Absorption reversal

It is 10:04 ET (22:04 ICT, winter). NQ has drifted down into the overnight low on your sheet. For five minutes sellers hit the bid hard: 600, 800, 900 contracts a minute, two to three times the half-hour average. Price prints one tick under the level and goes nowhere. Someone is buying everything that is sold.

Traders call this an absorption reversal and teach it as one of the cleanest setups on the footprint. This lesson shows how to trade it as a written rule, and what the data says about the print once you already have the level.

Key term

Absorption reversal (đảo chiều do hấp thụ): a setup that starts at a mapped level where heavy aggressive volume makes almost no price progress, and enters only after price moves back away from the level Definition. M3.02 owns the read of the print itself (effort without result, delta, range). This lesson adds the level, the trigger, the stop and the target.

Term you need: release. The release is the first bar that closes back away from the absorbed price, in the direction opposite to the aggressors, with delta turning that way. It is a trigger you write down before the session, not a fact about who won House rule.

Chart first: the same print at a level, three endings

How the setup works

The four parts

• Context. A level you mapped before the session: prior-day high or low, overnight high or low, an initial-balance extreme Context. The level tells you where to watch. It does not tell you that price will turn there: on our lab’s tests, pre-market levels were touched, held and broken at the same rate as random prices at the same distance Measured: no edgeNasdaq index (second box below).
• The print. Heavy aggressive volume at or within a tick or two of the level, a small range for that volume, delta strongly with the aggressor, and no extension Definition. Size thresholds such as “300 contracts” are conventions you choose and keep fixed, not calibrated numbers.
• Trigger: the release. No position during the absorption. Absorption can go on for many minutes, and you cannot tell from inside it whether the resting side is defending or simply filling a large order. The trigger is the release, inside a time window you set in advance (for example 15 minutes after the last absorption bar). If the window passes with no release, there is no trade House rule. Whether waiting for a release improves the result is not yet measured: M4.09 shows how to write it as a rule you can score.
• Invalidation and target. The stop goes a few ticks beyond the absorption extreme: if price trades through the absorbed price, the read was wrong House rule. Set the target in R (for example 2R), and note references such as VWAP or the IB midpoint only as places where you may manage the trade, not as places price is drawn to.

The story traders tell

Myth / hypothesis: “The absorber fills every order against the push. When the absorbed side gives up, their stops pile on and the release is violent, so absorption at a level is the best reward-to-risk reversal there is.” Lore It is a coherent story, and the tight stop is real arithmetic. What has not been shown is the forecast: on NQ, absorption at a key level did not reverse measurably more often than the same level without it (first box below). A good-looking ratio on paper does not help if the trade fails about as often as a touch without the print.

Why it is still worth knowing

You will meet this setup everywhere, so you need the words for it. The print describes the auction precisely: where effort met resting size, and at which price. It gives you a cheap, objective invalidation, the absorbed extreme, which is worth more than a stop placed at a round number. It keeps you out of the worst entry, buying or selling into the print while it is still printing. Footprint signals of this kind also tend to widen moves both ways, which matters for size and stops rather than direction PlausibleOne trader's study (M5.14). And it is a hypothesis you can test on your own levels with a written rule and a baseline (the drill below).

What the data says

Put the two boxes together. The level does not beat a random price, and the print at the level does not measurably beat the level alone. So the setup is graded Lore: a disciplined way to take a position with a defined stop, not an edge.

Worked example: long at the overnight low

Synthetic session, built to the figure above, not a real date. NQ 1-minute footprint, tick 0.25 points: $5 per tick on NQ, $0.50 on MNQ.

Before the open you write: “Overnight low 21,240.00. If sellers are absorbed there, long on the release: a 1-minute close above the absorption bars’ highs with positive delta, within 15 minutes of the last absorption bar. Stop 4 ticks beyond the absorption extreme. Target 2R.”

10:03–10:08 ET (22:03–22:08 ICT). About 3,900 contracts sell into 21,240.00–21,239.75. Delta runs −600 to −900 a bar. The low is 21,239.75, one tick through the level; no bar’s high exceeds 21,242.75. That is absorption by your definition. You do nothing.

10:12 ET. A bar closes at 21,243.50 with delta +420: the release, inside the window. Long 1 MNQ at 21,243.50.

• Stop: 21,238.75, four ticks under the absorption extreme. Risk 4.75 points = 19 ticks = $9.50 on 1 MNQ ($95 on 1 NQ), before costs.
• Target: 2R = 9.50 points, at 21,253.00. Session VWAP at 21,256.25 is noted as a reference, not the target.
• Invalid: a 1-minute close below 21,239.75 before entry cancels the plan.

At 10:31 the target fills: +$19 on 1 MNQ before costs. That is one ending of three, and the boxes say it is no more likely here than at the same level without the print.

The mirror and the failure

Short mirror (same rule, flipped). Overnight high 21,310.00, another synthetic morning. Five 1-minute bars print 400–600 contract buy imbalances at 21,308.00–21,310.00; delta per bar falls from +800 to +150 while price stalls. You wait. A final probe wicks to 21,311.50 and drops, and the next bar closes at 21,305.75 with delta −500: the release. Short 1 MNQ at 21,305.75, stop 21,312.50 (three ticks above the extreme): 6.75 points = 27 ticks = $13.50 on 1 MNQ ($135 on 1 NQ). Target 2R = 13.50 points, at 21,292.25.

The failure. Same short, same fill. Two bars later buyers come back, lift through 21,308.00, and a bar closes at 21,312.75 with delta +900. The stop at 21,312.50 fills: −1R, $13.50 on 1 MNQ plus costs and any slippage. The “wall” was not defending the high; someone was buying size, and the release was a pause. That is the planned loss, and it is why the stop is not optional: when price accepts through the absorbed price, the read is wrong and the trade is over.

Prop-firm note. The stop beyond the absorption extreme is the number your evaluation’s loss limits see. Size from the stop distance (RSK1.05), not from how strong the print looks; the Prop Firm course covers daily and maximum loss limits.

Common mistakes

Common mistake: Treating the level as the evidence. “Absorption at the overnight low reversed 7 of my last 10 times” says nothing until you know how often ordinary touches of the same level reversed. On our lab’s tests the levels themselves performed at the placebo rate.

Common mistake: Front-running the wall. Shorting into buy absorption because “it can’t hold forever” is guessing which ending comes next. Absorption can last far longer than your patience, and sometimes the resting side is accumulating rather than defending, so the eventual move goes with the push. The release is your written trigger; without it there is no trade.

Common mistake: Moving the stop when price trades through the absorbed price. That is the exact moment the setup tells you it is wrong. Widening the stop turns a planned 1R loss into an unplanned one.

Drill: write the setup so it can be scored

Use the rule from the worked example: release = a 1-minute close beyond the absorption bars’ opposite edge with delta in the new direction, within 15 minutes of the last absorption bar; stop 4 ticks beyond the absorption extreme; target 2R. Classify each synthetic event as trade, no trade (window expired) or no trade (invalid), and for any trade give the entry, stop, target and risk on 1 MNQ.

1. Sell absorption at 21,240.00 from 10:03 to 10:08 (extreme 21,239.75, highs at most 21,242.75). At 10:12 a bar closes at 21,243.50 with delta +420.
2. Buy absorption at a prior-day high of 21,330.00 (extreme 21,330.75, lows at least 21,326.25). For the next 25 minutes price chops between 21,326.50 and 21,330.00; no bar closes below 21,326.25.
3. Sell absorption at 21,240.00 (extreme 21,239.75). Before any release, a bar closes at 21,238.50 with delta −1,100.

What we don’t know yet

• Whether waiting for the release improves the result at a level compared with entering on the touch or not trading: not yet measured. Order-flow confirmation at a level is on the lab’s forward-test list (M4.09 writes the rule).
• Whether absorption read from order-book depth (resting size before and after the hits) behaves differently from absorption read off the footprint: not yet measured; the lab has no depth history yet.
• Our lab’s own replication of the public absorption results on NQ: planned, not run.

Recap

• An absorption reversal is a written setup: a mapped level, heavy aggression with little progress, then a release inside a time window.
• The release is the trigger; entering while the print is still printing is guessing the ending.
• The stop goes beyond the absorption extreme, and the target is set in R.
• Graded Lore: absorption at a level added nothing measurable beyond the level, and the levels did no better than random prices.
• Log triggered trades, stalls and breaks alike, so the record can be compared with a baseline.

Further reading

• Order Flow, Measured: All claims. A public, pre-registered practitioner study on NQ, 2018–2026. Read the absorption rows (FP-01, FP-02) and the “Absorption test” page.
• Our lab: pre-market level families vs placebo, 721 sessions and 226 pre-registered tests; no level family beat random prices at the same distance.
• M3.02 for the print itself; M4.01 for acceptance, the check behind the release and the invalidation.

Next: M5.03 applies the same frame to exhaustion into a level: a climax bar instead of a stalled one, and the same question about what the print adds.

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