Rejection entries

It is 10:15 ET (22:15 ICT). NQ has rallied 40 points into the prior day’s high at 25,640.00 and just printed 25,648.50 above it. Every tick higher feels like a better price to sell, and every tick higher could also be the start of the next leg. You already know the setup from M5.01: a probe beyond a mapped level that may fail. This lesson is about the click: what exactly has to happen before you fade the probe, where you are wrong, and how much that costs you, written so that you can score it afterwards.

Key term

Rejection entry (vào lệnh khi giá bị từ chối): an entry that fades an extreme after the auction has refused the new price: short after a failed probe of a high, long after a failed probe of a low Definition. It is one of the four entry families from M6.01. The setup is the failed probe; the rejection entry is the trigger and the invalidation you attach to it.

How it works

Three steps, in order

1. Probe (thăm dò): price trades beyond an extreme you mapped before the open: a prior-day high or low, the initial-balance extreme, the overnight high or low.
2. Stall (chững lại): the push stops. Write it as a check you can score, not a feeling. The house version used here: at least two 5-minute bars after the probe bar that do not exceed the probe high, each with a range smaller than the average of the three bars before the probe House rule.
3. Rejection (bị từ chối): a 5-minute bar closes back inside the level, below the high you probed (above the low, on the long side).

Enter on step 3, never on step 1 House rule. This is a process rule, not a forecast: on step 1 you have a question; after step 3 you have an answer you can write down, and a precise place where that answer is wrong. Whether waiting for step 3 pays better than entering on the wick, or than not trading at all, is not yet measured.

Many traders add an order-flow cue to the stall: delta turning negative while price holds near the high (a divergence), a thin print at the extreme tick, or a burst of buying that fails to extend. You may log these, but they are descriptions of the stall, not reasons to expect the turn; the evidence below says why. M4.01 owns the time, volume and close checks for acceptance; reuse them here rather than inventing new ones per trade.

Invalidation and size

The stop goes a small buffer beyond the probe extreme House rule. That is where the thesis “the auction refused this price” is plainly wrong: if price trades above the probe high again and holds there, the market accepted what you thought it refused. The distance is set by the probe, not by how much you would like to risk. Convert it to dollars before you size: if the distance at one micro contract is more than your per-trade risk budget, skip the trade rather than shrink the stop.

Size stays standard or smaller, never larger House rule. You are trading against the move that just happened, and the family’s appeal is a tight, well-defined stop, not a high hit rate. M6.09 owns the sizing formula.

Myth / hypothesis: “Fades work best late in a trend day, when initiative is exhausted, and the first touch of a level is where momentum is strongest, so never fade it.” Neither part has been tested in that form. The two open leads in the public practitioner study point elsewhere: delta divergence into a level’s first touch (+4.6 percentage points, CI −0.0 to +9.2, n = 567), and new 1-hour extremes made against the session’s direction so far (+2.5 pp, CI +0.6 to +4.4, n = 3,841, with a look-ahead check still pending). Both stay graded Lore until a forward test LoreOne trader's study. Treat “late in a trend” and “never the first touch” as hypotheses to log, not as filters with an edge.

Why it is still worth knowing. The three steps give you a shared way to describe what the auction did at an extreme, a trigger you can apply the same way every time, and a stop that is defined by the market rather than by your mood. That is what makes the trade scorable. Whether the trigger has an edge is a question your own log can start to answer.

What the data says

The level under the probe does not rescue it either. M5.01 shows the two results: on NQ, failed breaks of the prior day’s high or low did not reach the other side more often than comparable moves LoreOne trader's study, and in our lab’s placebo test mapped pre-market levels were respected no more often than random prices at the same distance Measured: no edgeNasdaq index. So the honest status of a rejection entry is a defined-risk rule whose edge is unproven. You can trade it small while you score it. You cannot call it high odds.

Worked example: the 25,648.50 probe

Synthetic session, NQ traded with MNQ (0.25-point ticks, $2 a point; NQ is $20 a point). It shows the reading, not a signal.

• Map, written before the open. Prior-day high 25,640.00. Scenario: if a probe above it stalls and a 5-minute bar closes back below, short to VWAP; no trade if the probe is accepted.
• 10:15 ET (22:15 ICT), probe. The bar trades to 25,648.50 and closes at 25,645.75. Step 1 only: no verdict.
• 10:20 and 10:25 ET, stall. Highs 25,647.75 and 25,648.25, both below the probe high; ranges 6.50 and 5.00 points against about 18 for the three bars before; delta −120 and +60 against +480 to +610 on the way up. Both stall checks pass. The flat delta is logged as a description, not as the reason for the trade.
• 10:30 ET, rejection. The bar closes at 25,632.00, back below 25,640.00. Step 3: short 1 MNQ at 25,632.00.
• Plan. Stop 25,651.00, 2.50 points (10 ticks) beyond the probe high: 19 points = $38. Target VWAP 25,604.00: 28 points = $56. Time stop: flat by 11:30 ET if neither trades.
• Break-even. Assume $1.50 of round-turn costs (check your own broker). A win nets $54.50 and a loss costs $39.50, so the trade needs to win 39.50 ÷ 94.00 ≈ 42% of the time to break even. How often this trigger reaches VWAP on NQ is not yet measured, so you do not know whether your rule clears 42%.
• Outcome on this synthetic day. VWAP trades at 11:00 ET: +$54.50 after costs. One day is one outcome, not evidence.

Note the size check. The same 19-point stop on 1 NQ is $380. The trigger’s precision is measured in points; your account’s survival is measured in dollars (M6.09).

The mirror and the failure

The long side (mirror). Below a mapped low, read the steps upside down: a probe under the level, two bars that make no new low with shrinking ranges, then a 5-minute close back above the level. Buy that close, with the stop a small buffer under the probe low and the target at the nearest reference above. Flip the figure to see it. Nothing in the evidence above favours one side, so treat the long version as equally unproven.

The failure. Same morning, different ending. After the short at 25,632.00, the 10:40 bar closes back above 25,640.00, and at 10:45 price trades through 25,648.50 to 25,651.25. The stop fills at 25,651.00: −$39.50 after costs, the planned loss. The next two bars close above 25,652 with volume building: late acceptance. The read at 10:30 was correct; the auction changed its answer. Late acceptance is a normal outcome of this family, not a mistake, and it hands the chart to the opposite family: a breakout entry (M6.02) above the accepted high, under its own rules.

Prop-firm note. In a prop evaluation the stop must be a real order placed with the entry, not a mental level. Evaluation accounts usually fill in simulation, and simulated fills are not live fills; tuning very tight brackets or automatic breakeven moves to win on simulated fills is treated as prohibited conduct by firms such as Topstep. M6.06 covers order types, and the Prop Firm course covers each firm’s rules; check your firm’s help center before you trade.

Common mistake: Shorting the first print above the level “because it is resistance”. That is step 1 of 3. Without an observed stall and a close back inside you have no trigger and no defined place to be wrong, only a guess that the top is in.

Common mistake: Treating the order-flow cue as the edge. “Delta divergence at the high” is one of the most repeated lines in order-flow teaching; on NQ the tested version came out at −1.1 percentage points. A cue that is easy to see is not a base rate.

Common mistake: Widening the stop after entry because the probe “has to fail”. The stop beyond the extreme is the whole definition of the trade. Moving it turns a planned $39.50 loss into an unplanned one.

Drill: score twenty rejection triggers

1. Before you open a chart, write your three steps with numbers: which levels count as extremes, your stall check, and the close that counts as back inside. Add your stop rule, target rule, time stop and cost per round turn.
2. In replay, find 20 probes beyond levels you would have mapped before the open. For each, record whether all three steps completed, and on which bar.
3. For every completed trigger, record which traded first, target or stop, and the result in R after costs.
4. Compute your win rate and compare it with the break-even rate your stop and target imply.

What we don’t know yet

• How often the full three-step trigger reaches its first target on NQ, against a placebo entry at the same time of day: not yet measured.
• Whether waiting for the close back inside beats entering on the wick or not trading: not yet measured.
• Whether the two open leads (divergence into a level’s first touch; extremes made against the session’s direction) survive a forward test: queued in our lab’s replication plan, not yet run; Lore until then.

Recap

• A rejection entry fades a probe that the auction refused: probe, stall, close back inside, written as checks with numbers, entered only on the third step.
• The stop sits a small buffer beyond the probe extreme; convert it to dollars before sizing, and skip the trade if it does not fit the budget.
• Size stays standard or smaller; the family offers a defined stop, not a high hit rate.
• On NQ, the stall cues traders read (divergence, exhaustion prints, failed heavy buying) showed no edge one by one, and the full trigger is not yet measured: score it before you trust it.

Quiz

Further reading

• Order Flow, Measured: all claims: the public practitioner study behind the evidence box (claims DA-01, DA-06, DA-07, FP-04).
• M1.07: a real NQ failed high from our lab’s tick recording, and why “trapped” describes the day without predicting the next one.
• Our lab: 226 pre-registered tests on 721 sessions found no pre-market level family respected more often than random prices at the same distance.

Prerequisites: M6.01 Setup vs entry, M5.01 Failed breakout at a level, M4.01 Acceptance. Next: M6.05 Continuation entries, the opposite choice: joining the move instead of fading it.

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