Opening reversal

Tuesday 20 October 2026, 09:45 ET (20:45 ICT). NQ opened at 29,594.00, below the whole of yesterday’s range, and has fallen 202 points in fifteen minutes, through the overnight low, on heavy selling. Then one 5-minute bar prints a low of 29,391.75 and closes 133 points higher. Is this the turn, or a pause before the next leg? This lesson writes the opening reversal as a rule you can check at the bar close, prices it before the click, and shows what the data says about it. Short answer for the last part: on NQ, no edge has been shown.

Key term

Opening reversal (đảo chiều sau mở cửa): an open drive that fails. Price drives one way from the 09:30 ET open, is rejected at the drive’s extreme, and closes a 5-minute bar back past the midpoint of the range traded so far Definition. The failed drive is the event; the trade is the position you may take against the drive once all three checks have passed.

Open drive (cú đẩy giá sau mở cửa): a hard, mostly one-sided move from the open in the first 20 to 30 minutes (M1.05 labels the four opening types) Definition. Midpoint of the range so far (điểm giữa biên độ đến lúc đó): halfway between the drive’s extreme and the opposite extreme traded since 09:30 Definition. We use it instead of the full 30-minute opening range because the trigger often comes before 10:00 ET, when that range is not finished.

How it works

The anatomy, written before the open

Every setup in this module has the same parts: context → trigger → invalidation → target → R:R, written before 09:30 ET so the decision at the bar close is a check, not a feeling. The numbers are conventions that make the checks scorable, not measured cut-offs House rule.

• Context: a drive that qualifies. At least about 0.35% of price from the open (about 100 NQ points at 29,000) within the first 20 minutes, mostly one way. Measure distance, time and delta.
• Trigger: rejection plus a close past the midpoint. The extreme is rejected (the bar that made it, or a later one, closes well away from it), and a 5-minute bar closes past the midpoint of the range so far. A rejection without the close is a pause; a close without a rejection is rotation.
• Entry: at the trigger bar’s close. Never during the drive.
• Invalidation: 2 points (8 ticks) beyond the drive’s extreme. If price trades there, the drive has resumed: exit, no debate.
• First target: the opposite end of the range so far, measured at the trigger.
• R:R and break-even: computed with costs before the click.

The arithmetic of this setup

Look at where the parts sit. You enter at or past the midpoint, your stop is beyond one end of the range and your target is the other end. So the reward is at most half the range, and the risk is at least half the range plus the buffer Definition. At the first target this setup always pays less than 1R, and its break-even win rate is always above 50% before costs. On 20 October the numbers were 135.25 points of risk for 80.50 points of reward: 0.6R.

That is not a reason to skip the setup; it is what the setup needs. A rule that must win more than 60% of the time needs evidence that it does, and that evidence does not exist yet (next section). Holding past the first target is a separate rule with its own exit, also not yet measured.

The patience clause

A drive can run further than feels reasonable. Buying a falling open because “it has gone too far” is fading initiative with no verdict and no stop that means anything. The setup starts at the trigger, not at your boredom House rule.

Myth / hypothesis: “The first reversal of the day carries extra fuel: the open’s traders are the most committed, a failed drive traps more size than any midday pattern, and their exits power the move; it rotates the whole opening range.” Lore Trapped size and stop locations cannot be observed, and the closest tests found nothing: the open type did not forecast direction on NQ, and failures at the overnight extremes did not reverse more often than other moves (boxes below). Treat the story as description, not as a reason the trade works.

Myth / hypothesis: “The open corrects overnight inventory, and the reversal starts once the correction is done.” Lore On NQ the measured direction was the opposite of the claim (box below).

Why it is still worth knowing

• Shared vocabulary. “Drove 202 points, rejected at the low, closed back above the midpoint” is a description anyone can check on the chart.
• Framing risk. The drive’s extreme is an objective invalidation, and the patience clause keeps you out of a move that is still extending.
• A testable rule. Written checks and fixed exits can be scored on replay and in your journal (see the drill): the only route from “not yet measured” to “measured”.

Where the open sits: use the base rates for the target

Did NQ open inside yesterday’s RTH range or outside it? M1.05 owns these numbers, counted by our lab on the Nasdaq-100 ETF (QQQ) from 2015 to 2025 MeasuredNasdaq index:

• Open outside the prior range: the far side was not traded on about 89% of sessions. A reversal back into the prior range is ordinary; a run to the far side is the rare case. On 20 October the open was below the prior low (29,653.75); the reversal reclaimed it, and the session high (29,903.50) stayed under the prior high (29,983.50).
• Open inside the prior range: exactly one side broke on about 74% of sessions, both sides on only about 9%. If the drive has already broken one side and failed, a target beyond the other side plans for the uncommon day.

Most of each rate is ordinary distance: a distance-only placebo gives 86.5% and 66.2% (box below). Use the rates to keep targets realistic, not as a signal House rule.

What the data says

So the honest status of this setup: a written rule with a defined stop, a first target that pays less than 1R, and no measured edge. You can trade it small while you score it. You cannot call it high odds.

Worked example: 20 October 2026, the drive that failed

Real NQ tape from our lab’s recording, traded as MNQ ($0.50 a tick, $2 a point). It shows the reading, not a signal.

• Before the open. Prior RTH range 29,653.75–29,983.50, prior value area low 29,718.25, overnight low 29,478.50. NQ is set to open below the prior range.
• 09:30–09:45 ET (20:30–20:45 ICT), the drive. Open 29,594.00. Price breaks the overnight low at 09:40 and reaches 29,391.75 at 09:45: 202.25 points in 15 minutes, delta −1,964 over three bars. It qualifies. You do nothing yet.
• 09:50 ET, the trigger. The 09:45 bar makes the low and closes at 29,525.00, with delta +1,149. The highest price since 09:30 was 29,605.50, so the midpoint is 29,498.75. Rejection ✓ (closed 133 points off the low), close past the midpoint ✓.
• Plan. Long 1 MNQ at 29,525.00. Stop 29,389.75: 135.25 points = $270.50. First target 29,605.50: 80.50 points = $161.00.
• Break-even. With about $2.50 of round-turn costs, a loss costs $273.00 and a win nets $158.50, so the trade must win 273 ÷ 431.50 ≈ 63% of the time to break even. How often this trigger reaches its first target on NQ is not yet measured.
• Outcome. 29,605.50 traded at 09:58 ET: +$158.50 after costs. Price went on to 29,876.50 by 10:25, a move no written rule here would have held for. One day is one outcome, not evidence.

M1.05 labels this same morning an open rejection reverse at 10:30; this lesson’s job is the earlier decision at the 09:50 close.

The mirror and the failure

The short side (mirror). A drive up that is rejected at its high and closes a 5-minute bar back below the midpoint is the same setup upside down: sell the trigger close, stop beyond the high, first target the low of the range so far. Flip the figure to see it. Same arithmetic, same evidence: none for either side.

The failure: a trigger that stopped out (29 September 2026). NQ opened at 30,708.75, inside the prior range, and drove down to 30,545.50 by 09:43 ET (163 points). The high since the open was 30,725.50, so the midpoint was 30,635.50. Closes stayed under it until the 10:15 bar closed at 30,660.50. Long 1 MNQ there, stop 30,543.50 (117 points, $234), first target 30,725.50. The best price after entry was 30,706.25, 19.25 points short. At 11:26 ET the stop traded: −$236.50 after costs. Every check had passed; the auction did not follow through. A full stop is a normal outcome of this setup, and the reason the stop exists.

The failure that costs nothing: a drive that pauses (23 September 2026). NQ opened at 31,006.50 and fell 104.50 points to 30,902.00 by 09:39 ET. The bounce stalled at 30,925.00, far under the midpoint (30,957.50). No trigger, no trade. Price made new lows at 09:45, broke the prior RTH low (30,763.50) at 10:10 and reached 30,642.75 by 10:25. Anyone who bought the first bounce was long into a drive that resumed. A plan for an accepted drive is M5.11 (trend-day continuation), itself Lore, not an edge.

Prop-firm note. In a prop evaluation the stop must be a real order placed with the entry, and a 135-point NQ stop is large next to most daily loss limits: size in MNQ from your dollar risk. Evaluation accounts usually fill in simulation, simulated fills are not live fills, and tuning brackets to profit from simulated fills is treated as prohibited conduct by firms such as Topstep. The Prop Firm course (PROP1.04) covers the rules; check your firm’s help center before you trade.

Common mistake: Calling every red morning an opening reversal. A 30-point dip on a balance day is not a drive: there is nothing to fail. If the drive does not qualify, the setup does not exist.

Common mistake: Trusting a hand-drawn R:R. An earlier version of this lesson promised about 1.7R with the stop beyond the drive’s low and the target at the top of the range. With the entry at or past the midpoint that is impossible. Price the trade from the three prices, every time.

Common mistake: Treating a base rate as an edge. “The far side is not traded 89% of the time” sounds like a trade, but a placebo that keeps only ordinary daily distances already gives 86.5%. It tells you how far a typical session travels, not where to enter.

Drill: score twenty opening drives

1. Write your drive definition, your two trigger checks, the stop buffer and the first target before you open a chart.
2. In replay, go through 20 or more NQ opens. For each, decide by 10:00 ET: no drive, drive with no trigger, or drive with a trigger.
3. For every trigger, record entry, stop, first target, R at the first target, and which of stop or target traded first.
4. Compute your win rate and compare it with the break-even rate your stops and targets imply after costs.

What we don’t know yet

• How often the trigger in this lesson reaches its first target before its stop on NQ, against a placebo entry at the same time of day: not yet measured.
• Whether the drive’s size, its delta or where the open sits against the prior range changes that rate: not yet measured.
• Whether holding past the first target adds anything after costs: not yet measured.

Recap

• An opening reversal is an open drive that fails: a qualifying drive, a rejection at its extreme and a 5-minute close past the midpoint of the range so far, all written before the open.
• Enter at the trigger close, never during the drive; the stop goes beyond the extreme, and a resumed drive ends the idea.
• At the first target the setup pays less than 1R, so it needs a win rate above 50% plus costs, and that win rate is not yet measured.
• Open-location base rates keep the target realistic; they are mostly distance, and on NQ the open type has not forecast the day’s direction.

Quiz

Further reading

• Order Flow, Measured: all claims: the public practitioner study behind the first two evidence boxes (claims OI-02, OI-10 and OI-11).
• Trading Key Market Auction Reversals: failed-auction anatomy on higher timeframes, in auction-theory terms.
• Our lab: on QQQ from 2015 to 2025, the far side of the prior range was not traded on 89.2% of outside opens, against 86.5% for a distance-only placebo.

Prerequisites: M1.05 for the opening types and open location, M4.02 Reading the open for the gap and the five-minute house rule, and EV3.01 on sample size. Next: M5.07 Breakout–retest continuation, the first continuation setup, where the question flips from “did it fail?” to “was it accepted?”.

Educational content only, not financial advice. Futures trading involves substantial risk of loss. Real-day examples are illustrations, not trade calls.