Details
The numbers and where they come from
In the US, brokers report short positions to FINRA as of two settlement dates each month, around mid-month and month-end. FINRA publishes the totals about a week or so later. So a short interest figure is always a snapshot from the past.
From that one number, three ratios do most of the work:
- Days to cover =
short_interest/average_daily_volume. How many days of normal volume it would take for all shorts to buy back. - Short percent of float =
short_interest/ shares available to trade. How crowded the short side is. - Change = this report vs the previous one (
change_percent). Whether shorts are adding or covering.
Short volume is different. FINRA also publishes daily short sale volume, the share of each day's trades that were marked short. Market makers short constantly to fill buy orders, so daily short volume is often 40% or more of total volume for ordinary stocks. It says little about positioning. Our schema keeps the two apart with record_type.
Worked example (illustrative numbers)
A small-cap stock with 40 million shares in its float:
| settlement_date | short_interest | average_daily_volume | days_to_cover | short_percent_of_float |
|---|---|---|---|---|
| first report | 8,000,000 | 1,000,000 | 8.0 | 20.0% |
| second report | 10,400,000 | 1,300,000 | 8.0 | 26.0% |
Short interest rose 30%, but so did volume, so days to cover held at 8. The crowding (26% of float) grew more than the exit risk. Now suppose good news triples volume for a few days to 3.9 million shares. Days to cover drops to about 2.7, which means shorts can get out much faster. Squeezes usually need high short interest and low liquidity at the same time, so the shorts can't all leave through the door at once.
Code
Screen for squeeze setups in Python
"""Example: rank stocks by short crowding and exit difficulty."""
import pandas as pd
si = pd.read_csv("short_interest.csv", parse_dates=["settlement_date"])
si = si[si["record_type"] == "short_interest"]
latest = (si.sort_values("settlement_date")
.groupby("symbol").tail(1)
.copy())
latest["dtc"] = latest["short_interest"] / latest["average_daily_volume"]
screen = latest[(latest["short_percent_of_float"] >= 20)
& (latest["dtc"] >= 5)
& (latest["change_percent"] > 0)]
print(screen.sort_values("short_percent_of_float", ascending=False)
[["symbol", "settlement_date", "short_interest",
"short_percent_of_float", "dtc", "change_percent"]].head(20))
The screen keeps stocks where shorts hold at least 20% of the float, would need five or more days to cover, and are still adding. That list is a watchlist, not a trade list. The trigger for a squeeze is usually a catalyst (earnings, news, a sharp move on volume), which this data doesn't contain.
Watch out
Common mistakes
Treating short interest as current. By the time it is published it is a week or two old. Shorts may have covered since. Check settlement_date, not the date you downloaded it.
Reading daily short volume as short interest. A 50% short volume day is often normal market making. It doesn't mean half the trading was new bearish bets.
Ignoring the float definition. Different sources compute float differently (excluding insiders, locked-up shares or not). The same short interest can be 15% or 25% of float depending on the source. Compare like with like.
Assuming high short interest means a squeeze is coming. Many heavily shorted stocks keep falling, because the shorts were right. Short interest raises the stakes; it doesn't pick the direction.
Details
Get the data
Short interest and daily short volume, kept separate. What you'll get:
short_interest,previous_short_interest,change_percent,average_daily_volume,days_to_coverandshort_percent_of_floatper settlement date, plus dailyshort_volumeandtotal_volume.
- Short interest data, available in early access. Full history is planned for Starter ($29/mo). Suggested dataset price: $49 one-off for full short interest history.
- A free short interest sample is available on request.
FAQ
Frequently asked questions
What is short interest?
Short interest is the total number of shares that have been sold short and not yet bought back, reported as of a settlement date. It shows how many shares are owed by short sellers. Traders usually read it as a percentage of float or as days to cover.
How often is short interest updated?
In the US it is reported twice a month, for mid-month and month-end settlement dates, and published about a week or so later. Daily short sale volume is published every trading day but measures something different.
What is a high days to cover?
Above 5 to 10 days is often considered high, meaning it would take that many days of normal volume for shorts to buy back. What counts as high depends on the stock's usual liquidity.
What causes a short squeeze?
A squeeze happens when a rising price forces short sellers to buy back shares, and their buying pushes the price higher still. It needs heavy short interest, limited liquidity and a catalyst that starts the first move.
Keep reading