Volume in Futures Explained: What It Measures, Why It's Cleaner Than Stock Volume and What It Can't Tell You
One print of 50 contracts is 50 volume and one tick. Why that distinction matters in futures, why NQ volume is cleaner than any stock's, what open interest adds, and four situations (roll week, time of day, news, data feeds) where the histogram lies.
Volume is the number of contracts that changed hands in a bar. One trade of 50 contracts adds 50 to the volume and 1 to the tick count; fifty trades of one contract each add the same 50 volume and 50 ticks. That distinction matters more in futures than anywhere else, because the bars on your chart can be built from either number, and because volume on an index future means something more specific than volume on a stock. If you're still getting comfortable with what a bar is, start with how to read a price chart and come back.
What one unit of volume is
Every futures trade has exactly one buyer and one seller, and the exchange counts the contract once. A print of 10 contracts means 10 contracts moved from one account to another; the print itself carries no "bought" or "sold" label. This is the first thing volume can't tell you on its own: which side was the aggressor. To get that, the feed has to record whether the trade hit the ask (a buyer lifted the offer) or the bid (a seller hit it), and that classification is what delta, footprint charts and the Bookmap heatmap are built on. Plain volume just counts.
The second thing to separate is trades from contracts. NinjaTrader's tick bars (a 2000-tick bar, say) close after 2,000 trades regardless of their size, its volume bars close after a set number of contracts, and time bars close on the clock; the bar types post walks through each. The practical consequence is that a tick chart speeds up when activity is many small orders and a volume chart speeds up when size shows up, and the two disagree most around the open and around news.
Why futures volume is cleaner than stock volume
A share of a large US stock can trade on more than a dozen exchanges, in dark pools, and inside a broker that matches the order against its own flow before it ever reaches a public venue. The volume you see on a stock chart is a consolidated tape stitched together after the fact, and a large share of it printed somewhere you could not have traded.
NQ trades in one place. The E-mini Nasdaq-100 has one order book on CME Globex, every trade prints to that book, and the volume in your NinjaTrader chart is the volume every other participant sees, with the same time stamps. There is no off-exchange NQ. The same holds for MNQ, the micro, which has its own book and its own volume, so comparing the two means comparing two markets, not one market reported twice. This is why volume-based tools (profiles, footprints, delta) are far more common in futures than in equities. The input is complete.
Volume, ticks, delta, open interest
| Number | What it counts | Where you see it | What it answers |
|---|---|---|---|
| Volume | Contracts traded in the bar | Volume indicator, volume bars, profiles | How much business, and where |
| Tick count | Individual trades in the bar | Tick bars (100T, 2000T) | How busy, regardless of size |
| Delta | Volume at the ask minus volume at the bid | Footprint, Bookmap, order flow tools | Which side was aggressive |
| Open interest | Contracts still open at the end of the day | CME daily report; some platforms, next day | Whether volume opened or closed positions |
Open interest is the row most people skip. Volume tells you how many contracts traded today; it doesn't say whether those trades opened new positions or closed old ones. Open interest, published by CME after the session, answers that, which is why it swings in the week before expiration when positions roll to the next contract.
Four places volume misleads
Roll week. Every quarter, volume migrates from the expiring contract to the next one over a few days around the rollover date. If your chart is still on the September contract, volume looks like it's dying; if you've just switched to December, the last three months look thin because most of that history traded on September. Neither is a market signal. NinjaTrader's merged contract stitches the two histories at the roll date, so the join looks clean; the few days when both contracts are active are still split between them.
Time of day. A 500-contract print in NQ at 3:00 AM ET / 09:00 CEST is a large order. The same print at 9:31 AM ET / 15:31 CEST is noise. Volume only means something against the same window on other days, and the sessions post shows how uneven the Globex day is: a spike in the European morning is measured against European mornings, not against the US open.
News. The minute after a high-impact release carries volume that includes stop-outs, forced liquidations and algorithms repricing everything at once. It is real volume, and it says nothing about whether anyone agrees on a direction. Treat post-release volume as a fact about the release, not about the level it happened at.
Feed differences. Some data feeds bundle trades that arrive in the same millisecond into one tick; others report each fill. The volume total comes out the same, the tick count doesn't, which is why a 2000-tick bar on two feeds can close at different moments. If your tick chart looks different from a friend's, check the feed before checking the strategy.
What to do with it
Volume works as context rather than as a trigger. A volume profile, whether from TradingView's free one or a paid add-on, tells you where the day's business was done: the point of control and the value area are volume sorted by price instead of by time. A volume spike at a key level says the level got attention; it says nothing about who won. A quiet bar at a level that "should" react is information too. Nobody showed up.
On my execution chart there is no volume at all. It lives in the profile I build in the morning and in the heatmap, and the one question I ask of it is whether a level is a place where the market did its business or somewhere it passed through. The bar-by-bar histogram doesn't answer that. The profile does.