Float, Short Interest & Supply Dynamics

Float tightness, SI% of float, days to cover. Score the squeeze setup. Do not invent the catalyst.

Track 6. Supply events that change the share count live in IPOs, Offerings & Buybacks (Course 56). Locate, borrow, and unlimited-loss anatomy live in Short Selling & Short Squeeze Mechanics (Course 47) — this page is not a second borrow-how-to. Next: Macro, Rates & Equity Regimes. Hub: stock courses.

A Tight Float Is Not a Squeeze

A small tradable supply is a constraint. A high short book against that supply is a constraint. Neither one is a catalyst, and neither one is a countdown. Traders who “buy the squeeze” because a screen painted a fat SI% are buying a label. The tape still needs a reason for shorts to cover, for new longs to compete for the same shares, and for you to be sized as if the name can gap through your stop — because it can.

This course is supply scoring: what float is versus shares outstanding; how to compute short interest as a percent of that float; why days to cover (SI ÷ ADV) is a scale label and not a timer; and how to grade a short squeeze setup without inventing the event. Course 47 already taught the short: locate, borrow, buy-in risk, unlimited loss on the short side. You will not re-learn how to borrow stock here. You will learn whether the supply stack even supports the story you are about to journal. Educational only; not personalized investment advice.

1. Float vs Shares Outstanding

Shares outstanding is the full claim count: every share the issuer has issued and not retired. Market capitalisation is usually price × that claim count (or a close cousin — diluted vs basic is a vendor choice). Float is the subset that is freely tradable in the public market: outstanding minus stock that is restricted, closely held, locked, or otherwise not available to trade. Write the identity before you trust a screener tile:

Float ≈ Shares outstanding − Restricted / closely held / locked shares
SI% of float = Short interest ÷ Float
Days to cover = Short interest ÷ Average daily volume

Three failure points hide in that first line. Vendor float is not a legal term of art. One data vendor subtracts 5%+ holders; another subtracts insiders and 10% holders; a third uses a “free float” index-provider definition. TODO:VERIFY the definition on the screen you are reading before you compare two names, or you are ranking two different objects. Restricted is not the same as “insiders will never sell.” Lockups expire. Form 144s print. Course 56’s secondaries, follow-ons, and lockup calendars change the float. A tight float last month is not a tight float the week a lockup or an ATM facility dumps supply. Outstanding itself moves — buybacks retire shares; primary issuance and stock-based compensation add them. If you did not check the latest 10-Q share count, you did not measure float. You measured a memory.

An IPO is the extreme case: a small public float against a large insider remainder, then a lockup clock. That is a Course 56 supply event with a Course 57 denominator. Do not collapse them. The IPO day is an issuance. The months after are a float that is still being manufactured.

Operator test: if you cannot say (1) outstanding, (2) what was subtracted, (3) the vintage of both, you do not have float. You have a marketing adjective. “Low float” without a share count is the same error as “cheap” without a multiple.

2. Short Interest as a Percent of Float

Raw short interest is a share count: how many shares have been sold short and not yet covered, as of a settlement date, as reported in a publication vintage. It is not the live locate tape, not the borrow fee, and not “how many people are betting against the company.” It is a lagged inventory statistic. TODO:VERIFY the current FINRA / exchange short-interest calendar before you treat a screen as today’s book. This page will not invent reporting dates, thresholds, or “average SI% for the Russell.”

The useful ratio is against float, not against outstanding, because the shorts have to cover in the tradable supply:

SI% of float = Short interest shares ÷ Float shares

A name with 40 million shares short against a 400 million float is 10.00% of float. A name with 4 million short against a 20 million float is 20.00% of float. The first number looks “bigger.” The second supply stack is tighter. Ranking by short-interest shares is how large-caps fake a squeeze story. Ranking by SI% of outstanding (instead of float) is how you understate tightness when insiders still hold the majority.

There is no DennTech SI% that “always squeezes.” A 25% SI% of float can sit for months while the borrow is easy, the tape is two-sided, and no catalyst exists. A 8% SI% of float can be violent if the float is tiny, ADV collapses, a halt prints, and a forced covering flow hits a one-sided book. Percentiles on a vendor screen are vendor percentiles. TODO:VERIFY them if you use them. They are not laws of covering.

Two cousins that are not SI%. Utilization (shares on loan ÷ lendable inventory) is a securities-lending print. It can be high while published SI% is stale, or low while the SI report still looks crowded. Borrow fee / rebate is a price of scarcity, intra-day, broker-dependent. Both belong to the locate market Course 47 already framed. Use them as current scarcity tells. Do not substitute them for the SI identity, and do not invent a fee that “guarantees” a squeeze.

3. Days to Cover Is a Label, Not a Timer

Days to cover (sometimes “short ratio”) is:

Days to cover = Short interest shares ÷ Average daily volume

If SI = 4,800,000 and ADV = 800,000, days to cover = 6.00. That sentence means: at this vintage’s average volume, the short book is six average sessions of volume. It does not mean the squeeze completes on session six. Covering is not a straight line at ADV. ADV itself explodes or dies when the name is in play. Shorts can add. New supply can print (secondary, ATM, lockup). Longs can sell into the same bid the shorts need. Treat 6.00 as a scale label — “the short book is large relative to typical volume” — the way you treat a high multiple as “expensive,” not as “the re-rating happens on Friday.”

ADV vintage is a second trap. A 20-session ADV and a 90-session ADV are different denominators. A name that just ran from 200k to 8 million shares a day will show a tiny days-to-cover on a short window and a fat one on a long window. Write the window. TODO:VERIFY the vendor’s ADV definition. Days to cover computed on a squeeze day’s volume is a joke: the denominator was the event.

Verify the percentage math — SI% of float, and the percent change in ADV when the name wakes up — in the percentage change calculator. Do not narrate 19.7% as “about 20% and therefore a squeeze.” The calculator does not know what float is. You still have to feed it the right two numbers.

4. Squeeze Setup Quality vs a Guaranteed Squeeze

A squeeze, in operator language, is covering demand hitting a thin offer — shorts buying back, optionally plus new longs, into a float that cannot absorb the flow without a large print. The setup is the supply stack plus a plausible reason for that flow. The event is the flow. Screens sell you the setup as if it were the event. It is not.

Quality of setup (a checklist, not a score that sums to a buy):

• Tight float, honestly measured. Small tradable supply vs the order you would actually need to cover or to enter. Vendor definition written down.
• Elevated SI% of float at a named vintage. High enough that covering would matter if it happened. Not a magic cutoff. Not “always.”
• Days to cover as a large label on a sensible ADV window — the short book is large vs typical volume.
• Current scarcity tells from the locate market (fee, utilization, “hard to borrow” at your broker). Course 47 objects. Stale SI with an easy borrow is a different animal.
• A real catalyst path — earnings, a regulatory print, an index add, a buy-in, a buyback, a short-report that failed — not “the SI% is high so it has to go.” No catalyst, no setup. You have a crowded name that can stay crowded.
• Tape that can actually gap. Wide spreads, one-sided Level 2, pre-market prints that do not mean a fill at your limit, possible trading halts. A halt is a halt, not a gift. Size as if you do not get the stop you drew.

What quality is not: a guaranteed squeeze, a target price, a “days until covering,” or a reason to ignore Risk 101. The names that historically squeezed and the names that historically did nothing with similar SI% are not a DennTech backtest. We will not quote hit rates. If a high-SI name goes nowhere, the setup was a constraint without a flow. That is the base case until the tape disagrees.

Volume is a character, not a proof of covering. Elevated volume can be new longs, can be shorts adding, can be a secondary printing into strength. Pair the supply stack with volume analysis as a question — who is the volume — not as a squeeze stamp. VWAP is an execution benchmark for the session you are in, not a squeeze indicator. Do not promote it into one.

5. Options Hedging Can Amplify. It Is Not an Official Named Event

Listed options on a name can add a second flow: dealers (and other hedgers) buying or selling the underlying as their net delta changes with price and implied volatility. That hedging can amplify a move that is already happening. It is a possible mechanical overlay. It is not an official named squeeze species on this site, and it is not a calendar you can trade from a cartoon of dealer positioning.

Rules for this page: do not journal an informal options-hedging label as if it were a defined event with a start time. Do not import crypto glossary language. Do not paste unsourced dealer-position or options-hedging-flow statistics. TODO:VERIFY any live dealer-position, open-interest, or hedging-flow figures from a named vendor, with the vendor’s definition, before they enter a journal. This course invents none of those stats. If you cannot name the vendor and the formula, you do not have an options-hedging thesis. You have a story that rhymes with a move you already wanted.

Practical consequence: if the equity setup in Sections 1–4 is weak, options chatter does not rescue it. If the equity setup is real and options are listed and active, treat hedging as possible extra convexity you did not size for — which is an argument to size smaller, not larger. Binary event risk (Course 28 family) still binds. A halt still binds.

6. What Each Number Answers

Use the right lens. Mixing them is how “20% SI” becomes a personality.

A workable sequence: share count and float definition → SI% of float at a named vintage → days to cover as a label on a named ADV window → locate-market scarcity (Course 47) → catalyst path → tape and halt plan → size. SI% is step two, not step one, and never step “therefore buy.”

7. Course 47 vs This Page — Do Not Duplicate the Borrow

Course 47 is the short: how locate and borrow work, why a short has theoretically unlimited loss, what a buy-in is, and why a squeeze is a covering event rather than a morality play. If you are short, that course owns the mechanics. This page owns the denominator those mechanics sit on.

If you are long a squeeze candidate, you still do not get a free pass on locate literacy. A hard-to-borrow name with a rising fee tells you the short is currently expensive to hold. It does not tell you the short will cover on your schedule. It does not tell you the offer will be there for your 66 shares at the limit you typed in pre-market. Session clocks live on the U.S. session hours 2026 page. Thin books plus extended-hours prints are how squeeze tourists get filled at prices that never existed in Core.

Margin, if you use it, is the post-PDT framework: typical ~$2,000 equity minimum to use a margin account, broker-dependent, house rules may be stricter (including RN 26-10 / Rule 4210(d)(2) as the current day-trading-margin conversation — not a $25,000 PDT floor as live law). Intraday, firms monitor equity versus risk during the session — see intraday margin requirements. None of that is a squeeze edge. A concentrated squeeze long on margin is a gap-risk product. Size from invalidation first.

Shorting into a “setup” you scored on this page is still Course 47 plus Risk 101. High SI% of float is not a reason to be short; it is a reason to ask whether the other side is already crowded. Being short a tight-float crowded name is how unlimited-loss diagrams stop being theoretical. This course will not teach you to hunt that as a personality.

EXAMPLE — Two Hypothetical Names, $10,000 Book

EXAMPLE. Illustrative filings-style figures, not live quotes, not a recommendation, and not a backtest. Account: $10,000 cash. Risk budget per idea: 1% of equity = $100. Shares always rounded down. Notional cap for this book: 10% of equity = $1,000 per name. Use the tighter of dollar-risk size and notional cap. No leverage in this example. Float definitions are stated; they are not a universal vendor standard — TODO:VERIFY any live screen against its own methodology.

Sanity-check SI% (20.00% vs 10.00%) and the percent the notional cap cut TIGHT (from 66 shares to 55: (55 − 66) ÷ 66 = −16.67%) in the percentage change calculator. Size both names from the tighter rule. Do not pick a winner. The pedagogical point is the opposite: the factor that looks like “more shorts” on WIDE is the wrong sort, and conviction about TIGHT’s 20.00% does not override a $1,000 notional cap on a $10,000 book. Use the risk calculator the same way: the live UI is crypto-branded; for stocks treat the unit as shares and round down. No stock-native calculator exists on this site. Stops are not a fill guarantee into a halt or a gap through $16.50.

8. Common Mistakes and Limits

• Treating days to cover as a countdown. 6.00 is a label. Covering is not a straight line at ADV. ADV is not stationary.
• Inventing an SI% that “always squeezes.” There is none on this page. Percentiles are vendor objects. TODO:VERIFY them if you use them.
• Ranking by short-interest shares. Name WIDE’s 40 million short shares lost to Name TIGHT’s 20.00% of a 24 million float. Sort the ratio, not the headline count.
• Using outstanding as the SI denominator when a large insider remainder is the whole story. The shorts cover in the float.
• Ignoring Course 56 supply. A lockup, secondary, or ATM is new float. Yesterday’s tightness is not a covenant.
• Re-teaching yourself to borrow on this page. Locate, fee, buy-in, unlimited loss: Course 47. This page scores the stack. If you cannot locate, you do not have a short. If you are long, the locate tape is a scarcity tell, not a fill.
• Promoting options hedging into an official named event. Possible amplifier. TODO:VERIFY dealer stats. No invented positioning. No crypto glossary import.
• Confusing volume with covering. New longs, shorts adding, and issuance all print volume. Volume analysis asks who; it does not stamp “squeeze.”
• Sizing from the story. In the EXAMPLE, TIGHT’s 1% budget wanted 66 shares; the notional cap allowed 55. The cap won. That is the correct outcome. Floor the share count. Never round up into a share you did not budget.
• Trading the setup in a session you did not map. Pre-market prints, halts, and Core are different books. Diagnosis is research. Execution uses the clock.
• Letting a “squeeze identity” override the book. Five tight-float names is a correlated gap cluster with a theme sticker — see portfolio basics. The label does not increase the dollars you are allowed to lose.

Limits. Supply math does not time entries. SI vintages lag. Vendor float definitions disagree. Borrow fees are not this site’s feed. Halts and gaps make stops a wish. Options hedging can amplify in either direction. This course is educational — not personalized investment advice, not a squeeze watchlist, and not a claim that any SI% outperforms from here. Historical squeeze anecdotes are anecdotes. They are not a DennTech hit-rate table.

Pre-Trade / Research Checklist

1. Outstanding share count from the latest filing. Note dilution, buybacks, ATMs.
2. Float: write what was subtracted and which vendor. TODO:VERIFY if the screen is unlabeled.
3. Course 56 overlay: lockup, secondary, buyback, IPO remainder. Is the float about to change?
4. Short interest shares, settlement date, publication vintage. Not “today’s locate.”
5. SI% of float = SI ÷ float. Write the percent to two decimals. No slogan rounding.
6. ADV window named. Days to cover = SI ÷ ADV as a label. Do not draw a date.
7. Locate-market tells (Course 47): fee, utilization, HTB at your broker. Scarcity now vs SI then.
8. Catalyst path in one sentence a skeptic could falsify. No catalyst → no squeeze thesis, only a crowded name.
9. Tape: spread, Level 2, halt risk, session you will actually trade. VWAP as execution, not as a squeeze stamp.
10. Options overlay, if any: possible amplifier only. No unofficial named-event label. No unsourced dealer stats.
11. Size: $100 on a $10,000-style 1% budget (or yours); notional cap; shares from the risk calculator (unit = shares, round down). Tighter rule wins.
12. Journal: “Float = ___ (def ___). SI% = ___. DTC label = ___. Wrong if price ___ or supply event ___. Sized to ___ shares.”

Tools for This Course

• Percentage Change Calculator — verify SI% of float and ADV regime shifts from the formula (as in the EXAMPLE: 20.00% vs 10.00%; the −16.67% share cut on TIGHT).
• Risk Calculator — live crypto UI; for stocks treat the unit as shares and round down. No stock-native calculator exists. Binding cap in the EXAMPLE was notional on both names.
• Stock Courses Hub · prev: IPOs, Offerings & Buybacks · next: Macro, Rates & Equity Regimes.

FAQ

Is float the same as shares outstanding?

No. Outstanding is the full claim count. Float is the freely tradable subset after you subtract restricted, closely held, or locked shares — and the subtraction is vendor-defined. TODO:VERIFY the screen’s methodology. Market cap usually rides outstanding, not float. An IPO remainder plus a lockup is a Course 56 supply clock sitting on this denominator.

Does a high days-to-cover number mean the squeeze happens in that many sessions?

No. Days to cover = short interest ÷ ADV is a scale label: how large the lagged short book is versus typical volume. Covering is not a straight line at ADV, ADV is not stationary, and new supply can print. Do not draw a date on the label.

What SI% of float always squeezes?

None that this course will ship. A high SI% of float is a constraint. A squeeze is covering flow into that constraint. Without a catalyst path and a tape that can actually move, you have a crowded name. Vendor percentiles are not laws. TODO:VERIFY any cutoff you did not compute.

How is this different from Course 47 on short selling?

Course 47 is the short: locate, borrow, buy-in, unlimited loss, squeeze as covering mechanics. This page is supply scoring — float, SI% of float, days-to-cover as a label, setup quality versus a guaranteed event. It is not a second borrow-how-to. If you cannot locate, you do not have a short; that sentence still lives in Course 47.

How do I size a squeeze candidate on a $10,000 account?

Same as any equity. 1% of $10,000 = $100 risk. Shares = floor($100 ÷ dollars to invalidation), then apply any notional cap and take the tighter number. In the EXAMPLE, Name TIGHT’s risk math wanted 66 shares at $18 with a $1.50 stop; the $1,000 notional cap allowed 55. Round down. The live risk calculator is crypto-UI; treat the unit as shares. A halt or a gap can skip $16.50. The SI% does not increase the $100.