Fear-Greed Extremes as a Buy Signal — How My Agent Caught the Bottom

The Problem

Everyone talks about "buying fear." Almost nobody does it.

The reason is simple: when the market is in extreme fear, you're terrified too. You're reading the same headlines. You're watching your portfolio bleed. The last thing your brain wants to do is deploy capital.

I've been in crypto since 2017. I've seen four capitulation events where extreme fear marked the bottom. I bought aggressively in exactly one of them. The other three, I hesitated, waited for "confirmation," and entered 30-40% above the actual bottom.

The problem isn't knowing that fear = opportunity. The problem is knowing in the moment — when your amygdala is screaming — and having a system that overrides the panic.

The Solution

I configured two agents to work in tandem. The first — part of my crypto market data stack — monitors the fear-greed index continuously and fires Telegram alerts when it hits extremes (below 20 or above 80). The second — Hari, my investment advisor agent — receives the alert and cross-references it with on-chain data, exchange flows, and funding rates before recommending an action.

The human still makes the final call. But the system ensures I never miss the signal, and Hari's analysis cuts through the emotional noise.

The Process

The alert configuration:

# Fear-Greed Monitor — runs every 4 hours
name: fear-greed-extremes
schedule: "0 */4 * * *"
channel: telegram

task: |
  Check crypto Fear & Greed Index.
  Current value + 7-day trend + 30-day avg.

  ALERT CONDITIONS:
  - Index < 20: 🚨 EXTREME FEAR — potential buy zone
  - Index < 10: 🔴 CAPITULATION — high-conviction buy zone
  - Index > 80: ⚠️ EXTREME GREED — risk management check
  - Index > 90: 🟡 EUPHORIA — consider taking profits

  On alert: tag Hari for cross-reference analysis.
  Include: what caused the drop, exchange inflows/outflows,
  stablecoin supply changes, funding rates.

When the alert fires, Hari's analysis looks like this:

# Hari's cross-reference logic (simplified)
async def fear_extreme_analysis(fear_value: int):
    signals = {
        "exchange_flows": await get_exchange_netflows(),     # Net BTC in/out
        "stablecoin_supply": await get_stablecoin_mint(),    # USDT/USDC minting
        "funding_rates": await get_perp_funding(),           # Negative = shorts paying
        "liquidations_24h": await get_liquidation_data(),    # Long vs short liqs
        "whale_accumulation": await get_whale_addresses(),   # Wallets >1000 BTC
    }

    bullish_count = sum(1 for s in signals.values() if s["bias"] == "bullish")

    return {
        "fear_greed": fear_value,
        "confirming_signals": bullish_count,
        "total_signals": len(signals),
        "recommendation": "DCA entry" if bullish_count >= 3 else "wait",
        "confidence": bullish_count / len(signals),
        "suggested_allocation": "25% of dry powder" if bullish_count >= 4 else "10%"
    }

Real alert that fired on a major drawdown:

🚨 EXTREME FEAR ALERT

Fear & Greed Index: 17
7-day trend: 42 → 31 → 24 → 17
30-day average: 38

HARI CROSS-REFERENCE:
✅ Exchange outflows: +18,400 BTC left exchanges (accumulation)
✅ Stablecoin minting: $2.1B USDT minted in 48h (dry powder)
✅ Funding rates: -0.03% (shorts paying longs — bearish consensus)
⚠️ Liquidations: $890M long liqs in 24h (pain not over?)
✅ Whale wallets: net +4,200 BTC in >1000 BTC addresses

Confirming signals: 4/5
Recommendation: DCA entry — 25% of dry powder
Suggested: BTC 60%, ETH 30%, SOL 10%

I followed the recommendation. Deployed 25% of available stablecoins across three entries over 72 hours.

The Results

The backtest across all instances where fear-greed dropped below 20 since the index was created:

Try It Yourself

1. Set up the fear-greed monitor with your preferred alert thresholds
2. Define your cross-reference signals — don't buy on fear-greed alone
3. Pre-commit a DCA plan: what % of dry powder at what levels
4. Let the agent deliver the signal and the analysis — you make the call
5. Journal every decision: followed the signal, ignored it, and what happened

The system doesn't remove emotion from investing. It gives you a structured counterargument when your gut says "wait."

The best trades I've ever made felt terrible at the time. Now my agent makes sure I don't miss them.