Market Profile: What It Is and How to Read It in Futures

What Market Profile is#

Market Profile is a way of organizing price by time. Instead of looking at one candle after another, it groups the session by price level and counts how long the market traded at each one. The result is a shape that often resembles a bell lying on its side: wide in the middle, narrow at the ends.

The unit of measure is called a TPO, short for Time Price Opportunity. A TPO is a block showing that price touched a level during a given period, usually thirty minutes. The more TPOs a level accumulates, the more time the market spent trading there.

The basic reading is straightforward. Where there are many TPOs, buyers and sellers accepted that price. Where there are few, price moved through because almost nobody wanted to trade in that zone. The profile does not say what will happen: it describes where trading has taken place.

The three levels you read first#

POC, Point of Control: the single price with the most time traded in the session. It is the day's center of gravity and the reference most watched when the market rotates.

Value Area: the band where roughly 70% of the session's time or volume is concentrated. Its upper edge is called VAH, Value Area High, and the lower edge VAL, Value Area Low. Outside that band the market spent little time.

Initial Balance, or IB: the high and low of the first sixty minutes of the regular session. The regular session, or RTH, is the U.S. stock exchange's trading hours, 9:30 a.m. to 4:00 p.m. New York time. The IB sets the day's opening frame.

How a session is classified by its profile#

These levels are used to label each day. A normal day opens, builds a wide initial range and barely extends it afterward. A normal variation extends the range on one side. A trend day pushes in one direction for almost the whole session. A neutral day breaks out of the Initial Balance both above and below.

The table splits ES sessions across those four categories and adds, for each one, how often the next day moved in the same direction.

It is best read with the days column in front of you. The categories do not carry equal weight, and that difference in sample size matters far more than any decimal.

What the split shows#

Measured in the regular session, 9:30 to 16:00 New York time, the most frequent type is the normal variation: 50.7% of ES sessions, across 2,127 days. On half of the days price extended the first-hour range on one side only.

The neutral day, the one that breaks out of the Initial Balance both above and below, is 27.5%, with 1,155 sessions. The trend day is 19.4%, with 814: one in five. The normal day, the one that never leaves the first-hour range, is the rare one: 2.5%, 103 sessions.

Correction of October 11, 2026: the first version of this table classified the nearly twenty-four-hour day, and there the neutral day took up three quarters of the series. That was an effect of the trading hours, not of the market: with that many hours, the range almost always ends up exceeding the first hour on both sides. This version measures the regular session only.

The next day: what the data does not promise#

The second percentage column measures how often the following session closed in the same direction. After a neutral day, 50.6% across 1,155 ES sessions; after a normal variation, 49.6% across 2,127.

After a trend day, 50.4% across 814 sessions, with a 95% confidence interval of 46.9 to 53.8%. A coin. The 40.8% for the normal day rests on 103 sessions: it is the smallest sample in the table and not enough for a headline.

Put plainly: in this series, knowing yesterday's day type does not anticipate today's direction. And today's type is only known at the close: at mid-session, the class the day was showing matched the final one in about half of the sessions. It is a useful result even if it disappoints, because it avoids building a routine on something the data does not support.

Do poor extremes and single prints attract price?#

Two classic profile ideas say the market leaves unfinished business: a poor high or low, one that ends with two or more letters on the last row instead of a tail, and single prints, interior stretches with one letter per row. The usual reading is that price comes back for them.

Measured in ES from 2010-06-07 to 2026-09-30: a poor extreme was exceeded within the next five sessions 79.1% of the time, over 1,234 cases. But a normal extreme was exceeded 73.6% of the time. The difference is there, four to seven points depending on the horizon, and it is small: any recent extreme gets taken almost three times out of four within a week.

With single prints there is no difference: they were filled within five sessions 72.8% of the time, over 776 cases, against 72.2% for a control stretch of the same size. Price comes back, but it comes back just the same to any stretch.

Are profile levels reached more often than any other price?#

The previous section compared a poor extreme with a normal one. The tougher comparison is still missing: a profile level against any price sitting at the same distance. If the level attracts price, it should be reached more often.

The table measures it in ES, regular session, from 2010-06-07 to 2026-10-09. At each session's close it takes the nearest unvisited level above and below, and checks whether the next session trades it. The control is not another level: it is how often the next session moved away from its close, in the same direction, by at least that same distance.

The POC of an earlier session was reached 23.1% of the time; any price at that distance, 23.0%, over 6,940 cases. The edge of an earlier value area, 33.0% against 32.5%, over 7,324. An unfilled stretch of single prints, 18.3% against 17.6%, over 3,898. A prior high or low, 26.1% against 26.0%, over 7,489.

The unfinished extreme, the one that leaves no tail, was reached less often than its control: 11.9% against 12.9%, over 3,792 cases. Once distance is controlled for, the four-to-seven-point advantage from the previous section disappears.

No difference exceeds one percentage point. Profile levels are reached just as often as any price at that distance. They are still useful as references to organize the day and name scenarios; what the data does not support is that they pull price.

Do the levels a session leaves untouched attract price?#

It is said that price goes back for the levels a session left untouched. We measured it in ES, NQ, YM and RTY, from 7-Jun-2010 to 23-Oct-2026 (RTY since 2017), with four levels: the prior close, the high and low of the Initial Balance and the overnight midpoint, each against a price at the same distance from the close, on the other side.

In the next session, ES traded those levels 18.2% of the time and the control 16.6% (4,416 cases): the same as chance. Within 5 sessions, 49.3% against 49.0%. The same in the other three indices.

One difference does hold up, and it is about the calendar: the levels Friday leaves are traded on Monday less than those of other days at the same distance. In ES, 13.0% against 19.6% (943 Mondays); in YM, 13.0% against 20.1% (944); and in RTY, 11.5% against 20.2% (477). In NQ, the same as chance.

It is frequency, not direction: it does not say where Monday goes, only that it reaches those prices less often.

What context does change: the size of the day#

There is one context reading that does leave a mark, and it is not about direction. When ES strings together two or more sessions of rising lows, the next session is smaller; when it strings together falling highs, it is larger.

Measured as the next session's range divided by the 20-session average true range: after a run of rising lows, 0.75 on average over 1,520 sessions, 0.24 less than the rest, with a 95% confidence interval of -0.28 to -0.20. After a run of falling highs, 1.10 over 947 sessions, 0.26 more than the rest, with an interval of +0.20 to +0.31. With no run, 0.93. All sessions, 0.91 over 4,179.

The difference holds in the most recent data: with sessions since 2022 only, it is -0.22 and +0.23.

Direction does not change. The next session closed higher 54.1% of the time after rising lows, 55.4% after falling highs and 54.5% overall. Context says how much room to expect, not which way. It helps with stop distance and position size, not with picking a side.

Does the first half hour set an extreme of the day?#

Another often-repeated line says the day's high or low is set in the first half hour almost every time. In ES, over 4,199 regular sessions, it happened 54.3% of the time, with a 95% confidence interval of 52.8 to 55.8%. A little more than half.

Broken down: the low was set in those 30 minutes in 29.8% of sessions and the high in 24.1%. Both at once, in 17 sessions, 0.4%. With the full first hour, which is the Initial Balance, the figure rises to 72.5%.

Two limits. At 10:00 you do not know which of the two extremes will hold, or whether either will: that is only known at the close. And since 2022 the figure is almost the same, 53.7%, so it is not something that is changing.

How the profile is used in practice#

The profile serves to describe the session's context, not to decide entries. Knowing whether price is trading inside or outside the prior day's Value Area changes what is reasonable to expect: inside, rotation; outside, a search for a new level.

The second use is organizational. The profile forces you to name, before the open, the levels that matter (POC, VAH, VAL and the Initial Balance extremes) and to write down what you would do in each scenario. That reduces improvisation.

The third is calibration. Each profile reading can be checked against a base rate, meaning the historical frequency with which that same scenario occurred under similar conditions. A reading without a base rate is just an organized opinion.

The limits of this measurement#

The day-type classification is a convention, not a physical measurement. Changing the threshold that separates a trend day from a normal variation moves the percentages. Two sources with different definitions will give different figures for the same days, and neither will be wrong.

The series is ES, regular session, runs from 2010-06-07 to 2026-10-09 and covers 4,199 sessions. It spans very different regimes, from years of low volatility to episodes of extreme stress. The average blends them all and represents none in particular.

There are also three things the profile does not measure. It does not measure who is behind each trade. It does not measure the orders that never got executed. And it does not measure the cost of trading: spread, commissions and slippage are left out of every percentage in this table.

Finally, the normal day row, with 103 cases against 2,127 for the normal variation, is the article's best reminder: a percentage without its sample sounds just as solid as one backed by two thousand observations, and it is not.

To close#

Market Profile answers the question of where trading has taken place, not what will happen. Read that way, it organizes the session and provides a shared vocabulary for talking about value, acceptance and rejection.

Everything above is informational and educational material about reading historical data. It is not personalized investment advice. Past results do not predict future results, and trading futures involves risk of loss.

Frequently asked questions

What happens if price opens outside the previous day's Value Area?

It indicates the market is not accepting the prior day's value and is probing for a new one. If price stays outside, the day tends to become more directional; if it returns to the range, the move outside turns out to be a false start. It is context for preparing scenarios, not a trade instruction.

How long does the Initial Balance last?

Sixty minutes from the open of the regular session. In U.S. index futures that means the 9:30 to 10:30 a.m. window, New York time. If you use a different duration, statistics calculated with a one-hour IB are no longer comparable.

Does Market Profile work outside of futures?

The concept applies to any market with a continuous auction. It is more convenient in futures because volume is centralized and the session hours are well defined, which is not the case in markets fragmented across many trading venues.

Does price come back to an unvisited POC?

In ES, no more than to any other price at the same distance. The POC of an earlier session was reached in the next session 23.1% of the time, against 23.0% for the control, over 6,940 cases since 2010. The same goes for value area edges and single prints.

What is the most frequent day type in ES?

In the regular session, the normal variation: 50.7% of 4,199 sessions since 2010. The neutral day is 27.5%, the trend day 19.4% and the normal day 2.5%. If the nearly twenty-four-hour day is classified instead, the neutral day rises to about three quarters, but that is an effect of the trading hours.

These numbers, instrument by instrument

• es
• nq
• rty
• ym
• nkd
• cl
• gc
• 6e

Where this is used in Perfiltrade

• Markets: price against the IB — Where price sits relative to the Initial Balance (above, below or inside), across 25 instruments.
• In-session Market Profile — Initial Balance, Value Area, POC and profile shape during the session, against their history.
• Week: a summary of what is coming — What is coming this week in a paragraph written by rules, plus an optional AI report with its sources.

The calculation and its limits are in the methodology.