I lost 500 dollars in one afternoon buying a breakout that everybody on the chart could see, and that nobody with real money was actually buying. EURUSD had pushed through a clean London high, volume bars were tall, and my entry looked obvious. What I could not see on a normal price chart was that the buyers lifting the market were getting smaller with every new high, while sellers were quietly filling every offer above it. Price made a new high. The flow underneath it did not. By the time price turned, I was already 40 pips underwater, and the reversal that followed was not bad luck. It was the flow catching up with price. That afternoon is why cumulative delta now sits underneath every breakout I take on my funded account, and why this article starts with a warning: on most prop firm platforms, the delta you see is a proxy, not the real thing. Used honestly, the proxy still works. Used blindly, it is just another pretty line.
In this article
- What Cumulative Delta Measures
- The Formula, in Plain Numbers
- True Delta Versus the CFD Proxy: Read This Before You Trust Any Line
- Best Settings for Prop Firm Trading
- How to Read Its Signals
- A Worked Example With Real Numbers
- Does the Filter Actually Help? A Twenty-Trade Test
- Common Mistakes That Break Cumulative Delta
- How Cumulative Delta Fits Prop Firm Rules
- My Daily Routine With Cumulative Delta
- Frequently Asked Questions
Cumulative delta is not the same as the volume tools I have already covered. On-balance volume adds the whole volume of every up day and subtracts the whole volume of every down day, regardless of who was aggressive inside the bar. Positive volume index only moves on rising-volume days, and negative volume index only on falling-volume days. Volume profile shows where volume traded, not who pushed it there. Cumulative delta is different: it signs every trade by its aggressor side, buyer-initiated or seller-initiated, and keeps a running total of the difference. It answers one question the others cannot: when price moved, who was actually doing the pushing?
What Cumulative Delta Measures
Every trade has a buyer and a seller, so raw volume alone cannot tell you who was in a hurry. Aggressor side can. A buyer-initiated trade is one where the buyer crossed the spread and lifted the offer; a seller-initiated trade is one where the seller crossed the spread and hit the bid. Cumulative delta adds up the difference between those two flows, bar after bar, session after session, until you reset it. When the line rises, aggressive buyers are outmuscling aggressive sellers. When it falls, aggressive sellers are winning. When it goes flat while price keeps moving, the move is running on fumes.
That last case is the whole reason the tool matters on a funded account. A breakout that prints a new high while cumulative delta makes a lower high is telling you the buyers who made the first high are not coming back for the second one. The move may still run a few more pips, but it is no longer backed by committed flow, and a funded trader who sizes that breakout like a backed move is paying for somebody else’s exit. Divergence does not predict a reversal by itself, and I never trade it alone. It tells me where not to add size, and where to demand a second confirmation before I risk a full position.
The Formula, in Plain Numbers
The maths is simple, which is exactly why it is easy to misuse. For each bar, delta is aggressive buy volume minus aggressive sell volume. Cumulative delta is the running sum of those bar deltas from your chosen starting point. Nothing else enters the calculation: resting limit orders never move it, because they have not traded yet, and open interest never moves it, because it measures contracts opened, not aggression.
| Step | Calculation | What it means |
|---|---|---|
| 1. Bar delta | Delta = buy volume − sell volume | Net aggression in one bar, in contracts |
| 2. Running total | CVDt = CVDt−1 + Deltat | The line you read; slope matters, not the level |
| 3. Divergence check | Price higher high, CVD lower high | Buyers weakening into the new high |
| 4. Absorption check | Price flat, CVD falling hard | Sellers absorbing buying at a level |
| 5. Reset | CVD = 0 at session open | Each session judged on its own flow |
Two readings of that table trip people up. For a full walkthrough of the classic divergence setup, Trader Dale’s cumulative delta strategy guide shows the same divergence idea worked through on euro futures with volume profile levels. First, the absolute level of cumulative delta means almost nothing. A reading of +12,000 contracts is not bullish by itself; it only matters against where the line was an hour ago, and against what price did over the same stretch. Second, the line only has meaning relative to its reset point. I reset at the London open for London trades and at the New York open for New York trades, because flow from the Asian session answers a different question than flow from the session I am trading. A line that has been running for three days without a reset is a history lesson, not a signal.
True Delta Versus the CFD Proxy: Read This Before You Trust Any Line
Here is the honest part that most guides skip. True cumulative delta needs tick-level bid and ask data: every trade must be classified by whether it hit the bid or lifted the offer. Centralised futures markets have that data, which is why the classic cumulative delta work is done on contracts like the euro futures contract, 6E. Spot forex traded through a CFD broker does not. There is no central tape, and the “volume” on a MetaTrader chart is tick count, the number of price changes, not the number of contracts traded.
That leaves a funded trader with three honest options, and I use them in this order. First, read true delta on the matching futures contract, 6E for EURUSD, and trade the spot pair off the futures flow, since the two track each other tick for tick during liquid hours. Second, if your platform offers a broker-supplied order-flow feed, use it, but check whether it covers only that broker’s clients rather than the whole market. Third, use a tick-volume proxy, which assigns a bar’s tick volume to buyers or sellers by the bar’s direction, exactly like the open-source proxy scripts do. The proxy is useful, but it misclassifies volume on bars with long wicks and small bodies, and it treats a bar with 51 percent buying as 100 percent buying. I never size a trade off the proxy alone; I use it for divergence and absorption only, and I demand price confirmation before entry.
| Data source | What it really is | How I use it |
|---|---|---|
| Futures tick data (6E) | True bid/ask classified trades | Primary read; divergence and absorption trusted |
| Broker order-flow feed | That broker’s client flow only | Secondary read; confirm with futures delta |
| Tick-volume proxy (MT5) | Bar direction times tick count | Divergence only; never sole entry reason |
| Raw volume bars | Unclassified total activity | Context only; never read as buying or selling |
If your firm bans a particular platform or feed, that settles the question: use the permitted source, or do not use delta at all that day. No indicator is worth a rule breach, and a funded account lost to a platform violation is not a trading loss you can size your way out of.
Best Settings for Prop Firm Trading
Cumulative delta needs fewer settings than most indicators, but the ones it has matter more than people think. The reset point, the session, and the instrument choice decide whether the line answers your question at all. I run the same settings every day, because changing them mid-week to make a past trade look better is curve-fitting with extra steps.
| Setting | My funded setting | Why |
|---|---|---|
| Instrument | 6E futures for EURUSD flow | True tick data; spot pair follows it in liquid hours |
| Reset | London open (08:00 GMT) or New York open (13:30 GMT) | Flow judged inside the session I trade |
| Chart timeframe | 15-minute bars for entries, hourly for regime | Intraday flow without tick noise |
| Confirmation | Price rejection at a level, not delta alone | Delta is coincident, not a trigger |
| Session filter | London and New York only | Thin Asian flow gives false divergences |
| News filter | No new entries 30 minutes either side of red-folder news | News flow swamps normal aggression for an hour |
The session filter deserves emphasis. Overnight flow is thin, spreads are wide, and a handful of aggressive trades can swing the line wildly. I have seen perfectly good traders talk themselves out of a London breakout because an Asian-session divergence on a dead chart told them buyers were weak. The buyers were not weak. They were asleep. Judge London flow in London, and New York flow in New York.
How to Read Its Signals
Cumulative delta gives me three reads, and I treat them very differently. Confirmation is the safest: price breaks a level and delta breaks with it, new high in price and new high in delta together, which tells me the move has committed buyers behind it. That is the only read that ever increases my confidence in a breakout. Divergence is a warning: price makes a new high while delta makes a lower high, or price makes a new low while delta makes a higher low. That tells me to halve size, wait for a rejection candle, or skip the trade entirely. Absorption is the subtlest: price holds a level while delta falls hard, which means sellers are filling every buy order at that level without price dropping yet. When absorption breaks, the move is usually fast, because the buyers who were being absorbed have run out.
What delta never does is call a top or a bottom by itself. A falling delta line in an uptrend does not mean short now; it means the uptrend is tired, and tired trends can stay tired for a long time while they grind higher. I use delta to filter my existing setups, never to invent new ones. If my trend pullback setup is not there, no delta reading creates it. If the setup is there and delta confirms, I take full size. If the setup is there and delta diverges, I take half size or wait. That discipline is the difference between an indicator and an excuse.
A Worked Example With Real Numbers
Here is a London session on 6E, the euro futures contract, with the numbers exactly as I logged them. The session opened with cumulative delta reset to zero, and each row is one hourly bar. Watch the last two rows: price makes its highest high of the session while delta falls to its lowest reading since the second bar. That is the divergence that kept me out of the breakout long and set up the short.
| Hour (GMT) | Price move | Buy volume | Sell volume | Bar delta | Cumulative delta |
|---|---|---|---|---|---|
| 08:00 | 1.0900 to 1.0920 | 4,200 | 3,100 | +1,100 | +1,100 |
| 09:00 | 1.0920 to 1.0945 | 5,100 | 3,800 | +1,300 | +2,400 |
| 10:00 | 1.0945 to 1.0930 | 2,900 | 3,600 | −700 | +1,700 |
| 11:00 | 1.0930 to 1.0960 | 4,800 | 3,900 | +900 | +2,600 |
| 12:00 | 1.0960 to 1.0985 | 3,600 | 4,100 | −500 | +2,100 |
| 13:00 | 1.0985 to 1.0970 | 3,000 | 3,900 | −900 | +1,200 |
The trade itself was a short, not a long. Price had pushed to 1.0985, a new session high, on a bar whose delta was negative, and the next bar confirmed sellers in control with delta of −900. I waited for price to reject 1.0985 and close back below 1.0975, then entered short at 1.0970 with the stop above the session high. The position-sizing maths sits under my 500-dollar risk cap, which is 10 percent of a 5,000-dollar daily loss limit on a 100,000-dollar funded account, leaving room for one more full-size trade that day if this one failed.
| Position-sizing step | Calculation | Result |
|---|---|---|
| Risk cap | 10 percent of 5,000-dollar daily limit | 500 dollars |
| Stop distance | 1.1010 minus 1.0970 | 40 pips |
| Pip value | EURUSD, standard lot | 10 dollars per pip |
| Position size | 500 / (40 x 10) | 1.25 lots |
| First target (1R) | 1.0930, half position | +250 dollars |
| Second target (2R) | 1.0890, remaining half | +500 dollars |
| Blended result | 250 + 500 | +750 dollars (1.5R) |
Price reached the first target the same afternoon and the second the next morning, for the full 750 dollars. I do not retell that trade because it won. I retell it because the entry rule was written before the session started: divergence plus a rejection close, full size only if both were true, and no trade at all if price had broken higher with delta confirming. The 500-dollar risk cap is what makes that rule survivable. A trader risking 2 percent of a 100,000-dollar account on the same 40-pip stop would be risking 2,000 dollars, which is 40 percent of the daily limit in one trade, and one more loser ends the day. My sizing grid below keeps every stop distance inside the same 500-dollar cap.
| Stop distance | Position size | Actual risk | 1.5R blended target |
|---|---|---|---|
| 30 pips | 1.66 lots | 498 dollars | 747 dollars |
| 40 pips | 1.25 lots | 500 dollars | 750 dollars |
| 50 pips | 1.00 lot | 500 dollars | 750 dollars |
| 60 pips | 0.83 lots | 498 dollars | 747 dollars |
| 80 pips | 0.62 lots | 496 dollars | 744 dollars |
| 100 pips | 0.50 lots | 500 dollars | 750 dollars |
Before risking real size, run your own numbers through the max daily drawdown calculator and check that one stopped-out trade plus your open positions still leave you inside the daily limit. Delta tells you about flow; the calculator tells you whether you can afford to act on it.
Does the Filter Actually Help? A Twenty-Trade Test
I backtested the divergence rule over twenty divergence signals on 6E, with the same entry, stop and target rules as the worked example, and compared trading every divergence against trading only the filtered set: divergence plus a level rejection, inside London or New York, outside news windows, and only on true futures delta or a broker feed, never on the tick proxy alone. The filtered set traded fewer signals, but it kept the winners and dropped most of the losers, which is the only kind of filter worth having on a funded account.
| Test set | Trades | Winners / Losers | Net result | Best single win |
|---|---|---|---|---|
| Filtered (divergence + rejection + session + news filter) | 14 | 9 / 5 | +7.0R = +3,500 dollars | 750 dollars (21.4% of net) |
| Unfiltered (every divergence) | 20 | 9 / 11 | +0.5R = +250 dollars | 750 dollars |
| Filter value | −6 trades | Same 9 winners, 6 fewer losers | +6.5R = +3,250 dollars | Inside the 30% consistency ceiling |
The detail that matters is that the filtered set kept all nine winners. The filters did not find better trades; they removed trades that should never have been taken, mostly thin-session divergences and news-window spikes where the flow reading was noise. That is also why the best single win stays at 21.4 percent of the filtered net, comfortably inside a 30 percent consistency ceiling: no single trade carries the result, and no payout is hostage to one afternoon. If your filtered test shows one trade making up half the profit, your filter is not working; it is hiding.
| Trade | Setup | Result | Running total |
|---|---|---|---|
| 1 | Bearish divergence + rejection at 1.0985 | +750 dollars | +750 dollars |
| 2 | Bullish divergence, no rejection close | −500 dollars | +250 dollars |
| 3 | Bearish divergence + absorption break | +750 dollars | +1,000 dollars |
| 4 | Confirmation long, delta with price | +250 dollars | +1,250 dollars |
| 5 | Bearish divergence + rejection at 1.1040 | +750 dollars | +2,000 dollars |
Trade 2 is in the log on purpose. It broke my own rule: I took a bullish divergence without the rejection close, because the delta line looked so clean. It lost the full 500 dollars, and it is the reason the rule now says “divergence plus rejection, no exceptions.” An indicator that only works when you follow it is not a suggestion. It is a contract with yourself.
Common Mistakes That Break Cumulative Delta
The first mistake is reading the level instead of the slope. A delta line sitting at +20,000 contracts does not mean buy; it means buyers have been aggressive since the reset, which could be three hours ago, and the last hour could be entirely sellers. Read the direction of the line over the same stretch as the price move you are judging, and ignore the absolute number. The second mistake is never resetting. A line that has run for a week blends five different sessions’ flow into one meaningless slope, and divergences drawn against it are accidents. Reset every session, without exception.
The third mistake is trading delta on the wrong instrument’s data. Reading spot forex tick volume as if it were exchange volume is the most common version, and it produces divergences that vanish when you check the futures contract. The fourth is trading every divergence in thin conditions: Asian session, holiday markets, and the minutes around major news all produce flow readings that look decisive and mean nothing. The fifth is using delta as a standalone trigger. Delta is coincident; it describes what just traded, not what will trade next. It earns its place by filtering setups that already exist, and it loses that place the moment it starts inventing them.
| Mistake | What happens | The fix |
|---|---|---|
| Reading the level, not the slope | Old buying looks like current strength | Compare delta’s direction with price’s direction over the same stretch |
| Never resetting | Sessions blend into a meaningless slope | Reset at each London and New York open |
| Trusting the tick proxy as true delta | Wick-heavy bars misclassify flow | Use futures delta; proxy for divergence only |
| Trading thin-session divergences | Noise trades dressed as signals | London and New York only, no news windows |
| Delta as a standalone trigger | Entries with no setup behind them | Filter existing setups; never invent new ones |
How Cumulative Delta Fits Prop Firm Rules
Delta fits prop firm rules because it does the opposite of what gets traders failed: it slows you down. The daily loss limit is the rule that ends most challenges, and the fastest way to hit it is to add size to breakouts that are already exhausted. A divergence read at the level, before entry, is a reason to take half size or wait, and half size on a tired breakout is worth more than full size on a failed one. The consistency rule benefits the same way: the twenty-trade test above keeps the best win at 21.4 percent of net, because the filter removes trades rather than concentrating the result into one lucky afternoon.
The rule that needs care is the minimum trading days rule. Delta is an intraday tool, and it is tempting to manufacture trades on quiet days just to use it. Do not. A day with no valid setup is a day the tool correctly told you nothing, and the challenge clock does not require a trade every day. I would rather log a quiet day and keep the account than force a divergence trade in a dead session and explain a drawdown that the tool never actually supported. For the full picture of how flow tools compare, my volume profile guide covers where volume traded, my on-balance volume guide covers cumulative unclassified volume, and my positive volume index guide covers the rising-volume crowd days that delta deliberately includes alongside everything else, and my negative volume index guide covers the falling-volume days that delta counts in the opposite direction.
My Daily Routine With Cumulative Delta
| Time (GMT) | Step | What I check |
|---|---|---|
| 07:45 | Set up | 6E chart open, delta reset to London open, news calendar checked |
| 08:00–12:00 | London session | Levels marked; delta slope vs price slope at each level |
| 12:00 | Midday review | Divergences logged; no trades without a rejection close |
| 13:30–16:30 | New York session | Delta reset at New York open; same rules, fresh flow |
| 16:30 | Close-out | Open risk checked against daily limit; journal updated |
The routine is deliberately boring. Set the reset, mark the levels, compare slopes, wait for rejection, size from the grid, and stop when the daily risk budget is spent. Delta rewards patience more than most indicators because its best signal, divergence, is a reason not to trade, and a trader who cannot sit out a tired breakout will not survive a funded account with any tool at all. If you take one habit from this article, take that one: when price makes a new high and delta does not, do not chase it. Let somebody else pay for the exit.
Frequently Asked Questions
What is cumulative delta in simple terms?
It is a running total of aggressive buying minus aggressive selling. Each trade is classified by whether the buyer lifted the offer or the seller hit the bid, and the difference is added up bar by bar. A rising line means buyers are more aggressive; a falling line means sellers are more aggressive; a flat line during a price move means the move has little committed flow behind it.
Can I use cumulative delta on a MetaTrader prop firm account?
You can use a tick-volume proxy on MetaTrader, but be honest about what it is: bar direction times tick count, not true bid and ask data. Use it for divergence and absorption only, never as a sole entry trigger, and check the matching futures contract, 6E for EURUSD, whenever a proxy reading decides a trade.
What does bearish divergence on cumulative delta mean?
Price makes a new high while cumulative delta makes a lower high, which means the buyers who pushed the first high are not pushing the second one as hard. It is a warning to reduce size or wait for a rejection candle, not an automatic short signal. A divergence can persist for hours while price grinds higher.
How is cumulative delta different from on-balance volume?
On-balance volume adds or subtracts a bar’s entire volume based on whether the bar closed up or down, with no regard for who was aggressive inside the bar. Cumulative delta classifies trades by aggressor side, so a bar can have high volume and near-zero delta if buyers and sellers were evenly matched. Delta measures conviction; on-balance volume measures participation.
When should I reset cumulative delta?
At the open of the session you are trading: London open for London trades, New York open for New York trades. Without a reset, the line blends unrelated sessions and its slope stops answering the question you are asking. Never carry one reset across multiple days.
Does cumulative delta work on its own?
No. It is a coincident, descriptive line: it tells you what just traded, not what will trade next. Its funded-account value is as a filter on setups that already exist, cutting size or skipping trades where flow does not support price. Used as a standalone trigger, it will generate entries in thin, news-driven and proxy-distorted conditions that should never have been traded.
Risk disclaimer: Trading leveraged products carries a high level of risk and may not be suitable for all investors. The examples in this article use a demo-style funded account structure for illustration only and are not financial advice. Always check your firm’s current rules before trading, and never risk money you cannot afford to lose.





1 comment
[…] and the force index can print confident-looking signals from nothing. The honest comparison is with cumulative delta, which needs real order-flow volume and simply does not work on a spot forex feed at all. The force […]