | Normal | Background noise |
| < $50M | Calm | Low volatility, leverage building |
6. Liquidation Cascade Anatomy
Typical cascade sequence:
1. Initial trigger (macro event, whale selling, technical breakdown)
↓
2. Price hits first liquidation cluster ($65,000)
→ $200M in long liquidations forced to sell
↓
3. Forced selling pushes price to next cluster ($63,000)
→ $300M more in long liquidations
↓
4. Cascade accelerates → high-leverage positions ($62,000-$60,000)
→ $500M in rapid succession
↓
5. Eventually: open interest drops 20-30%, funding rate flips negative
→ Leverage is "washed out" → bottom forms
↓
6. Recovery begins (short-term) as no more forced sellers remain
Trading around cascades:
- Before cascade: reduce leverage, set wider stops, avoid high-leverage longs near heavy liquidation zones
- During cascade: do NOT try to catch the knife; wait for OI to stabilize
- After cascade: when funding rate flips deeply negative + OI has dropped 20%+, contrarian long entry with tight risk
7. Exchange-Level Liquidation Differences
| Exchange | Liquidation Engine | Key Feature |
|---|---|---|
| OKX | Tiered auto-deleveraging | Partial liquidation (reduce position size, not full close) |
| Binance | Insurance fund + ADL | Largest insurance fund (~$1B+) reduces cascade severity |
| Bybit | ADL (Auto-Deleveraging) | ADL triggers when insurance fund depleted |
| dYdX | On-chain liquidation | Transparent, anyone can liquidate (MEV opportunity) |
Data Sources
| Source | Access | Data Available |
|---|---|---|
| CoinGlass | Free (limited) | Liquidation heatmap, 24h liquidations, OI |
| Laevitas | Free/Paid | Options + futures liquidation levels |
| Kingfisher (Coinalyze) | Paid | Real-time liquidation level estimates |
| Hyblock Capital | Paid | Professional liquidation heatmaps |
| OKX API | Free | Historical liquidation data |
| DeFi Llama | Free | DeFi protocol liquidation data |
Output Format
## Liquidation Analysis — [Asset] — [Date]
### Liquidation Overview (24h)
- **Total liquidated**: $XXX M
- **Long liquidated**: $XXX M (XX%)
- **Short liquidated**: $XXX M (XX%)
- **Largest single**: $XX M [exchange]
- **Market state**: [extreme / elevated / normal / calm]
### Key Liquidation Levels
| Price Level | Type | Est. Volume | Distance from Current | Priority |
|------------|------|-------------|----------------------|----------|
| $XX,XXX | Short liq cluster | $XXX M | +X.X% | High |
| $XX,XXX | Long liq cluster | $XXX M | -X.X% | High |
| $XX,XXX | Long liq cluster | $XXX M | -X.X% | Medium |
### Heatmap Summary
- **Strongest upside magnet**: $XX,XXX (short liquidation cluster, $XXX M)
- **Strongest downside magnet**: $XX,XXX (long liquidation cluster, $XXX M)
- **Asymmetry**: [upside magnet stronger / downside stronger / balanced]
### Cascade Risk
- **Downside cascade risk**: [high / moderate / low]
- [X clusters stacked within X% below current price]
- **Upside cascade risk**: [high / moderate / low]
### Post-Liquidation S/R Levels
- **Recent support formed**: $XX,XXX (long liquidation wipeout on DATE)
- **Recent resistance formed**: $XX,XXX (short liquidation wipeout on DATE)
### Trading Implications
- **Bias**: [upward magnet stronger → mild bullish / downward → bearish]
- **Risk**: [high leverage zone within X% → reduce position size]
- **Key level**: [$XX,XXX — if broken, cascade risk activates]
Notes
- Liquidation data is estimated, not exact — exchanges do not publish real-time liquidation level details for all users
- Heatmap providers use statistical models based on OI and leverage distribution to estimate liquidation prices
- Liquidation levels shift constantly as traders open/close positions — treat as dynamic zones, not fixed prices
- "Stop hunts" (price briefly touching a liquidation cluster then reversing) are common — market makers deliberately trigger clusters
- DeFi liquidations are fully transparent (on-chain) but CEX liquidations are opaque
- This framework is for research purposes only and does not constitute investment advice