3 / 7 · 7 min
The order block: the price a move started from
Not every candle before an impulse is an order block. One filter makes it usable, and without that filter a chart produces a dozen blocks a day.
An order block is the price a strong move started from. There is a mechanism behind it and it is simple: if the move began here, somebody entered with size here, and some of those who wanted in did not make it. Those unfilled intentions have not gone anywhere — when price returns to the same price they are likely to show up again.
How to mark it
Take the last candle AGAINST the coming move, immediately before the move itself. For a rally that is the last red candle before the impulse up; for a decline, the last green one before the impulse down. The zone runs from its open to its extreme. Everything else in order-block marking is a variation on that rule, and arguing about the variations is pointless: what matters is that your rule stays the same every time.
Marking it out in numbers
Price spent three days between 1,840 and 1,862. The last red candle before the break upward: open 1,848, low 1,844. Then an impulse to 1,930 without a single pullback — a move of more than two ATR. The block: 1,844 … 1,848. Two days later price comes back to 1,847 — that is a touch of the block. Entry from there, stop below 1,838: under the block plus room for noise. Note that the block is narrow — four points — so the stop comes out short. That is exactly why such an entry is taken: not because it "works" but because being wrong is cheap.
The filter without which everything falls apart
Every move has a candle before it. Mark them all and you will collect a dozen a day, and half of them will coincide with a reversal purely by the number of attempts. The rejection rule: a block counts only if the move that began from it BROKE structure — made a new significant high or low. A move that merely sloshed about inside a range creates no block.
Two conditions, both testable by number. The first cuts out small jerks: a move smaller than two ordinary candle ranges is not an impulse but ordinary noise. The second cuts out moves inside a range: if no high was taken out, the balance of orders did not change and there is no need to explain the move by somebody's entry. The 2 ATR threshold is not sacred — 2.5 works too — but it must be written down in advance and be the same for every markup.
Assuming a block works forever
An order that gets touched is filled and disappears — the mechanics here are exactly those of an ordinary level. The first touch of a block means something, the second less, the third usually nothing at all: whatever price was going there for has already been taken. Practical rule: a block lives until its first touch, after which it becomes an ordinary price from the past. Mark spent blocks in another colour, or in a week your chart will carry twenty "active" zones.
Adding a story about who was there
"A large player was accumulating here" adds nothing to the markup but adds confidence. And in this business confidence converts into position size. All the chart shows is that a move began here; who began it and why is not written there. The markup says "opposing volume appears here more often than average" — and that is everything it can say honestly.
The order block has a test most constructions lack. Having marked a block, look at the density: are large limit orders sitting at that price right now. If they coincide, you are no longer guessing — there really is other people's money there. If they do not, the block remains a hypothesis and the position size should reflect that. Same caveat as always: the book is alive, and a wall can be pulled a second before price arrives.
Order-book densityTwenty blocks blind
In bar replay find twenty order blocks by your written rule, marking them BEFORE you see the next candles. For each, record: did the move break structure, was it larger than two ATR, and what happened on price's first return. Then split the blocks into two piles — those that passed both filters and those that did not — and compare how many worked. If there is no difference between the piles, your filters are not filtering, and it is better to learn that on a practice account.
Price came back to the block and went straight through without pausing. What does that mean?
That the opposing volume is no longer there — either it was taken earlier or it never existed and the move started for another reason. That is a normal outcome, not a failure of the method: a block shifts probabilities, it does not guarantee a bounce. What matters is that you placed the stop BEYOND the block in advance and lost exactly what you intended to. A markup is judged not by how often it is right but by what its mistake costs.
Why does a block without a structure break not count?
Because without a structure break the balance of orders did not change: price went back and forth inside the old boundaries, and there is no need to explain that by somebody's large entry. Such "blocks" appear by the dozen — one for every move — and half will coincide with a reversal purely by the number of attempts. The filter exists not for strictness but to bring the number of marks down to something you can actually look at: three or four a week instead of thirty.