What Is Cumulative Delta in Futures Trading?

Cumulative delta is a running total. Take the difference between the volume that traded against resting offers and the volume that traded against resting bids, then carry that figure forward from one bar to the next instead of clearing it each time, and you have the cumulative version.

The count has to begin somewhere, and where it begins governs every value after it. That detail is the reason two traders can watch the same contract on the same day and read different numbers off their screens.

What is cumulative delta in futures trading?

Cumulative delta is the accumulated difference between buy-side and sell-side executed volume, summed across every bar since the count started. In Sierra Chart's volume version the quantity accumulated is ask volume minus bid volume, the span it accumulates over is either the whole of the chart's data or a single trading day, and the result is drawn as its own high-low candlesticks. The study runs on intraday charts only.

The sides come from who was in a hurry. Every fill pairs one buyer with one seller, as ATAS points out in its own write-up of the indicator, so the split is never a headcount. What gets counted is the aggressor: a market sell order that takes out a resting bid is booked to the sell side, and a market buy that lifts a resting offer is booked to the buy side. Those resting orders are the same ones stacked up in a depth of market ladder.

Sierra Chart's calculation section is specific about the boundaries. Ask volume is volume that printed at the best ask or above it; bid volume is volume that printed at the best bid or below it. A trade that occurs between the two quotes gets assigned by tick direction. The same raw split, shown at each individual price inside a bar rather than totalled, is what a footprint chart draws.

How is cumulative delta drawn as candles?

Cumulative delta is drawn as high-low candlesticks whose open, high, low and close are volume differences rather than prices — that is the display Sierra Chart documents for all three of its versions. Sierra Chart tracks the running difference as each bar forms and keeps its highest and lowest points, which it calls DifferenceHigh and DifferenceLow, then builds the candle out of those two extremes and the previous candle's close.

Laid out as arithmetic, the construction is:

• Open — the previous candle's close, or zero where there is no previous candle or the study has just reset. The open is pulled inside the candle's own high-low range when it falls outside it, so a value of zero is not guaranteed.
• High — the previous close plus DifferenceHigh.
• Low — the previous close plus DifferenceLow.
• Close — the previous close plus the bar's ask volume minus its bid volume.

Because the close is the running total, a run of up candles means the imbalance kept building in one direction. The colouring follows the usual candle rule of close against open, with the caveat Sierra Chart flags in its own text: the open and the close being compared are ask-versus-bid volume differences. A cumulative delta candle's colour therefore has nothing to do with the direction of the price bar beside it.

ATAS offers the same series in three shapes: bars, candles or a plain line.

Is there more than one kind of cumulative delta?

Yes. Sierra Chart's studies list carries three separate cumulative delta studies, and its troubleshooting section says plainly that comparing two charts starts with checking they are running the same one. Each sums a different quantity, so the three will not agree with each other on the same market.

The trades version counts executions rather than volume, so a one-lot trade and a two-hundred-lot trade each move its difference by one. Sierra Chart documents two companion studies, Number of Trades-Ask and Number of Trades-Bid, that show each half of that difference on its own.

The uptick/downtick version classifies on movement between consecutive prints: uptick volume traded at a higher price than the print before it, downtick volume at a lower one. Where the last price is unchanged, the platform's rule is that the print keeps whichever direction came before it.

Does cumulative delta reset at the start of each session?

Only if you tell it to. Sierra Chart exposes a Reset at Start of Trading Day input, and when it is switched on the running sum restarts from the session start defined in the chart's own session times — either the start time, or the evening start time where an evening session is enabled. A second input, Reset at Both Session Start Times, implies the first one and adds a further reset at the evening start time when the evening session is enabled.

There is a wrinkle worth knowing in the reset itself. Sierra Chart notes that the value the series opens on after a reset is whatever the first trade of the session, or the first intraday record, contributed on its own — so it does not start at zero, which the documentation says is normal and expected.

ATAS documents the same choice as a checkbox it calls Session Delta Mode: ticked, the indicator restarts each trading session; unticked, it runs over everything loaded in the chart. Another platform's published documentation offers the same fork under a setting it calls Indicator Range, chart range or session based, and adds Reference Points, user-defined times at which its count restarts from zero.

The setting decides what the number even means. Reset on, and the value answers "how far has the imbalance run today?". Reset off, and it answers "how far has it run since the oldest bar my chart happens to be holding?" — a question about your chart settings as much as about the market.

Why do two charts of the same market show different cumulative delta?

Because the total inherits everything that came before it, and two charts rarely start from the same place. Sierra Chart devotes a whole documentation section to this, and its headline point is that values can shift whenever a chart reloads — triggered by opening a chartbook, by a reload you ask for yourself, or by a fresh download of the history, whether partial or complete.

Its checklist for making two charts agree doubles as a list of everything cumulative delta silently depends on. Each has to match:

• which of the three study versions is loaded;
• the session times, together with whether a fresh bar opens at each session start and whether weekend data is loaded;
• the days-to-load or date-range start, where the study is running without a daily reset;
• the market data feed, which also has to be one that supplies consistent data;
• tick-by-tick storage in the intraday chart data files being switched on;
• the time zone under the data/trade service settings, where two separate installations are being compared.

Sierra Chart recommends leaving new-bar-at-session-start on, because with it off a reload can move the chart's first timestamp and drag every subsequent value with it. The drift also shows up visually: without a daily reset, saved chart drawings appear to wander as the bars slide underneath them, and the documentation offers no automatic realignment for it. The cures it suggests are to switch the daily reset on, or to pin the chart's start by extending days-to-load each day or fixing a date range.

On reconciling the study against a rival program, the page declines outright, citing the number of things that would have to line up for the two to agree exactly.

What data does cumulative delta need to be accurate?

Trade-by-trade data, and a feed that supplies the bid-ask split historically. Sierra Chart words the storage requirement differently for each study. The trades and up/down-tick pages state that tick-by-tick data has to be in the intraday chart data file or the result is not accurate. The volume page instead asks for an intraday data storage time unit of two seconds or less, and says tick-by-tick gives the highest accuracy — though its troubleshooting section then calls tick-by-tick data for everything loaded critical whenever the daily reset is switched off.

The historical half is the one that catches people out. The volume study's page states that it needs historical bid volume and ask volume from whichever data or trading service is connected, and that where the service does not provide it the study reads zero or simply stops moving for the periods the platform was disconnected. Its documented workaround is to stay connected so the platform records the split in real time as it happens, and its supported-services page is where each service's capability is listed.

ATAS says its own indicator is calculated from exchange tick data over a direct connection, and its 2019 article closes on the observation that not every market or exchange transmits correct delta data.

For anyone trading here, the included feed is Level 1 across CME, CBOT, COMEX and NYMEX — the best bid and the best ask at the top of the book — and the supported platforms page lists the four platforms you can use it from and which of them need a licence you buy yourself. Whether a given service also hands over the historical bid and ask volume a cumulative delta study wants is a question for that platform's own service documentation.

What is a cumulative delta divergence?

A divergence is the case where price and delta point opposite ways — price making a new high while the delta reading stays negative, or the reverse. Quantower ships an indicator that formalises it on bar delta, and its documentation states the two conditions precisely: a bar closing with both a lower high and a lower low while delta is positive prints a buy signal, and a bar closing with both a higher high and a higher low while delta is negative prints a sell signal.

Quantower's cumulative indicator is a separate one, COT High/Low, whose accumulation begins when price hits a new high, or revisits a previous one, and which sums bid-ask delta from that point. The page reads a rejected high as a negative COT High while price falls, and a price that stays high while COT High is negative and growing in magnitude as a sign that buy limit orders are holding it up. Quantower also warns readers not to confuse the indicator's name with the CFTC's weekly report, which is a different thing entirely — we cover that one in how to read the Commitments of Traders report.

What none of this documentation contains is a test. None of the platform pages cited here reports whether a delta divergence is followed by a reversal more often than chance, and ATAS's own article says the indicator produces no simple signals and calls it "not a magic wand". Treat a divergence as a description of what the tape did, and keep the evidence question open.

What does cumulative delta mean on a funded account?

Nothing, in rule terms. Cumulative delta is an analysis tool that lives on your chart; no rule here is measured from it, and the constraint that can actually end an account on performance is the drawdown line, which our help centre calls the only rule that closes an account on performance. The tool is neutral on whether you should be in a trade at all.

Where it can quietly hurt is consistency of interpretation. A cumulative delta of −40,000 means one thing when the count restarted this morning and something very different when it has been running since whatever date your chart was last reloaded with. If you are building a routine around a reading, fix the reset setting once and leave it alone, so today's number is comparable with the one you wrote down last week.

And the reading says nothing about the size you should carry. That is arithmetic on your account's own numbers, worked through in futures position sizing: the distance to your stop and the value of a tick set the contract count, whatever the order flow looks like.

FAQ

Is cumulative volume delta (CVD) the same as cumulative delta?

The name changes from platform to platform, while the documentation describes the same calculation. One vendor's knowledge base calls its own indicator Cumulative Volume Delta, or CVD, and defines it by the aggressor side of each trade, sellers set against buyers. Sierra Chart's equivalent is Cumulative Delta Bars - Volume, the running sum of ask volume less bid volume, and ATAS simply calls its own Cumulative Delta. Sierra Chart's other two versions sum trades and up/down-tick volume instead, and those are separate calculations that will not match a volume-based one.

Why does my cumulative delta change when I reload the chart?

Because the total is cumulative from the first bar the chart holds, and a reload can change which bar that is. Sierra Chart's documentation states that the output will likely change after a data reload for exactly this reason, and lists chartbook opening, manual reloads and re-downloaded historical data as the events that trigger it. Its recommended fix is to switch the daily reset input on, which pins the start of the count to the session rather than to the chart window.

Does cumulative delta show how many buyers and sellers there are?

No. Every executed trade has a buyer and a seller, a point ATAS makes directly in its own explanation of the indicator. What delta separates is which side crossed the spread: the market order is credited with the volume, and the resting limit order it traded against is not. A large positive cumulative delta means aggressive orders were lifting offers, and says nothing about how many participants stood on each side.

Can I compare cumulative delta between two trading platforms?

Sierra Chart's own troubleshooting section declines to help with this, citing the number of things that would have to line up before two charting programs agreed exactly. Its checklist covers the study version, the session times, how much history is loaded, the data feed, tick-by-tick storage and the time zone, and all of that has to match before two charts inside one platform agree. Across two platforms, treat the figures as separate series.