Day types in practice

It is 11:00 ET (22:00 ICT) on Tuesday 1 September 2026. NQ broke its initial-balance high of 29,179.50 at 10:30, and period C closed at 29,228.00. You bought 1 MNQ at 29,200.00 on the break. The IB low, 29,040.25, has not been touched. Is this a trend day you should hold, or a normal-variation day where you should take what you have?

M1.04 showed that nobody can label the day until 16:00. This lesson is about the five hours in between: you choose management from a written hypothesis, and you decide in advance what will make you drop it.

Key terms

Day-type hypothesis (giả thuyết loại phiên): a one-line written guess, made during the session, of the label the day is heading for, the management it implies, and the observation that would end it House rule. It is a planning tool, not a forecast with a measured hit rate.

Kill condition (điều kiện hủy): an observation written down before it happens that ends the hypothesis, for example “a 30-minute period closes back inside the IB” House rule.

Terms you need: the day types and acceptance. The five labels and the rule that assigns them are in M1.04: IB never left = normal; both IB sides broken = neutral; one side broken = trend if the range reaches at least twice the IB and the close is in the outer quarter on that side, otherwise normal variation. The time, volume and close checks for acceptance are in M4.01.

Chart first: the hypothesis meets the afternoon

How it works

Each label implies a way to manage, not an entry

The same long entry can be held on one day, scalped on another and skipped on a third. The entry does not change; the management does. The menu below is a set of process choices House rule. None of them has been tested by our lab as a profit rule on NQ.

• Trend day (ngày xu hướng): trail the stop instead of setting a fixed target, and add only on pullbacks. A finished trend day closed near its extreme by definition Definition, so on that day holding beat fading. That is also the trap: you find out it was a trend day at 16:00.
• Normal variation day (ngày biến thể bình thường): manage the morning extension, then stop starting new trades in its direction once it stalls. The afternoon is for managing what you hold. If you fade the afternoon extreme, do it only at a mapped level with a written confirmation rule (M4.09), and remember that mapped levels themselves did no better than random prices in our lab (M4.07).
• Neutral day (ngày trung tính): both IB sides have broken, so the first direction has already failed once. Think two-sided, keep size small and take profits sooner.
• Normal or non-trend day (ngày đi ngang): a narrow day with little trade. The hardest discipline is doing less: trade small or stand aside.

Myth / hypothesis: “On a neutral day the middle is a magnet and breakouts usually fail; on a balance day every breakout fails.” Lore Neither has been tested by our lab on NQ. The neutral label itself needs a second IB break, so “the first breakout failed” is part of how the day is named, not a forecast you can trade at 11:00.

The catch: you choose before the label exists

The menu needs a label, and the label arrives at 16:00 Definition. In our lab’s 58 NQ sessions from 17 August to 6 November 2026, the label so far at 11:00 ET matched the closing label on 24, and at 13:30 on 41 Context (M1.04 has the full count). A public practitioner study found no skill in forecasting the day type by 11:00 Lore (numbers below).

So expect your 11:00 hypothesis to be wrong more often than right, and build for that:

1. Choose management that is cheap to reverse. A trailing stop or a smaller size works on several day types; “hold to the close” works on one.
2. Write the kill condition before you need it. Under pressure, everyone finds a reason to keep the morning’s story.

The hypothesis habit

By 11:00 ET (22:00 ICT in summer, 23:00 ICT in winter), write one line with four parts: the label you are leaning to, the management it implies, the kill condition, and the time. For example: “11:00 ET: normal variation up. Hold the long, stop 29,150.00, no new longs above 29,268.00. Kill: a 30-minute close back below 29,179.50; then flat, no new longs.”

A hypothesis can be falsified; a feeling cannot. When the kill condition fires, cross the line out, write the revision and timestamp it. The morning sheet (M4.08) is where the hypothesis lives; the revision is where the discipline lives.

A revised hypothesis needs its own kill condition. Price wobbling back your way for one period does not bring the old one back.

Why it is still worth knowing. Day types give you words for what the auction is doing so far, and a written hypothesis turns “I have a feeling about today” into something you can check, journal and later count. That count is how you find out whether your own day-type reads add anything.

What the data says

Worked example: 1 September 2026 (real data)

An illustration on one real NQ session, not a tested rule. Prices are from our lab’s tick recording; results are before costs and slippage.

Entry. NQ trades above the IB high (29,179.50) at 10:30 ET. You buy 1 MNQ ($0.50 a tick, $2 a point) at 29,200.00 at 10:31. Stop 29,150.00: 50 points, $100 of risk.

11:00 ET (22:00 ICT). Period C closed at 29,228.00; the IB low is untouched. Your line: “Normal variation up. Hold, stop 29,150.00, no new longs above 29,268.00. Kill: a 30-minute close back below 29,179.50.”

11:49 ET. NQ reaches 29,317.75, the high of the day.

12:30 ET (23:30 ICT). Price first traded back below 29,179.50 at 12:21. Period F closes at 29,177.00, inside the IB. The kill condition has fired. Revision, timestamped: “12:30: back inside the IB. Not trending up. Flat, no new trades in the morning direction; neutral if 29,040.25 breaks.” You sell at 29,177.00.

Three ways to manage the same entry:

The rest of the day: G closed at 29,191.50, back above the IB high, the kind of wobble that tempts you to un-revise. H closed inside again at 29,142.25. At 14:42 ET (01:42 ICT) period K broke the IB low and made the day’s low of 29,002.00. NQ closed at 29,126.75, near the middle: a neutral day, 2.27 times the IB. Plan A without a stop, held to the close, would have lost 73.25 points ($146.50).

Both written plans beat the unwritten one on this day, and the trailing stop beat the hypothesis exit. One day proves nothing about any of them; M7.10 shows how to test a rule like this on your own sessions.

The mirror and the failure

Short side. Everything flips for an IB-low break: the hypothesis is “normal variation down”, the stop sits above the IB, and the kill condition is a 30-minute close back above the IB low.

Where the kill condition fails. On Thursday 24 September 2026 NQ poked 1.25 points above its IB high of 30,638.50 in period C and closed back inside. A trader leaning “balance” at 11:00 would plan to fade the IB edges. Then F broke out and closed at 30,697.75 at 12:30, and the day closed at 30,756.75 as a trend day up, 2.30 times the IB. A balance hypothesis needs its own kill condition, here a 30-minute close above the IB high. Even then, it is late by design: you revise after the evidence, never before.

How often did this kind of rule fire? In the same 58 NQ sessions, 26 had broken exactly one IB side by 11:00 ET. The “30-minute close back inside the IB” condition fired on 19 of them: 14 ended as normal variation, 4 as neutral and 1 (24 September) as a trend day in the original direction. Of the 7 where it never fired, 4 ended as trend days and 3 as normal variation Context. That is one autumn, counted after the fact, with no costs and no profit test: context, not evidence that the rule makes money.

Prop-firm note. Trailing a stop and exiting at market are order-handling decisions. On an evaluation account, check how your firm measures loss limits and what it says about simulated fills before you rely on either; the Prop Firm course covers both.

Common mistakes

Common mistake: Marrying the morning read: deciding “trend day” at 09:45 and defending it all afternoon while price rotates back through the IB. Defending a thesis is ego, not trading. Price can trade back inside the IB whether you allow it or not. The fix is mechanical: write it, watch the kill condition, revise in writing. Feelings do not get revised; positions in them get averaged down.

Common mistake: Treating a frequency as an edge. Normal variation was the most common label in our 58 NQ sessions (35), but that does not mean fading every IB break wins. That claim is untested here, and fading levels showed no edge in our lab’s tests.

Common mistake: Un-revising on a wobble. On 1 September period G closed back above the IB high one period after the kill. Re-entering the morning story there, without a new written condition, turns a small loss into the full stop.

Drill: write five hypotheses and keep score

Use five recent NQ or ES sessions in replay, or five live sessions on a simulator.

1. At 11:00 ET, write the four-part line: label, management, kill condition, time.
2. Note the exact time the kill condition fires, if it does, and write the revision.
3. After 16:00, apply the M1.04 rule and write the final label.
4. Score each day: did the hypothesis survive, and would its management have cost or saved points against a plain trailing stop from the same entry?

What we don’t know yet

• Whether choosing management by day type earns more than one fixed management rule (for example, always trailing) on NQ: not yet measured.
• Whether a written kill condition improves results on NQ after costs: not yet measured. The 26-session count above is description only.
• How often each day type occurs on NQ over many years: not yet measured; our count covers one autumn.

Recap

• A day type chooses your management, not your forecast: trail on a day that is trending so far, stop initiating when the extension stalls, trade small or stand aside on a narrow day. These are house rules, untested as profit rules.
• The label needs the close. On our 58 NQ sessions the 11:00 label matched the final label on 24; a public study found no skill in forecasting it by 11:00 (Lore).
• By 11:00 ET, write label, management, kill condition and time. When the condition fires, revise in writing and give the new hypothesis its own condition.
• On 1 September 2026 the kill condition cut the loss on a long that a “trend day” hold took to the full stop; one day proves nothing.

Further reading

• James F. Dalton, Eric T. Jones and Robert B. Dalton, Mind Over Markets (2nd ed., Wiley, 2013), chapters on day types. Read the definitions, then ask how each management idea would be tested.
• Order Flow, Measured: All claims. A public practitioner study on NQ; rows AP-08 (day-type forecast by 11:00) and RL-02 (trend days and VWAP, a description, not a forecast).
• Our lab: day-type labels for 58 NQ sessions, autumn 2026, how often the 11:00 label survived to the close, and the TPO study in which a “day type so far” filter did not help.

This lesson connects to: M1.04 (the day-type rule) · M4.02 (the open read that seeds the hypothesis) · M4.08 (the morning sheet, where the hypothesis is written) · M5.11 (trend-day continuation) · M7.10 (testing your own rule).

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