Author:
Bob Sloan, Founder and Managing PartnerShort interest aggregation is not a single number. Two families of metrics exist: SI% float aggregations, which represent the percentage of a company's float that is sold short, and dollar metrics, which refer to the total dollar value of shares sold short. Each answer is a different kind of question about how short sellers are positioned across an index or sector, forming the basis of broader short interest analysis.
SI% Float aggregations describe distribution and character. Mean, median, and notional weighted average each tell a different part of the story. The gaps between them are critical for analysis.
Dollar metrics describe scale and stress. Total notional short interest and dollar short P&L together identify whether the money behind a sector's short positioning is growing, profitable, or under pressure.
The S3 framework turns short interest data into a multi-dimensional signal, identifying where skepticism is concentrated, how unusual it is historically, and whether the short base is in control or under stress. This framework acts as a short squeeze indicator for identifying potential pressure points in crowded trades.
Characterizing a sector or index requires aggregating short interest across hundreds or thousands of individual stocks, and aggregation requires choices. A sector average of 5% SI% Float can mean very different things depending on whether that average is driven by a handful of heavily shorted names or spread evenly across the index. A rising total notional can reflect new short selling or simply price appreciation on existing positions. These distinctions require more than a single headline number to resolve.
This framework introduces the tools S3 uses to make those distinctions: three SI% float aggregation methods, two dollar-based metrics, and a z-score normalization layer that places any reading in a historical context. The Russell 3000 sector snapshot below shows all five metrics applied simultaneously, a starting point for identifying where positioning is concentrated, how it compares to history, and where the stress points are.
The foundation of S3's sector framework is short as a percent of float aggregation. Short percent float measures how much of a stock's tradeable supply has been sold short. At the stock level, it is intuitive. This metric is commonly referred to as SI % Float. At the index level, the question becomes how to collapse hundreds of individual readings into a number that meaningfully represents the group, and the answer depends on what you want to know.
Three aggregation methods each answer a different question. The mean gives you the overall level but is sensitive to a small number of heavily shorted names, inflating the average. The median filters those outliers out and tells you where the typical stock sits. The weighted average anchors the calculation to notional dollars so that stocks carrying the largest short positions receive proportionally more weight.
The three metrics are most informative when read together. A large mean-to-median gap indicates a skewed distribution, a handful of crowded names pulling the average above where most of the index actually sits. When the weighted average runs well above the median, the largest short positions are concentrated precisely in the most heavily shorted percent of float names.
Three aggregation methods, what each measures, its signal, and how to read them together.
Metric
What It Measures (Definition)
Primary Signal (Key Use)
Interpretation (Reading the Number)
Mean SI% Float (Simple average)
SI% Float summed across all constituents and divided by the number of stocks. Each name counts equally regardless of size.
The headline level of short interest in the index — but sensitive to outliers. A few names at 50%+ SI% Float can pull it well above where most stocks sit.
When mean exceeds median, a right-tail exists — a subset of heavily shorted names is inflating the average. The gap between mean and median shows the degree of skew.
Median SI% Float (Midpoint value)
The midpoint of the SI% Float distribution. Half of stocks are above this level, half below. Outliers at either extreme do not move it.
Where short interest stands for the typical stock in the index — unaffected by a handful of crowded names at the tail.
The reference point for other metrics. Is the mean running above it? Outliers matter. Is the weighted average above it? The biggest short positions are the most crowded.
Weighted Avg SI% Float (Notional-weighted)
Each stock’s SI% Float weighted by its share of total index notional short interest. Larger positions carry more weight.
The average SI% Float behind a dollar of short exposure — where the money is positioned, not just where names happen to be shorted.
Above median: the largest short positions are concentrated in highly-shorted names. Below median: large notional exposure is spread across names with modest float utilization.
Table 1: Three methods for aggregating SI% Float at the index and sector level: mean, median and notional short weighted average.
The simple average is the most common headline figure. Each stock counts equally regardless of market cap or position size. This makes the mean easy to calculate and easy to communicate, but it creates a meaningful vulnerability; a single biotech stock at 60% short percent of float, an extremely high reading, can drag an entire sector average well above where the vast majority of names actually trade. The mean is most useful as a starting point, a signal that something warrants investigation rather than a conclusion in itself.
The median is the midpoint of the distribution. Half the stocks in the index are above it, half are below it, and the value of an extreme outlier at either end makes no difference. This outlier-resistance makes the median the most stable single-number description of where short interest stands for the typical stock in an index. It is also the natural reference point against which the other two measures should be read.
The weighted average assigns each stock a weight equal to its share of the total index notional short interest. The result is an average that reflects where short dollars are actually deployed, not simply where names are the most short as a percent of float. A stock with $10 billion in notional short interest receives far more weight than one with $200 million. When the weighted average runs above the median, it signals that the largest short positions are concentrated in the most heavily shorted names, a pattern associated with elevated crowding risk and potential short squeeze indicators in heavily shorted names.
The mean tells you the level. The median tells you the typical stock. The weighted average tells you where the money is. Used together, they reveal whether a sector's short interest is broadly distributed or concentrated in a handful of names.
SI% Float describes the distribution of short positioning and changes in positioning as a result of adding or reducing the number of shares short. Dollar metrics describe its economic scale and the current stress on the short base, often referred to as notional short interest. Two measures matter: total notional short interest, which captures the absolute size of exposure, and dollar short P&L, which captures how that exposure is performing.
Total notional is shares short multiplied by current price, summed across all index constituents. It measures the absolute economic size of short exposure in dollar terms. One critical nuance: because price is a direct input, notional changes with the market, independent of any change in position size. A sector rally can inflate notional even if no new short is added; a sell-off can deflate it even if no one covers. The right diagnostic is SI% Float; if notional rises but SI% Float is flat, the price moved. If both rise, shorts are genuinely building.
Dollar-short P&L is the mark-to-market gain or loss on all outstanding positions, which short sellers are currently making or losing on their deployed capital. It moves inversely to price, which means it naturally complements notional: when a sector rallies, notional climbs and P&L turns negative simultaneously, both signals tightening at once. A large notion combined with a deeply negative P&L represents the most acute squeeze configuration. Shorts are underwater, heavily exposed, and covering can be forced regardless of fundamental convictions, one of the classic indicators of short squeeze conditions.
Short Interest Dollar Metrics
Two dollar-denominated metrics, what each measures, the signal it carries, and how they work together.
Metric
What It Measures (Definition)
Primary Signal (Key Use)
Interpretation (Reading the Number)
Total Notional Short Interest (Summed exposure)
Shares short times price, summed across the index. Price is a direct input — notional rises when stocks rally and falls when they drop, even if no one adds or covers.
How large is the short position, and is it growing or shrinking? Rising notional is often mistaken as a rise in shorting, but can simply reflect price inflation on existing positions.
Check SPX float before drawing conclusions. Notional up & SPX float: price move. Both up: shorts are genuinely building. Both down: real covering is happening.
$ Short P&L (Mark-to-market)
The gain or loss on all outstanding short positions at current prices. When stocks rise, shorts lose money. When stocks fall, they profit.
Shows whether shorts are in profit or pain. Moves opposite to price — when a sector rallies, notional climbs and P&L turns negative simultaneously. Both signals together explain positioning.
Negative P&L on large notional: shorts are underwater and exposed to squeeze risk. Positive P&L on rising notional: profitable shorts — indicates added size, not distress.
Figure 2 — Two dollar-denominated metrics for measuring short exposure and performance: definition, key signal, and how to read them together
Raw SI% float aggregates and dollar metrics are informative in isolation but difficult to compare across sectors and over time. A 5% weighted average SI% Float is elevated in utilities, where structural short interest is low, and unremarkable in biotechnology, where crowding is routine. Z-scores resolve such issues by normalizing each metric to its rolling history, expressing the current reading as a number of standard deviations above or below the historical mean through z-score normalization.
At the index level, a z-score above 2 indicates historically elevated positioning, a potential signal of crowding and squeeze vulnerability. Below −1.5 suggests unusually light positioning, which can reflect prior capitulation or a sector the short community has moved on from. Z-scores on notional and P&L identify whether the absolute dollar scale of short exposure is historically unusual. At the constituent level, the share of stocks with individual z-scores above 2 measures crowding breadth
— a harder-to-compress signal than any single aggregate.
The table below applies the full framework to the Russell 3000 and its eleven GICS sector sub-indices: mean, median, and notional-weighted average SI % Float describes the character and distribution of short positioning within each index. Total notional and dollar P&L describe economic scale and current stress on the short base.
The z-score column normalizes the weighted-average SI % Float to a 24-month rolling window, helping researchers evaluate short interest and aggregate stock returns across sectors.
How to read this table: a positive z-score alongside negative P&L on large notional is the highest-risk configuration for a squeeze
— shorts are elevated relative to history, underwater, and exposed.
Russell 3000 Sector Short Interest Aggregation
Index
Mean SI% Float
Median SI% Float
Wtd Avg SI% Float
Notional ($B)
$ P&L ($B)
Z-Score (Wtd Avg SI%)
R3000 Consumer Disc.
9.5%
7.2%
6.3%
$267
$7.6
0.3
R3000 Energy
9.1%
6.5%
6.7%
$77
-$5.2
-0.3
R3000 Technology
9.0%
6.7%
4.5%
$430
$18.2
0.2
R3000 Consumer Staples
7.7%
5.9%
5.5%
$69
-$1.9
1.7
R3000 Telecom
7.6%
5.8%
10.2%
$53
-$4.3
1.8
R3000 Basic Materials
6.5%
4.6%
7.3%
$45
$0.1
1.5
R3000 Industrials
5.9%
4.2%
4.7%
$240
-$2.4
0.9
Figure 3 — Russell 3000 sector short interest aggregation as of 4 March 2026: mean, median, and notional-weighted average SI% Float, total short notional, trailing one-month mark-to-market P&L, and one-year z-score on weighted average SI% Float.
Want to know more? Access this data in real time using S3’s BLACK APP & BLACK MAP