Short interest counts open short positions. Compare float and outstanding-share percentages, calculate days to cover, and avoid short-volume and publication-lag errors.
Short interest
Short interest is the number of a stock’s shares held in open short positions at a specified reporting date. It measures outstanding positions created through short selling, rather than the day’s short-sale activity. For a quant researcher, the useful questions are how large that position is relative to the stock’s tradable share base and trading volume, and when the information became available.
Three million shares can mean 15%, 10% or five days
Consider a fictional stock with the following inputs. All share counts refer to the same share class and use a consistent corporate-action basis.
Input
Assumed value
Reported shares short at the snapshot date
3,000,000 shares
Float under the chosen provider’s definition, at that date
20,000,000 shares
Total shares outstanding at that date
30,000,000 shares
Total share volume over the preceding 20 trading sessions
12,000,000 shares
Average daily volume over those 20 sessions
600,000 shares per session
Free float is the share base treated as available for public trading. It is smaller than total shares outstanding when the definition excludes restricted or closely held stakes. A float estimate is not a count of shares offered for sale at the current price, or of shares available to borrow. Check the provider’s exclusions and measurement date; a percentage from one provider need not be directly comparable with another’s. IBKR’s float-shares definition gives the basic trading-availability meaning.
The two percentage conventions answer different questions:
Short interest as % of float=10020,000,0003,000,000=15%,Short interest as % of outstanding shares=10030,000,0003,000,000=10%.
Neither calculation changes the 3 million shares short. A screen that says only “short interest: 15%” leaves the denominator unstated. A rising percentage can also reflect a shrinking float, so retain the raw share count as well as the ratio.
Days to cover divides the reported short position by average daily share volume:
Days to cover=average daily share volumeshares short=600,0003,000,000=5.
The units are trading-session volume equivalents. State the volume window: this example uses the preceding 20 sessions, a teaching convention rather than a universal vendor standard. Nasdaq’s Short Interest Report fact sheet uses volume over the current and previous reporting periods and describes its own rounding and display conventions.
Fictional 3 million shares short are 15% of 20 million float and 10% of 30 million outstanding. Holding shares short fixed, average daily volume of 300,000, 600,000 or 1,200,000 gives 10, 5 or 2.5 days to cover.
Original fictional comparison. The bars share a scale in millions of shares. The lower panel varies average daily volume while holding short shares fixed; a falling days-to-cover ratio need not mean anyone has covered.
Five days to cover does not mean shorts must close within five days, or that they could all exit at an unchanged price in five sessions. That interpretation would allocate the entire assumed daily turnover to covering, ignore other buyers and sellers’ responses and assume future volume resembles the past. Actual liquidity, available offers and price impact determine what can trade. With zero average volume, the mathematical ratio is undefined, not zero.
A position snapshot is not a transaction count
FINRA distinguishes short interest from its daily short-sale volume files. A short opened and closed within one day can contribute to that day’s short-sale volume without remaining in the later position snapshot. An older short can remain in successive snapshots without a new opening sale each day. Adding daily short-sale volume therefore does not reconstruct outstanding short positions. FINRA’s daily files also cover specified off-exchange activity, rather than all consolidated market trading. FINRA — Short Interest: What It Is, What It Is Not.
Likewise, reported share positions do not summarise every bearish economic exposure. A short exposure through derivatives is a different object from shares recorded short. A position’s size also does not identify its purpose: it may hedge another holding rather than express a standalone negative view.
The number has two clocks
The position date tells you what period the observation describes. Publication and receipt times tell you when a strategy could use it.
For the U.S. FINRA regime, checked 29 October 2026, firms report twice monthly. Rule 4560 specifies gross short positions at account level and the settlement-date basis; its scope and exclusions matter. The 2026 reporting calendar specifies filing by 6 p.m. Eastern on the second business day after the designated settlement date and separately lists publication dates. For example:
What the date means
Date in the checked FINRA calendar
Reporting settlement date
15 October 2026
Firms’ filing due date
19 October 2026, 6 p.m. Eastern
Publication date
26 October 2026
The public release thus describes positions from 11 calendar days earlier. These are dates from a specific U.S. reporting cycle, not a universal reporting or liquidation deadline.
FINRA’s October 2026 sequence: settlement snapshot on 15 October, firms’ filings due on 19 October by 6 p.m. Eastern, and publication on 26 October. A strategy cannot use that release before it becomes available.
Actual FINRA calendar example; schematic spacing. A publication date alone does not establish the intraday time at which a vendor or strategy received the release.
A backtest that attaches the released observation to 15 October and trades on it then introduces look-ahead bias. Preserve the snapshot date, publication time, vendor receipt time, corrections and the historically available float and volume inputs. Treat stale data as stale; a missing observation is not evidence of zero shorts. SelfMadeQuant’s Point-in-Time Data lesson develops the version and availability rules. Public preview; the full lesson requires membership.
Do not confuse FINRA’s broker reporting with the SEC’s separate institutional-manager Rule 13f-2/Form SHO regime. As checked on 29 October 2026, the SEC’s 3 December 2025 order grants temporary relief from Rule 13f-2 and Form SHO reporting effective 2 January 2026 and ending 2 January 2028. Older descriptions of the original compliance dates do not establish that those manager reports are currently available.
Crowding is a condition, not a forecast
A large position relative to float or normal turnover can motivate a crowding analysis. A short squeeze can arise when covering demand reinforces a price rise. But a large short-interest ratio supplies neither the probability nor the timing of that event. Shorts may be well funded, offset by other exposures or maintained for a long time. The negative information behind the positions may also be correct.
Interpret the snapshot alongside borrow availability, stock borrow fees, recall risk, trading conditions and the underlying catalyst. Reported short interest is not lendable inventory, and days to cover is not a broker’s margin or buy-in rule. The Short Selling lesson explains those loan and funding constraints. Public preview; the full lesson requires membership.
Check the denominator
A fictional stock has 2 million shares short. Its average daily volume increases from 400,000 to 1 million shares per session, measured over the same rolling-window convention. The next days-to-cover reading falls from five to two. Does this establish that short sellers bought back shares?
Answer. No. The numerator remains 2 million. The ratio changes from 2,000,000/400,000=5 to 2,000,000/1,000,000=2 entirely because volume increased. To assess a change in the outstanding short position, compare the dated share counts on a consistent corporate-action basis.
Sources
FINRA — Short Interest: What It Is, What It Is Not, 25 January 2023; accessed 29 October 2026. Position snapshots, transaction-volume differences, reporting coverage and provider-methodology cautions.
Nasdaq — Short Interest Report fact sheet, 2021, p. 2, “Data Fields and Definitions.” Short-share, average-volume and days-to-cover field conventions; a product definition, not a universal 20-session rule.
SEC — Release No. 34-104303, 3 December 2025, sections II–III; accessed 29 October 2026. Temporary relief from the separate Rule 13f-2/Form SHO reporting regime.
The share counts and volume sensitivities are original fictional illustrations. The FINRA dates are an actual calendar example; no security-specific short-interest observation or return forecast is presented.
