How we grade a falling knife — and why velocity is the tell

A stock down 30% in a week generates two opposite reactions from the same headline. The bargain hunter sees a discount. The risk manager sees a warning. They are reading the same number and disagreeing about what it means — and most of the time, the risk manager is right.

The reason is velocity. The speed and shape of a decline carries more information than its magnitude. A name that grinds down 30% over two quarters on deteriorating fundamentals is a different animal from one that gaps down 30% in three sessions on a broken thesis. We grade both, but we grade them differently, and we publish the rule so you can argue with it.

The four checks

When a name shows up in the cut down sharply, it runs four checks before we’ll attach a verdict. A “Falling Knife” flag means it failed at least two of them.

1. Is there a filing? A real dislocation usually has a primary-source cause — an 8-K, a guidance cut, a restatement, an insider sale cluster. If price moved hard and nothing was filed, that’s not information, that’s flow. Flow reverts. Information re-rates.
2. Does the flow confirm? We check whether options and dark-pool activity lean the same direction as price. A drop that institutions are quietly buying into is a different setup from one they’re hedging away from. Disagreement between price and flow is the single most common false-knife signal.
3. Is the balance sheet a clock? Velocity is most dangerous when the company has a deadline — a maturity wall, a covenant, a cash runway measured in quarters. A cheap multiple on a business that has to refinance into a closed window is not cheap.
4. Is short interest the cause or the effect? Rising short interest into a fundamental break is confirmation. Rising short interest on a crowded, already-broken name can be a squeeze setup. The trend matters more than the level.

Velocity, concretely

We don’t eyeball charts. We score the move against the name’s own distribution: how many standard deviations is the drawdown, over how many sessions, with what volume participation, and how does the shape compare — a single gap (event) versus a staircase (deterioration) versus a slide (distribution).

The question is never “how far did it fall.” It’s “what kind of falling is this, and who is on the other side of it.”

A worked example

Earlier this quarter a consumer name dropped 14% in two sessions. It screened cheap on a trailing basis and the message boards called it a gift. It failed three of the four checks: no new filing of substance, options flow leaning to downside puts, and a maturity inside eighteen months against a thinning cash position. We graded it Falling Knife. It was down another 19% within the month. No prediction was required — only a refusal to confuse a low price with a cheap one.

The opposite case is just as instructive. A networking-silicon name sold off 11% on a sector-wide derating with no company-specific filing, while two officers filed open-market purchases and three funds initiated positions. Price and flow disagreed, and the flow was the informed side. That one graded Deep Dive, not knife.

What this is not

This is not a prediction that every fast decline keeps falling. Some of the best entries we publish are violent selloffs where the flow and filings say the market overreacted. The grade isn’t “down, therefore avoid.” It’s a disciplined reading of why it’s down and who is acting on it. Velocity is the tell that tells you which question to ask.

---

This note is research, not advice. Examples are illustrative and may not reflect current positioning.