Grading the frameworks you already know

It is 10:00 ET (22:00 ICT, winter). NQ has just rallied 30 points in one five-minute candle, and your chart shows a shaded box under it: a fair value gap. The video you learned it from says price comes back to fill the gap and bounces from it. At 10:40 NQ dips into the box and bounces 30 points. It feels like the framework works. This lesson asks the EV1.04 question about that box: how often would a box of the same size, at the same distance, drawn by nobody, have filled and bounced too?

Key terms

Fair value gap, FVG (khoảng trống giá trị hợp lý): in three consecutive candles, a price range that the first and third candles leave untouched because the middle candle moved so far Definition. In an up move, the gap runs from the first candle’s high to the third candle’s low.

Order block (khối lệnh): the last candle of the opposite colour before a strong move, marked as a zone Definition.

Liquidity grab (quét thanh khoản): price trades a little beyond an obvious high or low, then comes back inside Definition.

Killzone (khung giờ “săn”): a fixed clock window, such as the London open or the New York open, that a framework says matters most Definition.

Distance-matched placebo (mức giả cùng khoảng cách): a zone of the same size, at the same distance from price, with no story attached Definition. EV1.04 built it for levels; here you build it for zones.

Term you need: ATR. Average true range: the average size of a day’s move over the last 14 days. A zone 0.1 ATR from price sits about a tenth of a normal day away, so price reaches it on most days by chance alone.

Chart first: what a fill rate is worth

How it works

Every framework term hides two claims

Each term carries a definition and a promise. The definition is geometry: three candles, a colour change, a clock window, graded Definition. The promise is what the course adds: the gap fills, the block holds, the grab reverses, the window moves. The promise is a market claim, and a market claim gets a grade.

Myth / hypothesis: courses sold as “smart money” or institutional concepts teach that large players leave footprints in these patterns: price returns to fill gaps because the move was “unfair”, order blocks hold because institutions defend their orders, highs are swept because stops sit above them, and the day’s real move starts inside a killzone. Graded Lore Lore: our lab has not yet measured any of these promises (search date 2026-11-09), and we know of no public test against a placebo.

Grade each promise, and say what would test it

Sort the cards first, then flip each one to see the test that would settle it.

The two cards that are not Lore are both narrower than the claim beside them. Volatility near the open is graded Plausible from standard literature pointers PlausibleOther market, and a busy window is not a direction. The sweep result is plausible on NQ PlausibleNQ, but it is one aggressive order eating several price levels on the tape (the live list of trades), not price poking above yesterday’s high.

Why fill rates look high

Price wanders, so a zone near price gets revisited on most sessions, like any random line Definition. On a driftless random walk, a price 0.1 ATR away is revisited before the close about 85% of the time; at 0.5 ATR, about 41%. A course can quote a large fill rate without the gap adding anything. Ordinary opening gaps behave the same way: our lab measured in sample on QQQ (2013–2026) that small ones under 0.25% filled about 85% vs about 19% for gaps over 2% (out of sample the largest-gap bucket was 26.7% on n = 15) MeasuredNasdaq index, and REG4.01 tests whether the prior close adds anything beyond distance.

The unfalsifiable rescue

Watch for rules that turn every outcome into a win. The gap fills: it worked. It does not: the move was “too strong”. Price slices through: now it is an “inverse” gap. A claim that reads every result as success cannot be tested, so it cannot rise above Lore Lore. Ask for the failure rule, written before the session.

Why it is still worth knowing

• Shared vocabulary. Many traders use these words; you can now name exactly what they drew.
• Describing the auction. A gap marks a stretch where one side traded fast on little volume, the same idea as single prints in a profile (M2.08).
• Framing risk. A zone gives you a stop and a size in ticks before you enter.
• A hypothesis you can test. Every card ends with a test; running it turns a belief into a number.

What the data says

Worked example: scoring one bullish gap

A synthetic day, labelled as such. NQ 5-minute chart, ATR 360 points. Before the session you write your rules: a fill is any trade inside the zone before 16:00 ET; a hold is a move 20 points away from the zone’s near edge before price trades 10 points beyond its far edge, within 60 minutes.

• Candle 1 (09:45 ET): high 20,412.50.
• Candle 2 (09:50 ET): runs from 20,410.00 to 20,441.25.
• Candle 3 (09:55 ET): low 20,421.75.

The gap is 20,412.50 to 20,421.75: 9.25 points, or 37 ticks of 0.25. At 10:00 ET NQ trades 20,446.00, so the gap’s top edge is 24.25 points below price (about 0.07 ATR).

The placebo: the same 9.25-point zone, the same 24.25 points from price, on the other side: 20,470.25 to 20,479.50. Draw both before 10:01.

What happened: at 10:40 NQ dips to 20,419.50, inside the gap: filled. By 11:15 it trades 20,452.00, 30.25 points above the gap, without ever trading 20,402.50: held. At 13:20 NQ rallies to 20,474.00, inside the placebo zone: placebo filled. By 13:55 it is at 20,448.00, 22.25 points below that zone, without trading 20,489.50: placebo held.

Your row: gap filled and held; placebo filled and held. The box nobody drew did exactly what the gap did. One row proves nothing; forty start to answer the question.

The mirror and the failure

Short side. A bearish gap forms when candle 1’s low sits above candle 3’s high, say 20,388.75 and 20,379.25: a 9.50-point zone above price. The placebo goes the same distance below price. The hold rule flips: 20 points down, away from the zone, before 10 points above its top.

Where it fails. On a trend day NQ climbs 200 points, closes on the high and never revisits the gap. The course calls that a “strong” day; your sheet says “not filled”. If the placebo was not filled either, the day says nothing about gaps, only that it trended.

Do it on the platform

Prop-firm note. A stop “below the order block” is still a stop in ticks against a firm’s loss limits: size it with RSK1.05, and read PROP1.03 before paying for an evaluation on an unmeasured setup.

Common mistake: treating a base rate as an edge. “Most gaps fill” is a base rate until you see what a same-size zone at the same distance did on the same days.

Common mistake: keeping only the clean examples. Log every gap your rule finds, not the ones you remember.

Drill: twenty gaps, twenty placebos

1. Write your fill and hold rules (points away, points through, minutes) before you start.
2. On replay or live NQ 5-minute charts, mark the first fair value gap after 09:45 ET each session for 20 sessions, with its placebo.
3. Fill one row per session: gap filled, gap held, placebo filled, placebo held.
4. Compare the two fill rates and the two hold rates.

Recap

• Each framework term is a definition plus a promise; the definition is geometry, the promise is Lore until tested.
• A fill rate means little without the rate of a same-size zone at the same distance on the same days.
• The busy-open part of killzones is graded Plausible from US futures references; the direction part is not yet measured.
• The one measured neighbour, the NQ tape sweep, is plausible, narrower than a liquidity grab, and in a descriptive study without modelled costs its ~1.92-tick move is smaller than a typical round trip in costs.
• A rule that reads every outcome as success cannot be graded above Lore.

Further reading

• Andersen and Bollerslev (1997), “Intraday periodicity and volatility persistence in financial markets”, Journal of Empirical Finance 4 (standard literature pointer on the U-shape in intraday volatility on S&P 500 futures, not re-fetched in our lab).
• Order Flow, Measured, the public practitioner study (dannyorderflow.notion.site), claim OB-03 on sweeps. One practitioner’s pre-registered study; intervals not clustered by session; at most Plausible.
• Freedman, Pisani and Purves, Statistics (4th ed.), the chapters on chance variation: why 20 rows give wide intervals.
• Our lab: NQ sweeps through 3 or more price levels continued slightly less than matched single-price bursts over 53 sessions; the forward test is pending and in this descriptive event study (no costs modelled) the ~1.92-tick move is smaller than typical round-trip costs.

Before this: EV1.02 and EV1.04 build the placebo you used here. Next: EV3.01 shows how many rows your gap sheet needs before the difference means something; M5.05 teaches the sweep result in full.

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