What Is Volume in Trading? How to Read It, With Real SPY Data

Michael Whitman · Trading since 2017, full-time since 2020 · YouTube creator, 137K subscribers · Published 2026-09-05 · Updated 2026-11-02 · Reviewed 2026-11-01

Volume is the number of shares or contracts that changed hands during a period, drawn as a bar under each candle. It is the one common chart input not calculated from price, so it shows how much trading stood behind a move but never which side won it.

How it works

Each trade adds its size once. If one trader buys 300 shares from another, the volume for that bar rises by 300, not 600. A buy and a sell are the same event seen from two sides, which is why the phrase “more buyers than sellers” never describes a volume bar: the count of shares bought and the count sold are always equal.

The bar’s height is the total for that period. On a daily chart it is the whole session; on a five-minute chart it is five minutes. Stocks and funds count shares, futures and options count contracts, and spot forex has no central record at all, so a currency chart’s “volume” is usually one broker’s tick count rather than a market total.

What it adds is a second measurement. A candle shows how far price traveled. The bar underneath shows how much business it took to get there. Moving averages, the relative strength index (RSI) and most other tools are arithmetic on price, so volume is the only independent reading most charts carry.

How to read a volume bar

Compare it with the bars around it. A raw share count means nothing until it has a baseline: a million shares is a busy day for a small company and a quiet five minutes for SPY. The usual yardstick is relative volume: today’s volume divided by a recent norm, often the median or average of the previous 20 sessions. A reading of 1.0 is a normal day; 2.0 is twice normal.

Use the same time of day when reading intraday bars. A five-minute bar at 9:30 a.m. and one at 1:50 p.m. live in different worlds, as the data below shows, so an opening bar several times taller than the lunchtime bars is ordinary rather than news.

Then ask what the bar sits under. A tall bar under a big candle says many participants were active while price moved far. A tall bar under a small candle says plenty traded and price went nowhere, which can happen on options expiration days and index rebalances, when large orders cross for reasons that have nothing to do with the chart.

A worked example

Friday 30 October 2026 on SPY. The fund opened at $761.38, traded between $750.36 and $761.46, and closed at $755.72, down 1.77% from Thursday’s $769.36. Yahoo Finance records 77,625,800 shares traded.

Step one, the baseline. The 20 sessions before it ran from 2 October to 29 October. Sorted, the two middle values are 35,048,400 and 36,054,300 shares, so the median is 35,551,350.

Step two, the ratio. 77,625,800 ÷ 35,551,350 = 2.18. Using the 20-session average instead (37,659,235 shares) gives 2.06. Either way the day traded about twice its recent normal.

Step three, the context. The two days before it were 0.81 and 1.13 times normal, and 28 October, the day the Federal Reserve announced its rate decision, was one of the quiet ones. Yet 2.18 times normal made 30 October only the 50th-busiest of 2026’s 209 sessions, because the year had heavier stretches. Measured against its own previous 20 sessions, though, no 2026 session had a higher ratio. At the close, those shares were worth about $58.66 billion.

What the reading says, and what it does not. It says the fall happened with far more participation than the month’s typical day. It does not say sellers outnumbered buyers, because each of the 77.6 million shares had one of each. Nasdaq’s own record shows 77,696,480 shares, within 0.1% of Yahoo’s figure, so the reading does not depend on the data feed.

When volume runs heavy

The edges of the session. Across 77 sessions of SPY five-minute bars from 10 July to 30 October 2026, the first bar of the day (9:30 to 9:35 a.m.) carried a median 6.18% of the day’s five-minute total and the last bar (3:55 to 4:00 p.m.) carried 8.49%. If trading were spread evenly across the 78 bars, each would carry 1.28%. The quietest bar was 1:50 p.m., at 0.56%, so the last bar was about 15 times busier than the slowest one.

Half an hour at each end outweighs two hours in the middle. The first 30 minutes carried a median 14.23% and the last 30 minutes 21.01%, while the two hours from noon to 2:00 p.m. carried 17.74%. And the bars still miss part of the day: summed, the 78 bars came to a median 84.47% of Yahoo’s daily total. The closing auction at 4:00 p.m. and trading outside regular hours are not in regular-session bars and would account for at least part of the gap.

The days of the week. Over the ten years from 31 October 2016 to 30 October 2026, 2,514 sessions, each day’s volume was divided by the median of the 20 sessions before it. The median of that ratio was 0.90 on Mondays, 0.93 on Tuesdays, 1.00 on Wednesdays, 1.01 on Thursdays and 1.08 on Fridays. In 426 full five-session weeks, Friday was the busiest day 144 times and Monday 70 times, where an even split would give each about 85.

Options expiration explains part of Friday, not all of it. The 116 third Fridays of the month, when standard monthly options expire, had a median ratio of 1.29. The other 389 Fridays still came in at 1.03, above every Monday-to-Thursday median. A triple witching Friday is the extreme case of the same calendar effect. Nasdaq’s volume record gives the same order, 0.90 on Mondays and 1.08 on Fridays.

The original data

The breakout test. A common rule says a breakout on above-average volume is more significant than one on light volume. To test it on SPY, a breakout here is a daily close (adjusted for dividends) above the highest close of the 20 sessions before, counting a close only when none of the five sessions before it was also one. Testing every session from 8 March 1993, the first with 25 sessions of history, gives 314 breakouts from 19 May 1993 to 5 October 2026, each followed for 10 sessions. Each was sorted by its relative volume: that day’s shares divided by the median of the 20 sessions before.

Heavy breakouts did not do better. Breakouts on 1.5 times normal volume or more were higher 10 sessions later 53.12% of the time, and lighter-than-usual breakouts 65.50% of the time, against 61.81% for any session. The heavy group’s median change, +0.87%, was close to the +0.70% of any session. With 32 heavy cases, each rate sits within about one standard error of the any-day figure, a gap that chance alone produces easily, so the honest reading is no measurable difference, not a reversal of the rule.

The broken level held about as often either way. The share still above the broken high was close to 70% in every group. This is one index fund on daily bars, where a breakout is a move of the whole market; single stocks breaking out on news may behave differently. Every breakout is in the SPY breakout file.

What counts as heavy drifts over years. SPY’s median session traded 256.7 million shares in 2008 and 54.6 million in 2026 up to 23 October, across 8,492 sessions since January 1993, so a fixed share count goes stale and only a recent baseline works. Measured in dollars the order flips, because each share now costs far more.

The biggest days lean toward falling days. Of SPY’s 50 heaviest sessions since 1993, 31 (62.00%) closed lower, against 45.25% of all sessions, and their median move either way was 4.13% against 0.53% for a typical day. The record was 10 October 2008, at 871,026,300 shares. A huge bar marked a big move far more reliably than it marked a direction. The 50 heaviest sessions and the yearly medians are published, and the weekday ratios are in the daily relative volume file.

How the data was checked. Nasdaq’s SPY history covers 2,512 of the same sessions; its daily volume was within 1% of Yahoo’s on 2,393 of them, and every session since 1993 is in the daily SPY volume file. Yahoo’s five-minute download of 1 November 2026 had 27 empty bars on 21 sessions, each followed by a bar holding the session’s running total of shares so far, so the time-of-day figures use an earlier clean download for 10 July to 2 October and drop the three later sessions with the fault (6, 8 and 26 October). The time-of-day file lists every bar’s median share.

When it fails

A tall bar read as buying

This is the most common misreading, and the arithmetic rules it out. Every share in a bar has a buyer and a seller, so a tall bar on a rising candle is not “buyers stepping in” any more than it is sellers stepping out. It says the move mattered to many participants, and on SPY’s heaviest days the move was more often down than up.

Volume used as the confirmation that decides a breakout

On SPY’s daily chart it did not separate the breakouts that kept going. The heavy group was the smallest, and it was higher 10 sessions later less often than an average day, within chance. A trader who skipped every light breakout would have skipped the group that rose most often. Volume can still describe how a level broke; it is not a filter that has earned trust on this evidence.

The wrong baseline

Comparing across times of day, weekdays or years produces false alarms. The 9:30 a.m. bar is always tall, Friday is usually busier than Monday, an expiration Friday is busier still, and the 2008 normal would look like a panic today. Relative volume only works when the baseline matches the bar being judged.

A feed that is quietly wrong

Volume data has more faults than price data. The empty-bar problem above appears in a free download that looks complete at a glance, intraday bars miss the closing auction, and forex volume is one broker’s count. A wrong baseline produces a confident wrong reading.

Climax bars found afterward

The largest bar of a run often prints near its end, when everyone who meant to act has acted. At the moment it prints there is no telling it apart from the first bar of a bigger move; only hindsight sorts them, which is why exhaustion is easier to name than to trade.

Questions people ask

What does 1 volume mean in trading?

One share of a stock or fund, or one contract of a future or option, changing hands once. In forex, a platform’s “1 volume” usually means one lot, which many brokers set at 100,000 units of the base currency; the broker sets it, not an exchange.

What volume is good for trading?

There is no universal number. What matters is enough trading for a tight bid-ask spread and a reading compared with the same instrument’s recent sessions. SPY’s 35.6 million shares was a normal day in October 2026; in 2008 it would have been unusually quiet.

Is higher trading volume good?

For getting in and out, yes: busier markets usually have narrower spreads. As a signal, not by itself. On SPY, heavy breakout days were no more often followed by gains than light ones, and the heaviest sessions on record leaned toward falling days.

Does volume show buyers or sellers?

Neither. Each traded share has one buyer and one seller, so volume measures how much changed hands. Tools such as accumulation distribution guess at direction by where the close sits in the bar, which is an estimate layered on top.

Related

Candlesticks is the page above this one: the bar only means something once you can read the candle it sits under.

Support and resistance is where volume is most often read, because a level that drew heavy trading is a different kind of level from one price drifted through.

Market structure supplies the price at which any volume read is wrong, which a volume bar cannot do on its own. And VWAP uses volume as a weight to give the average price paid across the session.

This page is educational, not financial advice. Test every idea on your own charts before risking money.