How I Use AI to Track Whale Trades and Make Consistent Money

The Setup

Most people lose money following "Whales" because they don't understand hedging. If a fund owns $100M of Apple stock and buys $1M in Puts, that's insurance, not a directional bet. This strategy separates the insurance from the real conviction bets using a 6-step filtering pipeline that runs every 4 hours.

What the AI Found

1. 1.Step 1 — Market-wide scan found Tesla, Meta, and Nvidia with large clustered call activity and minimal puts, signaling positive directional bets. QQQ and SPY were heavy on puts — classic portfolio insurance, not crash bets.
2. 2.Step 2 — Filtered for trend alignment. Tickers where bullish flow matched an uptrend (price above 20-day SMA) were kept. Bearish flow in uptrends was rejected as hedging. Meta and ORCL passed with bullish flow + confirmed uptrend.
3. 3.Step 3 — IV filter eliminated overpriced setups. Meta's IV Rank was at the 46th percentile — well below the 70th percentile cutoff — meaning premiums were still reasonably priced before the crowd piled in.
4. 4.Step 4 — AI sentiment check on news and social media. No binary events (earnings/FDA) within 7 days for Meta. Sentiment score positive, aligning with the bullish flow. No contradictions flagged.
5. 5.Step 5 — Strike adjustment for retail risk management. Meta's original whale strike was Dec 5 expiry, padded to Dec 19 (+14 days). Strike stayed the same since it was already within 2% of ATM.
6. 6.Step 6 — Black-Scholes scoring ranked Meta Dec 19 $675 Call as the top trade. High R/R ratio, positive sentiment, reasonable IV cost.

The Trade

META $675 Call, Dec 19 expiry. Adjusted from the original whale flow (Dec 5 expiry) with 14 extra days of breathing room. Score combined Risk/Reward strength, sentiment alignment, and IV cost.

The Result

2 days later, Meta announced a 30% cut in metaverse budget, shifting resources to AI. Stock jumped 3%, contract up 100% in 3 days. Overall strategy stats: 56% win rate, +85% average winner, -30% average loser. The wins are modest but consistent — the 6 layers of risk filtering are the edge.

The Prompt

Step 1: Scan the entire market for unusual institutional options activity. Filter for premiums above $50K and expiry within 90 days. Cluster by ticker and show me the top 20 by total premium, broken down by calls vs puts.

Step 2: For each ticker from Step 1, pull detailed flow data and the 20-day SMA. Reject any ticker where the flow direction contradicts the price trend (bearish flow + uptrend = hedge, bullish flow + downtrend = hedge). Keep only where flow matches trend.

Step 3: For the remaining tickers, check IV Rank. Reject anything above the 70th percentile — I don't want to pay inflated premiums.

Step 4: For what's left, grab recent news and social sentiment. Flag and reject any ticker with a binary event (earnings, FDA, lawsuit) in the next 7 days. Score sentiment from -1 to 1 and reject if sentiment contradicts the flow direction.

Step 5: For the surviving trades, adjust the strikes: push expiry out by 14 days for breathing room, and move strikes to within 2% of ATM.

Step 6: Run Black-Scholes on each adjusted trade. Calculate max loss, max profit, probability of profit, and breakeven. Rank by: Score = (Risk/Reward) + (Sentiment Score) - (IV Cost). Show me the final ranked list.