Delta is the volume traded at the ask minus the volume traded at the bid, so it measures which side was more aggressive. Cumulative delta adds that up across a session, giving a running total of net aggression. Divergence is when price makes a new extreme and cumulative delta does not follow.
Delta is the tool most new order flow traders fall in love with and then get hurt by, because it produces a signal roughly every four minutes and most of them lead nowhere.
What delta actually counts
Every trade has a buyer and a seller. Delta is not counting how many of each exist, it is counting which side crossed the spread.
Somebody who buys at the ask wanted in badly enough to pay the offered price rather than wait. That is aggression, and it adds to delta. Somebody who sells at the bid did the same in the other direction, and subtracts from it.
So positive delta means aggressive buyers dominated. Negative means aggressive sellers did. Nothing more complicated than that.
What it deliberately ignores is the passive side, which is a real limitation. When a thousand contracts trade at the ask, a thousand were also sold by someone sitting there patiently offering them. Delta credits the impatient party and says nothing about the patient one, even though the patient one is frequently the more informed.
Delta, cumulative delta and delta by price
Three flavours, easy to confuse:
- Delta for a single candle. Net aggression during that bar.
- Cumulative delta across the session. A running total, usually drawn as its own line or candle series under price.
- Delta by price inside a footprint. Net aggression at each individual level within a bar.
Cumulative delta is the one people mean when they talk about divergence. Delta by price is where absorption shows up, covered in absorption, exhaustion and stacked imbalances.
Divergence, and what it really means
The classic setup: price grinds to a new session high, but cumulative delta makes a lower high than it did at the previous peak.
Read literally, that says price went up while net aggressive buying got weaker. Which raises the obvious question. If buyers were less aggressive, who moved price up?
Usually the answer is that sellers stepped back rather than buyers stepping in. Price drifted higher through thin resting supply on modest participation. That kind of high is fragile, because it was not built on anyone actually wanting the price.
Occasionally the answer is something else entirely: a large passive buyer was absorbing on the way up, which produces a similar looking divergence and means close to the opposite thing.
This is why I treat divergence as a question rather than a signal. It tells me the move and the aggression behind it disagree. Finding out which explanation is correct takes looking at the footprint.
ES price with cumulative delta plotted below, at a session high where price made a higher high and cumulative delta made a lower high. Mark both highs on each pane so the divergence is unmistakable, and show what price did afterwards.
The traps
Divergence is common and mostly meaningless
On any given session you will find a dozen divergences. Most resolve into nothing. If you trade every one you will be short in uptrends all day.
The filter is location, as with everything else. A divergence at a random price in the middle of a range is noise. A divergence at the value area high, on the third test, into a level that already rejected twice, is worth acting on.
Cumulative delta resets, and the reset point matters
Where your platform starts counting changes what the line looks like. Session start, daily open and rolling windows all produce different pictures of the same market.
Pick one and stay with it. Traders who switch reset points mid analysis end up comparing readings that are not comparable.
It gets distorted in thin markets
Overnight and around lunch, small orders move delta a long way because there is so little volume overall. A dramatic looking delta swing at three in the morning is often one participant.
The passive side is invisible
Worth repeating because it causes real misreads. Strongly positive delta with price going nowhere does not mean buyers are winning. It means buyers are trying and something is stopping them, which is bearish, not bullish. The number alone points the wrong way.
How I actually use it
Three jobs, none of them entry signals.
Confirming the state. Cumulative delta trending persistently one way supports a read of imbalance. Delta oscillating around a level supports balance. This feeds directly into the state call described in balance vs imbalance.
Checking trend health. An uptrend with cumulative delta making higher lows alongside price is being driven by real buying. An uptrend where delta has gone flat while price keeps rising is being driven by absent sellers, which is a much weaker foundation.
Flagging tests worth watching. When price approaches a level I already care about and delta is diverging into it, I pay closer attention to what the footprint does on arrival. The divergence gets me looking. The footprint decides.
What it replaces
If you came from momentum indicators, cumulative delta does the job RSI was attempting, with actual transaction data instead of a formula on closing prices.
RSI at 78 tells you price rose quickly recently. Cumulative delta flattening while price rises tells you the buying behind the move is thinning out. The second is a statement about participants. The first is arithmetic on the price series. That difference is covered further in indicators vs order flow.
It is still not a signal. Nothing on a footprint chart is. It is a better quality input into a decision you make.
Frequently asked questions
What is delta in trading?
Delta is the volume traded at the ask minus the volume traded at the bid. It measures which side was more aggressive, meaning which side crossed the spread rather than waiting. Positive delta means aggressive buyers dominated, negative means aggressive sellers did.
What does delta divergence mean?
It means price made a new extreme but cumulative delta did not confirm it. Price rose while net aggressive buying weakened. Usually that indicates the move happened because sellers withdrew rather than because buyers arrived, which makes the level fragile. It can also indicate a large passive participant absorbing, which means close to the opposite.
Is cumulative delta a leading indicator?
It is coincident rather than leading. It reports aggression as it happens rather than predicting it. That still puts it ahead of anything calculated from closing prices, but it does not tell you what is about to occur. Treat it as higher quality current information, not a forecast.
Why does delta sometimes contradict price?
Because delta only counts the aggressive side. If buyers are aggressively lifting the offer and price is not rising, a large passive seller is filling every one of them. The contradiction is the useful part: it reveals a participant who does not show up in the delta number at all.
What should cumulative delta reset on?
Most traders use the session open, which keeps the reading tied to the current day's participation. Daily and rolling resets are also valid. What matters is consistency, because the same market produces very different looking cumulative delta depending on where the count begins.
Can I trade delta divergence on its own?
I would not. Divergences appear many times a session and most lead nowhere. Used alone they will put you against trends repeatedly. Used as a reason to examine a level you already considered important, they are genuinely valuable.