Lazy Trader AIOpen any trading Discord and you'll find someone breathlessly posting screenshots of “MASSIVE UNUSUAL FLOW” on a ticker. The premium is real. The size is real. And most of the time, none of it means what they think it means.
Roughly 60% of unusual options flow is hedging. Another good chunk is market-maker delta management, calendar rolls, or a portfolio manager rebalancing risk. The signal — actual directional positioning by someone who thinks they know something — is the smaller, quieter slice.
This is a guide to filtering for the slice that matters.
“Unusual” alone isn't a signal. Unusual plus aggressive ask-side execution, no offsetting hedge, and a high vol-to-OI ratio is what you actually want to watch.
What real institutional flow looks like
Big positioning leaves a fingerprint. The same trade placed differently can mean wildly different things — and the difference is mostly in the execution.
Sweeps that lift the ask
A sweep is an order routed across multiple exchanges simultaneously to fill quickly. When that sweep is on the ask side, the buyer is paying up — they don't want to wait. Patience is cheap; urgency is expensive. Urgency usually means conviction.
Volume that dwarfs open interest
If a contract has 500 open contracts and 4,000 trade in a single hour, that volume is overwhelmingly new positioning. Compare that to a contract with 80,000 OI where 4,000 contracts trade — same volume, far less informative.
The four filters we use
This is the filter stack we ship inside Lazy Trader AI. It's deliberately strict — we'd rather miss noise than alert on it.
- Volume / Open Interest ≥ 3×. The contract is being newly built, not redistributed.
- Premium ≥ $1M total. Filters out retail-sized sweeps that look dramatic on a chart.
- ≥ 70% of the order on the ask. Buyer-initiated, not a market-maker spread fill.
- No matching hedge in the underlying. If a $5M call sweep coincides with a $5M short in stock, you're watching a delta-neutral hedge, not a bet.
Apply all four together and the alert volume drops by roughly 90%. What's left is the fingerprint we want.
Two case studies
The good signal: $TSLA, October 7
$8.4M in $700 strike calls, expiring in three weeks. Sweep, 76% on the ask. Vol/OI of 3.2×. No corresponding short in the stock. Stock was flat that morning.
Twelve days later: Q3 deliveries beat by 9%. The contract went from $4.20 to $19. The sweep was placed by someone who, at minimum, had a strong view on the print.
The stock didn't move. The flow did. That's the asymmetry you're looking for.
The fakeout: $XYZ, November 14
$12M in put sweeps. Looks bearish. But same day, the underlying saw matching long stock buying at the same notional. Net delta: zero. This was a hedged position being rolled, not a bet.
The naive “unusual flow” feed flagged it as bearish. The position closed flat the next week. No edge for anyone who copied it.
Common mistakes
- Treating size as the signal. A $20M block can be one institution rebalancing pension exposure. A $1.5M ask-sweep on a thinly-traded weekly is more interesting.
- Ignoring expiration. Long-dated LEAPs and weeklies are different animals. Weeklies ≤ 14 days out are where short-term conviction lives.
- Forgetting context. The same sweep one week before earnings means something different than the same sweep three months before.
- Following without sizing. Even a perfect signal can be wrong. Real institutional traders are right ~55% of the time. Risk accordingly.
Putting it together
The shortcut: don't watch all unusual flow. Watch aggressive, premium-heavy, urgent flow with no offsetting hedge. That's the slice with predictive value.
Lazy Trader AI's flow alerts apply this filter stack automatically. When something clears it, you get a push — not a firehose.
Vol/OI ≥ 3× · Premium ≥ $1M · Ask-side ≥ 70% · No matching hedge · Expiration matters · Size your bets like the signal is right ~55% of the time.
This article is for educational purposes and is not financial advice. Lazy Trader AI is a market-monitoring tool, not a brokerage.