A scanner flags a stock at 25% short interest and ranks it near the top of the squeeze-candidate list. Another stock sits at 25% too, three rows down. Same percentage, same scanner column, treated as roughly the same setup. They aren't. One of those 25% readings can represent a million shares that the stock's own daily volume can barely absorb in a week; the other can represent a hundred and fifty million shares that a single active session chews through without much drama. The percentage is the same. The float behind it is not, and the float is the number that decides which stock is actually dangerous to be short — or worth being long against.
What the percentage is actually built from
Short interest is a straightforward count: the total number of shares currently sold short and not yet bought back, reported by FINRA twice a month. On its own it's just a share count, so it gets turned into a percentage by dividing by the float — the shares actually available to trade, after you strip out what's locked up with insiders, founders, and long-term holders who aren't selling. A company can have 200 million shares outstanding and a float of only 15 million if most of the stock sits with management and a handful of institutions. Short selling itself only works against that tradable float, because that's the pool of shares a broker can actually locate to lend.
That division is where the information gets lost. A scanner column that reads "Short % of Float: 25%" has already thrown away the two numbers that produced it. You can't get them back without pulling the float and the raw short-interest count separately — and until you do, you don't actually know what you're looking at.
There's a second reason not to take the number at face value: it's old by the time you see it. FINRA requires member firms to report short interest twice a month, and the published figure usually lags the settlement date by a few trading sessions. A stock can have covered a meaningful chunk of its short position, or added to it, in the days between that settlement date and the morning a scanner shows you the reading. The percentage isn't wrong, exactly — it's a real number from a real filing — it's just a snapshot of a position that's already had a week or more to move, on top of being a ratio that hides the float it was built from.
The same percentage, three different floats
25% short interest at three float sizes
| Float | Shares short (25%) | Typical daily volume | Days to cover | What it implies |
|---|---|---|---|---|
| 4,000,000 (micro) | 1,000,000 | 150,000 | 6.7 days | Thin volume can't absorb forced buying quickly — a spike session can compress a week's worth of covering into hours |
| 40,000,000 (mid) | 10,000,000 | 3,000,000 | 3.3 days | Enough daily liquidity that covering can spread across a few sessions without a violent single-day move |
| 600,000,000 (mega) | 150,000,000 | 40,000,000 | 3.75 days | Days-to-cover looks similar to the micro-float row, but the float is deep enough that nothing structurally forces simultaneous covering |
Those are illustrative numbers, not a specific ticker, but the pattern holds in real scanners constantly: the mega-float row and the micro-float row can land within a day of each other on days-to-cover, and a ranking built on the raw percentage alone has no way to tell you that the micro-float setup is the one where a single active session can actually force the issue, while the mega-float setup needs sustained buying across more sessions than the number by itself suggests — because "40 million shares a day" is a small fraction of a 600-million-share float, even though it's a large number in isolation.
Notice what the mid-float row is doing in that table, too: it has the lowest days-to-cover of the three, which by the simple "lower is scarier, higher is safer" reading would make it the least interesting setup. It isn't. Days to cover alone doesn't tell you whether the float is small enough for a surge in demand to matter, or whether it's just a stock that happens to trade a lot relative to its (still substantial) short position. You need the float number sitting next to it to know which story you're in.
The second number: how fast the float actually turns over
Days to cover answers "how many sessions of normal volume would it take." It doesn't answer whether normal volume is likely to show up at 2x, 5x, or 10x on the day it matters. That's float turnover — average daily volume divided by float — and it's the number that tells you whether a stock's liquidity can plausibly spike hard enough to run through a short position fast.
Float turnover, worked from the table above
The micro-float name doesn't need an unusual amount of buying to meaningfully reshuffle who holds the float — a session at 3x normal volume is still a small absolute number of shares, but it's now a real chunk of everything available to trade. The mega-float name would need a genuinely extraordinary session, in dollar terms, to do the equivalent. That's the gap a bare short-interest percentage can't show you, because percentage and float turnover are answering two different questions that happen to get collapsed into the same scanner column.
Think about why that gap exists in the first place. A float is just the pool of sellers available on any given day. A thin float means a small pool — not many shares, and often not many distinct holders willing to part with them at the current price. When buying pressure shows up, whether from short covering or ordinary demand, it has to find sellers inside that small pool, and a small pool runs out of willing sellers quickly, which is what pushes the price up fast enough to force even more covering. A deep float is the opposite kind of pool: there's almost always another seller sitting a few cents higher, which is exactly what keeps a surge in demand from turning into a vertical move. The short-interest percentage never changes between those two cases. The depth of the pool it's measured against does all the work.
Same volume headline, opposite reaction
Put a plausible, identical-looking volume spike against both ends of that spectrum and the price reaction isn't close to the same shape.
Nothing about the short-interest percentage changed between those two charts. What changed is whether the float was shallow enough for the buying to run through it or deep enough to shrug it off — which is exactly the variable a scanner's single percentage column hides.
This also cuts the other way, and it's worth saying plainly: a thin float with heavy short interest isn't automatically a buy because it might squeeze. It's a stock where the range on both sides is wider than the percentage suggests — the same shallow pool of sellers that can send it sharply higher on a surge of buying can send it sharply lower just as fast once the surge fades, because there was never much depth holding the price up either. Treat a thin float as a sign that normal-looking moves will be bigger than usual in both directions, not as a one-way signal.
What to actually pull before trusting the number
Before you treat a short-interest reading as a squeeze setup
-
You pulled the float, not just the percentage — The same 25% means a different trade at a 4-million-share float than at a 600-million-share float — you need the raw number to know which one you're looking at.
-
You calculated days to cover against current average volume — Shares short ÷ average daily volume. A high number means a lot of forced buying would have to happen relative to how the stock normally trades — a low number means the float can absorb it without much strain.
-
You checked float turnover, not just days to cover — Average volume ÷ float tells you how much of the float moves on a normal day, which is the context days-to-cover needs to mean anything at all.
-
You're trading the percentage alone because it's at the top of a scanner list — A ranking built on short-interest percentage with no float or volume context can put a slow, deep-float setup above a fast, thin-float one — the sort order isn't telling you what you think it's telling you.
The bottom line
A short-interest percentage is a ratio, and a ratio is only as informative as the numbers that built it. The same 25% sits on top of a million shares at one float and a hundred and fifty million at another, and those aren't comparable trades just because the scanner column matches. Pull the float, calculate days to cover against real average volume, and check float turnover to see whether a spike can plausibly run through the position fast — that's the work a bare percentage skips, and it's the difference between a genuine squeeze setup and a number that happens to look like one. Liquidity is the thing that was missing from the percentage all along; the short-interest column just hides it better than most.
