Why Your Footprint Shows a Different Volume Than Your Broker (and Which Number to Trust)

Look at one contract, one instant. Your footprint column totals one number. The candle it sits on shows another. Your broker’s DOM has a running figure that agrees with neither, and TradingView in the next tab shows a fourth. Same symbol, same second.

None of them is broken. They are four correct answers to four slightly different questions, and the trouble only starts when you assume there is supposed to be one. This is where each number comes from, and which one your order flow should actually be built on.

There is no "volume" — there are at least four

"Volume" is a single word doing four jobs. On a live order-flow screen you are usually looking at all four at once without noticing:

• Candle volume — every contract traded in that bar. If the bar was built tick-by-tick, this is a hard count. If it was built from a minute feed, it is whatever that feed reported for the minute.
• Footprint volume — the same contracts, but split into bid volume + ask volume per price. Add the two sides back up and you should land on the candle volume. When you don't, the split and the total came from different data paths — see the next section.
• DOM volume — a running tally the broker ladder keeps, often cumulative for the session and reset on the broker's own clock. It is a different accumulator, not a different market.
• Exchange / settlement volume — the official figure the exchange publishes for the trading session, after it reconciles spreads, blocks and corrections. It is the most authoritative and the least useful in real time, because it lands after the close.

Four numbers, four definitions. The mismatch you are staring at is almost always two of these being compared as if they were one.

The biggest gap is tick replay versus minute bars

This is the one that catches everyone, and it hides in your chart settings rather than your broker. A footprint needs, for every trade, which side was the aggressor — was it lifting the ask or hitting the bid. That information exists only in the individual trades. A one-minute bar has already thrown it away.

So a platform has two ways to draw a historical footprint:

• Replay the ticks — pull the individual trades, rebuild the bars, keep the real bid/ask split. Correct, heavy.
• Estimate from the minute bar — take OHLCV and guess how the volume split between bid and ask. Cheap, and wrong at exactly the moments you care about (absorption, a violent one-sided push).

This is also why our bridge reads NinjaTrader's own tick stream rather than a rebuilt minute feed: the trades it hands us already carry the aggressor side, so there is nothing to estimate. If you want the mechanics of why that feed is missing entirely on some accounts, that is a separate entitlement problem.

The session boundary decides the daily total

Two platforms can count the exact same trades and still print a different daily volume, because they disagree on when the day starts. Futures do not run on your calendar. A CME equity-index contract trades almost around the clock, and the exchange trading day rolls at 17:00 Central Time — the session runs 17:00 to 16:00 CT with a maintenance break in between. The volume the exchange attributes to a day is the volume of that session.

A platform that resets its daily counter at local midnight, or at 00:00 UTC, or at the equity cash open, is summing a different window of the same tape. Nobody miscounted; they drew the day's edges in different places. Before you trust any "daily volume" comparison, check that both tools cut the session at 17:00 CT — most order-flow mismatches at the day level are this and nothing else.

Two feeds rarely agree to the single trade

Even tick-by-tick, on the same session, two correct feeds can differ by a handful of contracts. The tape is not as singular as it looks:

• Different vendors, same exchange. Rithmic, CQG and a broker's native feed all carry CME data, but they timestamp, batch and occasionally consolidate prints differently. The differences are small, and they are real.
• Spread and implied trades. A calendar-spread leg or an implied-in trade can be counted by one feed and suppressed by another. On a roll day this alone moves the total.
• Filtered ticks. Some feeds collapse several fills at one price and timestamp into a single print. Fewer prints, same contracts — but a footprint that counts prints instead of size will read differently.

Crypto is the same problem wearing a different hat. There the question is not the session but which venues are consolidated: a Binance-only tape and a multi-exchange tape are both correct and will never match, because they are not watching the same market.

Which number to trust

For order flow, the useful volume is the one your decisions are actually built from: the tick-by-tick, aggressor-tagged count from your execution feed, cut on the same session your exchange uses. That is the volume your cumulative delta and your footprint imbalances are derived from. A tidier figure from a delayed, consolidated, calendar-day feed is not more accurate — it is answering a question you are not trading.

This is a deliberate choice on our side, and worth naming plainly: we reconcile our footprint's daily volume to the number your broker platform shows — the NinjaTrader figure, when you run the bridge — rather than to some abstract "canonical" total. Not because that number is the one true volume, but because it is the one you already read a hundred times a day. A footprint that disagrees with the ladder next to it makes you doubt the tool, even when the tool is right. Agreement, here, is a feature.

If you want the wider operational picture — feeds, entitlements, what a funded account changes — the pillar of this cluster is order flow on a prop-firm account.

Nothing here is financial advice. Feed behaviour, session times and platform settings change — verify against your own broker and data vendor before relying on any of this.