Footprint Patterns, Honestly

≈33 min read · Real gold futures examples with XAUUSD equivalents · Free

In this chapter: You will meet the best-known footprint patterns — exhaustion, finished and unfinished auctions, absorption, effort versus result, trapped traders, delta-POC shift and CVD divergence — and you will learn to define each one precisely, to notice when different sources use the same word for different things, and to treat every pattern as a way of reading context rather than as a standalone signal. The most important skill in this chapter is a question: how often does this pattern appear, and when it appears, how often does it actually mean anything more than an ordinary bar would? By the end you will be able to read a full gold footprint bar step by step, and you will see how a single bar can trigger two "signals" pointing in opposite directions.

A word before we start. Footprint patterns are usually taught with carefully chosen screenshots: the exhaustion top that marked the high of the day, the absorption at the low that held. Those pictures are real, but they are a selection. The bars where the same pattern appeared and nothing happened are rarely shown. In our own testing of footprint rules on gold, the honest conclusion was the same one you will meet throughout this book: patterns on their own did not reliably predict direction after costs. That does not make the footprint useless. It makes it a reading tool. This chapter teaches the patterns as vocabulary for describing what happened at important places — not as a list of triggers.

Gold note: All examples in this chapter use hypothetical GC (gold futures) prices around 4210. XAUUSD (the spot CFD) has typically traded roughly $27–30 below front-month GC, so 4210 GC is very roughly 4180–4183 on XAUUSD. The gap changes over time; always measure it live and treat it as approximate.

Exhaustion: Thin Volume at the Extreme

Exhaustion describes the idea that the aggressive side has "run out" at the end of a move. Price makes a new high, but at the very top rows of the bar almost nobody trades. The buyers who were pushing price up have stopped arriving. On a footprint, this shows up as very little volume in the extreme rows compared with the rest of the bar — for example, compared with the bar's POC row. Some vendors describe it as a drop-off in follow-through volume.

In auction terms (Chapter 7), the auction has reached a price where there are no more aggressive buyers willing to transact. It is a one-bar cousin of excess at the end of a session profile.

A hypothetical exhaustion top

The bar pushed to a new high at 4214.0, but the top two rows hold only 11 contracts out of 410, and the top row holds 2 contracts against 133 at the POC. Price then closed five ticks below the high.

Making it precise

To study a pattern honestly, you have to define it precisely enough that a computer could find every instance. Here is a made-up teaching definition of an exhaustion top (the numbers are illustrative, not a recommended setting):

1. The bar makes the highest high of the last several bars.
2. The top row's total volume is less than one-fifth of the POC row's volume.
3. The top two rows contain less than 15% of the bar's aggressive buying (Ask volume).
4. The bar closes at least a third of its range below the high.

The bar above passes all four: top row 2 versus POC 133; top two rows' Ask = 8 out of 213 total Ask (under 4%); close 4213.5 is five ticks below a high that sits six ticks above the low.

The bottom version is the mirror image: a new low, very thin volume in the lowest rows, little aggressive selling there, and a close well above the low.

The honest problem: most bars look like this

Here is the uncomfortable fact. Thin volume at the extremes is the normal shape of almost any bar. Volume gathers in the middle of a bar because price spends most of its time there; price usually touches the high and low only briefly, so the extreme rows naturally hold few contracts. Draw the volume histogram of a typical one-minute bar and it bulges in the middle and thins at both ends.

That means "low volume at the extreme" is mostly a description of an ordinary bar, not of a special event. When we tested a reasonable-looking exhaustion rule on gold, it fired on a large share of all one-minute bars — far too often to say anything meaningful on its own about the end of a move. A signal that is lit up on a large fraction of bars cannot, by itself, tell you that this bar is special.

When exhaustion is worth noticing

Exhaustion becomes interesting only when context makes the bar unusual:

• it happens at an important reference level — the previous day's high (PDH), the overnight high (ONH), the value area high (VAH);
• it comes after an extended move, not in the middle of rotation;
• the bar carries meaningful total volume — a New York bar, not a 40-contract Asian-session bar;
• and ideally the next bars show no follow-through: price does not come back to test the high.

Even then, exhaustion raises the possibility of a pause, not the certainty of a reversal. Aggressive buyers stopping is not the same thing as aggressive sellers arriving.

Key idea: Thin volume at a bar's extreme is normal. Exhaustion is worth attention only when the location, the prior move, the bar's volume and the following bars make it unusual — and even then it suggests a pause, not a reversal.

Common mistake: "Exhaustion on the footprint means price will reverse." This is a popular claim, not proven. A related mistake is confusing exhaustion with absorption: exhaustion is aggression stopping; absorption is aggression continuing but being soaked up by passive orders. They look different on the footprint and mean different things.

Finished vs Unfinished Auction at Bar Extremes

In Chapter 7 you met poor highs and poor lows on a session profile. The same idea exists at the scale of one footprint bar, and it relies on the edge rows that had no diagonal partner in Chapter 8.

Finished auction

At the top row of a bar, look at the Bid side. If it is zero — for example 0 × 5 — then at the high, only aggressive buyers traded; no aggressive seller hit a bid at that price. The market pushed up until buying stopped, and then it turned. That high is called a finished auction. At the low, the mirror case is an Ask of zero, such as 7 × 0.

Unfinished auction

If both sides traded at the extreme — say the top row is 12 × 9 — then even at the highest price, some aggressive sellers were hitting bids there. For a trade to happen on the bid at the high, the offer must at that moment have been one tick higher, and nobody bought it. The auction in that direction did not reach a clean end. That is called an unfinished auction, and the common interpretation (found in several vendors' educational material) is that price may return later to "finish" it.

Two definitions, two answers

Bar C shows the problem. The simple definition says an extreme is unfinished if both sides are greater than zero. Under that rule, 1 × 2 — three contracts in total — counts as unfinished. In practice, that is noise.

A threshold definition requires both sides to be meaningful. For example (illustrative only): the extreme is unfinished only if the smaller of its two numbers is at least 10 contracts. Under that rule, Bar B (12 × 9) is finished, because 9 is below 10, and Bar C is clearly finished. Raise or lower the threshold and labels change.

Neither rule is "correct". The point is that "unfinished auction" means different things in different tools, and the same row can be labelled finished by one and unfinished by another. You will see an example of exactly this in the walkthrough at the end of the chapter.

"They always get revisited" — and the baseline

The popular claim is that unfinished auctions almost always get revisited. This is a popular claim, not proven. Even vendors who teach the concept note that there is no guarantee, and that in strong trends such levels can be left behind.

There is also a statistical trap. An unfinished high on a one-minute bar is usually only a few ticks from the current price. Price wanders a few ticks back and forth constantly. So "the unfinished high got tested in the next few bars" will happen very often by chance alone. To claim the pattern adds anything, you would have to compare it with a baseline: how often does price return to any price the same distance away, in the same time? If the two rates are similar, the unfinished label added nothing.

Key idea: Finished = one side is zero at the extreme. Unfinished = both sides traded there. An unfinished extreme is a possible reference level, not a target. Definitions differ, and "it got revisited" must be compared with how often a random nearby price gets revisited.

Common mistake: Labelling a 1 × 2 row as an unfinished auction, and counting the "revisit" of an unfinished high that was only one tick away as a success for the pattern.

Absorption: Three Definitions, One Word

Absorption at its core means: passive limit orders soak up the other side's aggression, and price does not move despite the aggression. A classic description: lots of aggressive selling hits the bid, but the bid keeps getting refilled at the same price and price does not go lower. Delta becomes more and more negative, yet price stays put. Sometimes an iceberg order — an order that shows only a small part of its size and refills as it is consumed — sits behind this, but not necessarily.

The trouble is that the word "absorption" is used for at least three different things, with very different quality of evidence.

Definition 1: the order-book definition (most direct)

You watch the order book (DOM or liquidity heatmap, or full order-by-order data called MBO, "market by order") and see a resting limit order at one price being consumed again and again and refilled, while price fails to pass through it. This is the most direct evidence, because you can see the passive orders themselves. A footprint alone cannot show this, because a footprint does not show resting limit orders — only the trades that hit them. Chapter 10 covers these tools.

Definition 2: the footprint proxy (indirect)

On the footprint, you see heavy volume and strongly one-sided delta in one zone, with little price progress. For example, large Bid numbers (aggressive selling) in the bottom rows of a bar, but the bar closes well above its low. Here is a hypothetical case:

Aggressive sellers traded 267 contracts in the bottom two rows, yet price never got below 4205.1 and the bar closed three ticks above its open, five ticks above its low. Something absorbed that selling — very possibly resting bids. But the footprint cannot confirm it. The same picture could also be the end of a selling wave that simply ran out, with buyers stepping in afterwards.

This is a proxy: indirect evidence that is consistent with absorption, not proof of it.

Definition 3: candle shape (weakest)

"A long wick plus high volume equals absorption." This definition calls almost every high-volume pin bar absorption. When we looked at candle-shape proxies of this kind, they fired on many bars and were rarely precise. They are the weakest of the three.

Presence is not outcome

Even with the best evidence, absorption shows the presence of a large passive participant. It does not tell you the outcome. Whether that level holds or breaks depends on how much size the passive trader still has left and how long they want to defend the price — and that is invisible. Absorption can, and regularly does, break when the passive size runs out. A common teaching guideline is not to act on absorption alone, but to wait and see whether the other side actually starts participating (for example, aggressive buyers appearing after absorbed selling).

Key idea: Absorption = passive orders absorbing aggression while price stalls. The order book confirms it most directly; the footprint offers an indirect proxy; candle shape alone is weak. Always ask which definition a source is using. Absorption shows presence, not outcome.

Common mistake: "Absorption at the low means a safe buy." This is a popular claim, not proven. Also avoid calling every high-volume pin bar "absorption", and assuming the footprint can directly show iceberg orders.

Effort vs Result: Wyckoff on the Footprint

Richard Wyckoff, an early twentieth-century market analyst, described several "laws" of market behaviour. One of them is the law of effort versus result: volume is effort; price movement is result.

When they agree — heavy volume and a large move in the same direction — the move is "in harmony". When they disagree — very heavy volume or delta but little price movement — someone on the other side was neutralising the effort. Often that is the absorption you just met.

A sharper measure on the footprint

A candle chart only has total volume. The footprint gives a sharper measure of effort: delta, the net aggression of one side. A simple ratio compares effort with result:

Efficiency = |close − open| (in ticks) ÷ |bar delta|

That is, how many ticks of movement each contract of net aggression bought.

Both bars had almost the same net aggressive buying. Bar X turned it into twelve ticks ($1.20 on the price, $120 per GC contract). Bar Y turned it into one tick. If Bar Y printed right at the overnight high, you would describe it as big effort, small result — aggressive buyers did not move price, which suggests passive sellers were present. Some traders call this an aggression failure.

What counts as "low" efficiency? There is no universal number. You would need to know what is typical for that session: measure the efficiency of ordinary bars in the same session on your own data, and look for bars far below that norm.

Context decides

At a major resistance level, effort without result plausibly reflects passive sellers defending the level. In the middle of a quiet range, the same numbers may be nothing more than normal back-and-forth trading where buyers and sellers took turns.

Limitations

1. The ratio is sensitive to close minus open. A bar with a large range that closes near its open has an efficiency near zero, even if it travelled a long way in both directions.
2. Normalise by session. A delta of 100 in the Asian session and a delta of 100 at the New York open are different amounts of effort relative to what is normal.
3. It is a reading rule, not a formula that guarantees anything. The same idea underlies Volume Spread Analysis (VSA), where it is also a way of reading bars, not a mechanical edge.

Key idea: Effort (volume, delta) should produce result (price movement). When it does not, someone was on the other side. Measure "low" against the same session's norm, and interpret it only at meaningful locations.

Common mistake: "High volume and a small bar always means reversal." Not true. Also avoid comparing raw delta across sessions, and ignoring that a bar can have a big range but a small body.

Trapped Traders

Trapped traders are traders who entered aggressively at a price and then watched price move straight against them. On the footprint, the picture for trapped buyers is: strong buying aggression — a buy imbalance or a buy stack, and strong positive delta — in the upper part of a bar, but the bar closes below that zone, near its own low. Anyone who bought aggressively in that zone is now holding a loss.

Here is a hypothetical example (diagonal imbalances at 3:1):

There is a three-row buy stack at 4218.1–4218.3 and a bar delta of +160 (403 Ask against 243 Bid), yet the bar closed at 4217.8, three ticks below the bottom of the stack and in the lower part of its range. The buyers who lifted the offer at 4218.1–4218.3 are underwater.

The possible logic

If price keeps moving lower, some of those buyers may be forced out — by choice or by stop orders — and closing a long position means selling, which can add to the downward pressure. And if price comes back up to their entry zone, some may sell just to get out at break-even, which can make that zone act as resistance.

A made-up teaching definition

For research, you need precision. An illustrative definition of trapped buyers (numbers are for teaching only):

• a buy stack whose lowest row is above the bar's close;
• positive bar delta;
• the close sits in the lower third of the bar's range.

The example above passes: the stack's lowest row (4218.1) is above the close (4217.8); delta is +160; the range is 4217.7–4218.4 (seven ticks), and the close is one tick above the low. The mirror image describes trapped sellers.

Important limitations

1. "Trapped" is an assumption about intent. The footprint cannot tell whether those aggressive buys were new long positions or short sellers closing out. If they were shorts covering, nobody is trapped — they are now flat.
2. Trapped traders may be hedged, or may simply be willing to hold through a larger loss.
3. The pattern is only known when the bar closes, and the very next bar can reverse straight back up.
4. The same footprint shape is common in unimportant places, such as the middle of an Asian-session range. The idea carries most weight at a failed breakout of a reference level — for example, a push above the previous day's high that is rejected back below it.

Key idea: Trapped traders = strong aggression in a zone, then a close on the other side of it. The logic is that forced exits may add fuel and the entry zone may become a reference. But intent is invisible, so "trapped" is always an assumption.

Common mistake: "Trapped traders always cause a strong move the other way." This is a popular claim, not proven. Also avoid assuming every aggressive buy was a new position, and seeing a "trap" in every bar with an upper wick.

Delta-POC and Its Shift

In Chapter 8 you met the volume POC of a bar: the row with the most total volume, the place of greatest agreement. A second, related measure is the delta-POC: the row with the largest absolute row delta, |Ask − Bid|. It marks where one-sided aggression was most concentrated.

The two can sit at very different prices. The volume POC may be a row where both sides traded heavily and delta was close to zero; the delta-POC is the row where one side clearly dominated.

At 4212.3, both sides traded heavily — agreement. At 4212.4, aggressive buyers clearly dominated — aggression.

Delta-POC shift

The delta-POC shift compares this bar's delta-POC with the previous bar's, in ticks. Read the sign of the delta at the delta-POC as well as its position.

From 10:02 to 10:04, the concentration of aggressive buying moved higher: traders were willing to buy urgently at rising prices. At 10:05 the delta-POC moved up again but its sign turned negative — the strongest aggression in that bar was now selling, at higher prices. That is a description of where aggression concentrated and how it moved. It can sit alongside POC migration (Chapter 8) and cumulative delta.

Honest limitations

1. In thin bars the delta-POC jumps between rows; large shifts may be pure noise.
2. Ties decide the output. When two rows have almost the same |delta|, the tie-break rule your software uses (for example "pick the lower price") determines the delta-POC — the market did not.
3. It is a derived metric. It does not come from the classic auction or Wyckoff literature, and there is no public evidence that it predicts anything. Treat it as a descriptive column.

Key idea: Volume POC = where most trading happened (agreement). Delta-POC = where one side was most aggressive (aggression). Its shift from bar to bar, with its sign, describes how aggression is moving.

Common mistake: Confusing delta-POC with volume POC; reading meaning into one- or two-tick shifts in thin bars. "Delta-POC shift predicts the next bar's direction" is a popular claim, not proven.

CVD Divergence — and Why It Often Fails

Cumulative volume delta (CVD), from Chapter 3, is the running total of delta from a starting point called the anchor — for example the start of the session. CVD divergence means price and CVD disagree: for example, price makes a higher high while CVD makes a lower high.

Price went six ticks higher on the second push, but the running total of aggressive buying minus aggressive selling was 430 contracts lower. The classic interpretation: "price rose, but fewer aggressive buyers pushed it there, so the move is weak."

That sounds compelling. On a chart in hindsight, the examples that "worked" look beautiful. In practice it fails often, for at least five mechanical reasons.

Reason 1: passive orders move price too

CVD counts only aggression. If sellers with resting limit orders raise their offers or pull them, price can rise on very little aggressive buying. The higher high is real; the lower CVD just means it took less aggression to get there.

Reason 2: absorption can last a long time

There is no law that says a divergence must "resolve". A large passive participant can keep absorbing aggression for a long time, and the market can stay divergent while price continues in the same direction.

Reason 3: anchor dependence

CVD depends on where you start counting. If both swing points come after the anchor, moving the anchor earlier or later shifts the whole CVD line up or down by the same amount, and the comparison between the two highs does not change. But traders very often compare points that sit on different sides of an anchor — a New York high against an overnight high, today against yesterday — and many platforms reset CVD each session. Then the anchor decides whether a divergence exists at all. A divergence on a session-anchored CVD may not exist on a weekly-anchored one.

Reason 4: frequency

On short timeframes, small divergences form all the time. The ones that "worked" and the ones that did not look identical at the moment they form. You only know which kind it was afterwards.

Reason 5: data quality accumulates

Bid/ask classification errors in fast moments (Chapter 8) are small for any one bar, but CVD adds them all up. Over a long session, those small errors accumulate into the running total.

A better use

Some of the best-known order-flow educators state plainly that CVD is not a tool for rule-based signals. A more useful approach is to treat a divergence at a reference level as one piece of the puzzle — for example, a second test of the overnight high on weaker CVD — and to read it together with acceptance or rejection at that level and with the footprint at that moment.

Key idea: CVD divergence is price and running delta disagreeing. Passive orders can move price without aggression, divergences need not resolve, anchors matter when comparing across them, and divergences that fail look identical to ones that work when they form. Use it at levels, alongside other evidence.

Common mistake: "CVD divergence is a reliable reversal signal." This is a popular claim, not proven. Also avoid comparing CVD highs and lows with different anchors on different days, and remembering only the divergences that worked — a form of survivorship bias when reviewing charts.

Base Rates and False Positives: How to Judge Any Footprint Pattern

Nearly every footprint pattern "works" when you scroll back through old charts. That is not because the pattern works; it is because our memory and our eyes are drawn to the examples that fit. To judge a pattern honestly, you need three numbers.

Number 1: fire rate

On what share of bars does the pattern appear? The higher the fire rate, the more the pattern is just a description of an ordinary bar rather than a special event. Remember the exhaustion lesson: a rule that is true for a large fraction of bars cannot, by itself, single out a special moment.

Number 2: precision

When the pattern appears, how often does the claimed outcome actually follow? The outcome must be defined precisely and in advance — for example, "price moves 5 ticks against the prior direction before it moves 5 ticks further in the same direction, within the next 10 bars."

Number 3: baseline

How often does that same outcome follow any bar — or a random bar? This is the number most people never compute, and it is the one that matters most. Precision without a baseline is meaningless.

A worked example with invented numbers

These numbers are made up purely to illustrate the arithmetic. Suppose you study 1,000 one-minute bars:

• Fire rate = 260 ÷ 1,000 = 26% of bars.
• Precision = 143 ÷ 260 = 55%. "The pattern is right 55% of the time!"
• Baseline = 535 ÷ 1,000 = 53.5% for all bars (and 392 ÷ 740 ≈ 53% for bars where the pattern did not fire).

The pattern adds about one or two percentage points over doing nothing at all. That is very little information — and with a sample like this it may not even be distinguishable from chance. Then add the spread and commissions: a tiny edge before costs can easily become nothing after costs. A screenshot of the 143 successes would look convincing. The table shows what it is really worth.

Common traps

• Conflicting detectors. One bar can trigger a bearish pattern and a bullish pattern at the same time — you will see exactly this in the walkthrough at the end of this chapter. If you treat both as signals, any outcome "confirms" one of them.
• A loose outcome definition. "Price reacted" is true almost every time. If the outcome is not defined precisely before you look, the pattern cannot fail.
• Choosing settings after seeing results. Adjusting the ratio or threshold until the pattern fires on the nice examples is overfitting (Chapter 12).
• Ignoring costs. Even precision slightly above the baseline may leave nothing after spread and commissions.

The practical conclusion

Read the footprint with the question "at this reference level, what did aggression do?" — not with the instruction "pattern X fired, act on it".

Key idea: Judge any pattern with three numbers: fire rate, precision and baseline. A high fire rate means little information; precision without a baseline means nothing; define the outcome first and never tune settings after seeing results.

Common mistake: Believing claims like "this pattern has an 80% win rate" when nobody has defined the outcome, the baseline or the data. That kind of claim is a popular one, not proven. Reviewing old charts and counting only the successes, or judging a pattern from a handful of screenshots, are the same mistake.

Try it: Pick one pattern from this chapter and write a precise definition and a precise outcome on paper. Before you look at any results, also write how you will measure the baseline. Then go through one week of GC one-minute bars and fill in the 2 × 2 table yourself.

Common Footprint Mistakes

Most footprint mistakes come from a single root: reading detail without context. Here are ten of the most common, each with a practical check that replaces it.

There is also a behavioural mistake: overtrading because the footprint seems to show a "signal" in every bar. A busy gold footprint shows dozens of small events per minute. Most of them are noise. The footprint is most useful when you bring a small number of specific questions to a small number of important places in the day — the touch of yesterday's high, the test of the value area low, the break of the initial balance.

Key idea: Context first, footprint second. Compare bar volume with the session, read GC with diagonal imbalances, fix your settings and definitions, separate delta from direction and divergence from reversal, count fire rates, and read sequences rather than single bars.

Common mistake: "The more I watch the footprint, the better I trade." More watching tends to produce more imagined signals. Another: "The footprint alone is enough to trade." That is a popular claim, not proven. A third: copying a video creator's settings, built for a different market, onto gold.

Try it: Take one 30-minute stretch of GC. First look at the footprint with all its colours and count how many "signals" you think you see. Then mark only five reference levels on the chart and look at the footprint only at the moments price touches one of them. Compare how many questions you actually needed to answer.

Full Walkthrough: Reading One Gold Footprint Bar

Now let us read a complete bar in a fixed order, using everything from Chapters 7, 8 and 9. The bar is hypothetical — a one-minute GC bar with symbolic prices — and the numbers are designed to teach, not to represent a typical session.

• Open = 4210.4, High = 4211.0, Low = 4210.2, Close = 4210.6 (roughly 4180–4184 on XAUUSD, using the approximate gap).
• Teaching settings: diagonal imbalances, ratio 3:1, minimum volume 20.

Step 1: context first

Where is this bar? Is it sitting on the previous day's low, testing the overnight high, at the value area edge — or in the middle of nowhere? Without this, everything that follows is only half meaningful. Write the location down before reading a single number.

Step 2: totals

• Volume = 630 contracts.
• Ask total = 353; Bid total = 277; bar delta = 353 − 277 = +76.
• The bar is green: open 4210.4, close 4210.6, two ticks higher.

Step 3: POC and delta-POC

• Volume POC = 4210.4 (162 contracts), in the lower part of the bar. The main agreement happened low.
• Delta-POC = 4210.6 (+59). But look at 4210.4: −58. The two largest one-sided rows are almost a tie in size (59 versus 58), with opposite signs. Had one contract been classified differently, the delta-POC would have been the selling row instead. This is the tie problem from earlier in the chapter in action.

Step 4: imbalances

Check every row diagonally:

• Sell imbalance at 4210.4: Bid(4210.4) = 110 ÷ Ask(4210.5) = 30 = 3.7. Heavy aggressive selling at the bar's busiest price.
• Buy imbalances at 4210.6, 4210.7 and 4210.8: 81 ÷ 26, 71 ÷ 22 and 64 ÷ 20, all just above 3. That is a three-row buy stack.
• No other row qualifies: for example 4210.9 is 22 ÷ 18 = 1.2, and 4210.3 is 24 ÷ 9 = 2.7 (below 3).

Step 5: extremes

• High 0 × 5: no aggressive seller traded at 4211.0 → finished under any definition.
• Low 9 × 4: both sides traded. Under the simple rule (both > 0) it is unfinished. Under the illustrative threshold rule from earlier (the smaller side must be at least 10), it is finished, because 4 is below 10. The same row, two labels.

Step 6: exhaustion at the top?

Apply the made-up exhaustion definition from the start of this chapter, and suppose this bar did make the highest high of the last several bars:

• Top row volume 5 versus POC 162: far below one-fifth. ✓
• Top two rows' Ask = 5 + 22 = 27, out of 353 total Ask: about 7.6%, below 15%. ✓
• Close 4210.6 is four ticks below the high, in an eight-tick range: half the range, more than a third. ✓

So the bearish exhaustion pattern fires.

Step 7: absorption at the bottom?

Now apply an equally made-up footprint absorption proxy: "at least a fifth of the bar's total volume is aggressive selling in the bottom third of its rows, and the bar closes in the upper half of its range."

• The bottom third of the nine rows is 4210.2–4210.4. Aggressive selling there = 9 + 60 + 110 = 179, out of 630: about 28%. ✓
• Close 4210.6 is exactly at the midpoint of 4210.2–4211.0 — just inside the "upper half" if the midpoint counts. ✓

So the bullish absorption proxy fires too — on the same bar.

Step 8: trap, and effort versus result

• Trapped buyers? The buy stack's lowest row is 4210.6, which is the close itself, not above it. The illustrative trap definition requires the stack to sit above the close, so no trap.
• Effort versus result: two ticks of movement on +76 delta gives an efficiency of 2 ÷ 76 ≈ 0.026 ticks per contract. Whether that is low depends on what is normal for this session; with a modest delta and a modest move, nothing here stands out.

The honest story

One bar, two patterns pointing in opposite directions: bearish exhaustion at the top and bullish absorption at the bottom. If you pick whichever one suits your existing opinion, you will always find "confirmation".

The honest reading puts the numbers into a story without forcing a direction:

Sellers were aggressive at 4210.4 — the busiest price in the bar — but price did not go lower. Buyers answered with a three-row stack from 4210.6 to 4210.8, but at 4211.0 nobody followed them. This was two-sided rotation inside one minute.

Anything more than that has to come from outside the bar: the reference level it sits on (Step 1), and what the next bars do. Does price hold above the stack and come back to retest 4211.0, or does it slide back to the 4210.4 selling and through it? The footprint has told you who was aggressive, where and how hard. It has not told you what happens next.

Key idea: Read every bar in the same order: context → totals → POC and delta-POC → imbalances → extremes → effort and result → story. Expect conflicting patterns. The story comes from the bar plus its location plus the bars that follow.

Common mistake: Starting from the imbalances instead of the context; picking the one "signal" that matches your bias; and forgetting that a symbolic teaching bar has different scale from real GC bars in different sessions.

Chapter summary

• Footprint patterns are vocabulary for describing what happened, not standalone triggers. On their own, they did not reliably predict direction after costs in our own testing.
• Exhaustion (thin volume at the extreme) describes the normal shape of most bars; it is worth noticing only with context — level, prior move, volume, follow-through — and suggests a pause, not a reversal.
• Finished extremes have zero on one side; unfinished extremes have both sides trading. Definitions differ (simple vs threshold), and "revisits" must be compared with a baseline.
• Absorption has three meanings: order-book refills (direct), footprint proxy (indirect), candle shape (weak). It shows presence, not outcome, and it can break.
• Effort vs result compares delta with movement; low efficiency must be judged against the same session's norm and read at meaningful locations.
• Trapped traders = strong aggression in a zone, then a close on the other side. Intent is invisible, so "trapped" is an assumption; it matters most at failed breakouts.
• Delta-POC marks the most one-sided row; its shift describes how aggression moves. It is a derived, descriptive metric with tie and noise problems.
• CVD divergence fails often: passive orders move price, absorption lasts, anchors matter when comparing across them, failing and working divergences look identical when they form, and classification errors accumulate.
• Judge any pattern with fire rate, precision and baseline, with outcomes defined in advance and costs included.
• Most mistakes come from reading detail without context. Read sequences, at a few important places.
• A single bar can trigger opposite patterns at once. The honest reading is a story, completed by the level and the next bars.

Checklist

• Before reading a bar, I write down which reference level it is next to.
• I compare the bar's volume with the average for its session before taking any pattern seriously.
• I have written, precise definitions for each pattern I use, and I know which definition a source uses before I trust its claims.
• I do not call a 1 × 2 extreme "unfinished", and I compare revisits of nearby prices with a baseline.
• I separate exhaustion (aggression stopping) from absorption (aggression soaked up).
• I measure effort versus result against the same session's typical efficiency.
• I treat "trapped traders" as an assumption about intent, not a fact.
• I check delta-POC ties and ignore one- or two-tick shifts in thin bars.
• I check the CVD anchor before calling a divergence, and I look at divergences that failed as well as those that worked.
• For any pattern, I know (or plan to measure) its fire rate, precision and baseline, with costs.
• I read bars in a fixed order and expect conflicting patterns.

Quiz

1. A footprint "exhaustion" rule fires on a large fraction of all one-minute bars. How is it best used, and why?
2. The top row of a bar reads 0 × 7 (bid × ask). Is this high finished or unfinished? What about a low row reading 9 × 4 under the simple "both sides > 0" rule?
3. Which data source can most directly confirm absorption — orders being repeatedly refilled at one price — and why can the footprint alone not confirm it?
4. Bar delta is +300, but the bar closes only one tick above its open, right at a major resistance level. What would Wyckoff's law call this, and what is one honest caution about interpreting it?
5. A pattern is followed by a 5-tick reversal 55% of the time. Random bars are followed by the same reversal 53% of the time. How much does the pattern add, and what else must you consider?

Quiz answers

1. As context at important levels, not as a standalone signal. Thin volume at a bar's extremes is the normal shape of most bars, so a rule that fires that often describes ordinary bars and carries little information on its own. It becomes interesting only with location, a prior extended move, meaningful volume and lack of follow-through.
2. The 0 × 7 high is finished (no aggressive selling at the high). The 9 × 4 low is unfinished under the simple rule, because both sides traded — although under a stricter threshold definition it could be labelled finished.
3. Order-book data (the DOM, a liquidity heatmap, or full market-by-order data). Absorption is about resting limit orders being consumed and refilled; the footprint shows only executed trades, not the resting orders themselves, so it can offer only an indirect proxy.
4. Effort without result (high effort, small result — sometimes called an aggression failure). Cautions: judge "low efficiency" against the same session's norm; the ratio is sensitive to close minus open; and it suggests the presence of passive sellers, not a guaranteed reversal.
5. Very little information — about two percentage points over the baseline, which may not even be distinguishable from chance with a modest sample. You must also consider how the outcome was defined, how many bars were tested, whether settings were tuned after seeing results, and whether anything is left after spread and commissions.

Education only. Not financial advice. Trading involves substantial risk of loss.