Bid, ask & delta

It is 10:30 ET (22:30 ICT, winter). NQ has just printed the morning’s new low at 20,408.00. You look at the 5-minute bar that made it and see delta +60: more contracts traded at the ask than at the bid, yet the bar closed down. Twenty minutes earlier the first low printed on delta −420.

Is the selling over? This lesson covers what that number measures for certain, and the one thing it cannot do: tell you what the next bar will do.

Key terms

Bid and ask (giá chào mua / giá chào bán): the best resting buy price and the best resting sell price Definition. A trade at the ask means a buyer crossed the spread and lifted an offer; a trade at the bid means a seller hit a resting bid. The exchange tags every trade with the side that initiated it (the aggressor flag, M0.07).

Delta (chênh lệch khối lượng chủ động): volume traded at the ask minus volume traded at the bid, counted per price, per bar or per session Definition. Every trade has one buyer and one seller, so delta does not count “more buyers than sellers”. It counts which side initiated. Positive means buyers were the aggressors; negative means sellers were.

Term you need: CVD. Cumulative volume delta is the running sum of bar deltas through the session, drawn as a line. M3.05 covers it in full.

Chart first: effort and result

How it works

Delta is effort; price is the result

Delta measures aggression, not direction. A bar can print +500 and close red. Buyers lifted offers for 500 contracts more than sellers hit bids, but every lift met a resting seller who refilled (absorption, M3.02), or sellers kept lowering their offers so buyers paid less each time. The footprint tells you buyers tried; the close tells you they did not move price up. Never read delta without price.

The 10:30 bar in the figure is the same thing in reverse. Two stories fit it equally well. In one, resting buyers soak up the selling at the lows (the textbook story behind the “absorption” label). In the other, the bids are thin, so very little selling moves price a long way. Delta cannot tell you which story is true.

What delta explains, and what it does not

This is the most useful fact about delta, and it has been measured. Order flow explains price over the same interval. In NYSE stocks, a peer-reviewed study found that order-flow imbalance, a close cousin of delta built from the order book, explains about 65% of the price change in the same interval MeasuredOther market. On NQ, a public practitioner study found that a 1-minute bar’s delta explains 43.6% of that same bar’s price change (R²) and almost none of the next bar’s PlausibleOne trader's study.

Tested as a forecast, delta fails Lore. Net session delta did not predict the next session’s direction, extreme 5-minute delta did not predict the next 15 minutes, and a bar whose delta disagrees with its close (the +500 red bar) did not predict a reversal.

So delta is a description of how the bar was made: by initiators pushing, or by price moving with little aggression behind it. It is not a prediction.

What to watch instead of the number

1. Delta against price. Bars where delta and price agree are made by initiators. Bars where price moves with little or opposite delta are made by something else: thin liquidity, or resting orders moving Definition.
2. Delta trend. In the figure, the first leg down was made on heavy selling and the second on almost none. Price moving further on shrinking delta is “thinning”: later prices were reached with less aggression.
3. Delta extremes and location. The session’s largest prints mark where the most aggression arrived. Where they print (at yesterday’s value edge, in the middle of a chop day) changes what you are looking at, not what happens next.
4. Per-price delta. The same aggression split by price (M3.06): at the bar’s lows or at its highs.

Myth / hypothesis: “Thinning delta and delta extremes mark the turn.” Lore This is the most common way delta is taught. Tested on NQ in the same public study, a higher price high with a lower cumulative-delta high did not reverse more often than other pivots (−1.1 pp [−2.6, +0.4], n = 17,332), and end-of-move volume surges leaned slightly against reversal (−2.8 pp [−5.0, −0.7], n = 14,546). Divergence at a key level was inconclusive. Treat it as a hypothesis you could test, not a rule.

Why it is still worth knowing

Delta is the vocabulary of every footprint, CVD line and delta profile in this module. It tells you whether a move is pushed or drifting, which is context for managing a position you already hold. And it is the cleanest example of explains versus forecasts (EV3.03): a 43.6% same-bar R² sounds like an edge until you check the next bar.

Worked example: the low at 20,408, long side

Synthetic session, built to the figure, not a real date. NQ, 5-minute bars, tick 0.25 points, $5 per tick per contract.

From 10:00 to 10:10 ET NQ falls from 20,440.00 to 20,416.00 on bar deltas −220, −340 and −420: 24 points on −980 of net selling. A bounce to 20,424.00 prints +260. The second leg makes a lower low at 20,408.00 by 10:30 on −150, −90 and +60: 16 points on −180 net.

The description, which is all delta gives you: the new low took less than a fifth of the first leg’s selling, and its bar closed down on positive delta. The tempting conclusion, “sellers are exhausted, buy the low”, has no measured edge.

What you can do with it:

• If you are short from above 20,424, the thinning is a reason to run the management rule you wrote before the trade (M6), not a reason to reverse.
• If you want to be long, the shape is a reason to watch, not to buy. A long needs its own written trigger at a level (M4.09), for example acceptance back above the 20,424.00 bounce high, invalid on a return below 20,408.00. That trigger is a house process House rule, not a tested edge.

The mirror and the failure

Short side. Flip the figure: a higher high made on shrinking buying, the last bar closing up on negative delta. Same reading, same limits.

The failure. Same shape on another synthetic morning: at 10:35 ET the next bar prints delta −640 and NQ trades 20,396.00, twelve points under the “exhausted” low. A trader who bought 20,408.25 on the divergence is 49 ticks under water, $245 per contract; a trader who waited for a written trigger never entered. In the public study this shape did not reverse more often than an ordinary pivot: that is what no edge looks like.

Common mistake: Treating delta as a buy/sell signal: “delta positive, so it must go up.” Delta describes who pushed this bar; price decides who won it. Read delta with price and at a location, as a description, or not at all.

Drill: measure your baseline

Delta’s scale depends on the instrument and the time of day: +200 on a 5-minute bar can be large overnight and a rounding error just after the RTH open. For two weeks, log your instrument’s bar delta on one timeframe, split into overnight and RTH. Write down the 50th, 90th and 99th percentile of the absolute values.

Recap

• Delta = ask volume − bid volume: a measure of who initiated, not of where price goes. A positive-delta bar can close down.
• Delta explains its own bar (NQ 1-minute R² 43.6%, one trader’s public study; about 65% for order-flow imbalance in stocks) and almost none of the next.
• Net session delta, extreme delta, delta that disagrees with the close and divergence were tested on NQ as forecasts and showed no edge.
• Use delta to describe how a move was made and to manage a position by a written rule; “extreme” needs your own measured baseline.

Further reading

• Order Flow, Measured: All claims. A public, pre-registered practitioner study on NQ, 2018–2026. Read the delta rows (MR-02, DA-01, DA-03, DA-05, DA-07, DA-09).
• Cont, Kukanov and Stoikov (2014), “The Price Impact of Order Book Events”, Journal of Financial Econometrics 12(1). Order-flow imbalance explains price changes over the same interval; that is explanation, not forecast.
• Our lab: same-interval delta is descriptive; its forecasting use, CVD divergence included, is graded Lore.

Next: M3.05 adds delta up through the session (CVD) and applies the same test to divergence; M3.06 splits delta by price.

Educational content only. Not financial advice. Futures trading involves substantial risk of loss.