Trade size distribution is a breakdown of the sizes of individual executed trades in a security over a period, showing how many trades were small (odd-lot or retail-sized) versus how many were large (block-sized, more typical of institutional activity). Traders sometimes use it to infer whether retail or institutional-sized participants appear more active in a name, but trade size alone is an imperfect proxy -- institutions routinely split large orders into many small pieces to avoid showing their true size.
Key Takeaways
- Trade size distribution sorts every executed trade in a security over a period into size buckets, from small odd-lot prints to large block trades.
- It is built directly from the tape of executed trades (prints), not from quotes or order-book depth.
- A shift toward more large trades is commonly cited as a sign of increased institutional-sized activity, and a shift toward small trades as a sign of more retail-sized activity.
- The signal is imperfect: institutions frequently break large orders into many small child orders to reduce market impact, which can hide real institutional flow inside the small-trade buckets.
- Trade size distribution is best used alongside other order-flow and volume tools, not read in isolation.
What Is Trade Size Distribution?
Trade size distribution is a breakdown of the sizes of individual executed trades over a period -- for example, how many trades were small (odd-lot or retail-sized) versus how many were large (block-sized, more typical of institutional activity). Instead of looking only at total volume for a stock or a single average trade size, this view separates every print on the tape by how many shares it represented, then groups those prints into size ranges so a trader can see where activity is concentrated.
The idea sits alongside other tape-reading and volume-based tools covered in the Quotes, Spreads & Liquidity section: quotes show where participants are willing to trade, and trade size distribution shows how big the trades that actually happened tend to be. Traders sometimes use trade size distribution to infer whether retail-sized or institutional-sized participants appear more active in a given security, though trade size alone is an imperfect proxy for that question.
How Trade Size Distribution Is Built
Trade size distribution is constructed from the raw tape of executed trades for a security over a chosen period, such as a single trading day. The process has three basic steps:
- Collect every print. Each individual executed trade -- a single match between a buy order and a sell order -- is recorded along with its share size.
- Sort trades into size buckets. Each trade is assigned to exactly one size range. A simple version might use two broad buckets -- small (odd-lot or retail-sized) trades versus large (block-sized) trades; more detailed versions can use finer or dollar-value-based buckets instead of raw share counts.
- Aggregate and compare. The number of trades (and often the total volume) in each bucket is summed and can be expressed as a percentage of all trades or all volume for the period, making it possible to compare how much activity came from small prints versus large prints.
Because it is built purely from executed trades, trade size distribution differs from measures built off quoted size or order-book depth -- it describes what actually traded, not what was merely offered or bid at any moment.
Worked Example
Hypothetical example -- for education only. Suppose a trader pulls every executed trade for a stock during one trading session and sorts them into small and large size buckets. A simplified summary might look like this:
| Size bucket | Number of trades | % of all trades | % of total volume |
|---|---|---|---|
| Small (odd-lot/retail-sized) | 6,300 | 87% | 31% |
| Large (block-sized) | 920 | 13% | 69% |
In this hypothetical table, small odd-lot trades make up the majority of the trade count (87%) but a smaller share of total volume (31%), while block trades are only 13% of the count but 69% of the volume. That gap between trade count and volume share is typical: a relatively small number of large prints can move a large fraction of the day's shares. A trader reviewing this table might note that a meaningful portion of volume came from larger prints, while being careful not to assume every one of those larger prints came from a single institutional order rather than, say, an algorithm intentionally sizing its child orders to look small.
How Traders Use It
Traders sometimes reference trade size distribution as one input when trying to gauge the mix of participants active in a security:
- Gauging participant mix. A distribution skewed toward small, odd-lot-sized trades is sometimes read as a sign of more retail-sized activity, while a distribution with a larger share of block-sized volume is sometimes read as a sign of more institutional-sized activity. This is a commonly cited heuristic, not a precise measurement of who is actually trading.
- Context for a volume spike. When volume rises sharply, checking whether the increase came from many small trades or from a handful of large ones can add context to what kind of activity is driving the move.
- Comparing across sessions. Watching how the distribution shifts from one day or week to the next can highlight changes in the pattern of participation over time, rather than relying on a single day's snapshot.
None of these uses amount to a reliable read on trader intent or a guarantee about future price direction. Trade size distribution is best treated as descriptive context, used alongside other order-flow and liquidity tools rather than as a standalone signal.
Limitations and Common Mistakes
- Order-splitting hides true size. Institutions can and do break large orders into many small pieces specifically to avoid showing large trade sizes, so real institutional flow can appear as a stream of small prints rather than one large block.
- No universal size thresholds. What counts as "small" or "large" varies by data provider, by the price and typical liquidity of the security, and by convention -- the same trade can land in different buckets depending on which definition is used.
- Count versus volume can tell different stories. A bucket with a high trade count is not the same as a bucket with a high share of total volume; conflating the two can lead to a misleading read of where activity is concentrated.
- Not a substitute for order-book or quote data. Trade size distribution describes what already executed. It does not show resting liquidity, cancellations, or the depth of the book, all of which are separate considerations covered elsewhere in this section.
- Not a timing or direction signal on its own. A shift in the distribution does not indicate whether the underlying activity is buying or selling, only that the mix of trade sizes has changed.
Why Trade Size Is a Weak Signature
The inference this display invites is that large prints indicate large participants and small prints indicate small ones. That inference has been unreliable for a long time, because orders of any size are routinely broken into pieces before execution, which means a substantial order can arrive as a long series of ordinary-looking trades.
What the distribution does describe reasonably well is how fragmented execution was over a period. A shift toward smaller prints is a fact about how trading was conducted, and it is a much weaker basis for conclusions about who was doing it.
Comparisons across securities are particularly fragile. A price of a few dollars and a price of several hundred produce very different share counts for the same capital, so a distribution measured in shares is partly measuring the price.
Venue coverage completes the caveats. Activity executed away from the venues your feed carries, or reported under different conventions, is either absent or arrives in a form that distorts the shape.
Frequently Asked Questions
What is trade size distribution?
Trade size distribution is a breakdown of the sizes of individual executed trades in a security over a chosen period, such as a trading day. It groups trades into buckets -- for example, odd-lot or retail-sized trades versus block-sized trades more typical of institutional activity -- so a trader can see how order flow is split across the two ends of the size spectrum, rather than just looking at total volume.
How is trade size distribution calculated?
It is built by taking every executed trade (print) in a security over a period, recording its share size, and sorting those sizes into buckets, such as small (odd-lot or retail-sized) versus large (block-sized) trades. Each trade falls into exactly one bucket. The count or total volume in each bucket is then compared, often as a percentage of all trades or all volume, to see where activity is concentrated.
Does a high number of large trades mean institutions are buying?
Not reliably on its own. A cluster of large, block-sized prints is commonly cited as a sign that institutional-sized participants are more active in a security, but trade size alone is an imperfect proxy. Institutions frequently break large orders into many small child orders specifically to reduce market impact and avoid showing their true size, which means a large real institutional order can appear in the data as dozens or hundreds of small trades instead of one large one.
What counts as a large trade versus a small trade?
There is no single official cutoff, and thresholds vary by data provider, security price, and typical liquidity in that name. Traders commonly treat trades of roughly 100 shares or fewer as odd-lot or retail-sized, and trades in the thousands of shares (or a dollar-value equivalent) as block-sized. Because these thresholds are conventions rather than fixed rules, the same trade can be classified differently depending on which dataset or definition is used.
Can trade size distribution be manipulated or misleading?
It can be misleading even without any intent to manipulate anything. Order-splitting algorithms used by institutional desks are designed to disguise large orders as a stream of small ones, which pushes true institutional flow into the small-trade buckets and understates it. Retail brokers also batch and route orders in ways that can affect print sizes. Traders who rely on this data should treat it as one imperfect signal among several, not a direct read of who is trading.
Is trade size distribution the same as volume profile?
No. Volume profile shows how much trading volume occurred at each price level, answering where trading happened. Trade size distribution shows how many trades (or how much volume) occurred at each individual trade size, answering how big each execution was. They are both built from the same underlying tape of executed trades but organize that data along different axes and are used to answer different questions.
How does order-splitting by execution algorithms affect the distribution?
Algorithms break large orders into many small executions specifically to avoid signalling size, which moves institutional activity into the small-trade portion of the distribution. The result is that the distribution reflects execution technique as much as participant type. A shift toward smaller average trade sizes over time is consistent with more algorithmic execution rather than with a change in who is trading.
What does a bimodal distribution suggest?
Two clusters, one of very small trades and one of much larger ones, is a common shape and generally reflects two different execution styles operating in the same security rather than two groups of participants with different views. Blocks negotiated and printed separately produce the upper cluster; continuous algorithmic execution produces the lower one. Describing the shape is more defensible than inferring intention from it.
How should the distribution be normalized for comparison across securities?
Absolute share counts are not comparable between a security trading at a few dollars and one trading at several hundred, because the same dollar commitment produces very different share quantities. Expressing trade sizes in dollar terms, or as a proportion of the security's own average trade size, puts them on a common basis. Comparing raw share distributions across price levels mostly compares the prices.
Related Reading
- Quotes, Spreads & Liquidity -- the parent hub covering how quotes, spreads, and liquidity shape trade execution.
- Market Structure & Trade Execution -- the broader hub on how orders move through markets.
- Liquidity Gaps, Thin Books, and Price Discontinuities -- how thin order books can cause abrupt price moves.
- Bid Price vs. Ask Price: How Quotes Actually Work -- the mechanics of the quotes that trades execute against.
References
- CMT Association -- professional body for chartered market technicians, covering technical and market microstructure analysis methodology.
- CFA Institute Research and Policy Center -- research on market microstructure and trading behavior.
- TA-Lib documentation -- open-source reference for volume- and trade-based technical analysis calculations.
- SEC Investor.gov -- U.S. Securities and Exchange Commission investor education resources on market structure and trading.