3 / 6 · 11 min
Order book walls: why a wall is not a support level
A large limit order is visible and measurable. But its size in dollars means nothing until you divide it by the coin's turnover — and it holds price only while it stands.
A wall is a threshold, not an impression
Every price in the book has size sitting on it. Sometimes one level holds so much that neighbouring rows are invisible next to it: not 300 coins but 30 000. That is a density, a wall. The word large means nothing on its own, so behind ours there is a number: we count an order from $100 000 standing no further than 3% from the current price. We listen to 6 venues with an order book and 300 coins by turnover. Knowing these thresholds matters not out of curiosity but so you understand what you do NOT see: an $80 000 order on a small coin is an event, yet it will not reach our map.
The number that matters: the wall divided by turnover
The size of a wall in dollars is the least useful thing you can say about it. Four million sounds impressive on any coin, but on one it is a minute of ordinary trading and on another it is half a day. Comparing walls across instruments by money is pointless for exactly the reason comparing open interest by money is pointless: the scales are different. The useful number comes from a division.
1440 is the number of minutes in a day. The result answers how long the market needs to eat this order if it trades at its usual pace and nothing else changes. It is not a forecast but a translation of a sum into something you can picture — into time.
Two walls at the very same moment
Measured on our density map. First: bitcoin, a sell order of $4 million right next to price. Bitcoin's daily turnover is $5.98 billion, so the wall is 0.07% of turnover — under a minute of trading. Second: ARPA, a sell order of $610 thousand half a percent above price. The coin's daily turnover is $1.03 million, so the wall is 59% of turnover — about 14 hours. In money the first is 6.6 times larger. In significance the second is almost 900 times larger. The bitcoin order will vanish within a minute and leave no trace on the chart; the ARPA order is a wall the coin will have to work through for half a day.
Three numbers that read a wall
| Number | What it answers | When the wall carries weight |
|---|---|---|
| Share of daily turnover | how long the market must eat it | 5% and above |
| Distance to price | will price even reach it today | inside the daily range |
| Age | has it survived a single approach | longer than a few minutes |
None of the three works alone, and that is the main thing to take from the table. A wall at 40% of turnover standing 2.9% away on a coin with a daily range of 1% is an order price will simply not reach today. A fresh wall half a percent from price is an intention that no approach of the market has tested yet.
Why this is not support
A support level on a chart is memory of the past: price already bounced here. It cannot be cancelled, because the past has already happened. A wall is an obligation of the present. It exists while the order stands and disappears in two ways: it gets filled, or it gets pulled. The second takes its owner one second and costs nothing. So calling a wall support swaps one thing for another: what you take for footing answers to a button press. A wall changes not direction but speed. It tells you where a move will slow down, and that is all it tells you.
Three outcomes when price arrives at a wall
- It is eaten gradually and does not go away — a genuine buyer of size stands behind it. Price slows and you gain time to decide. This is the only outcome in which a wall behaves the way people think it does.
- It disappears seconds before price arrives — it stood in order to be seen. Nothing happened except that you made a decision based on it.
- It is pushed straight through — often the strongest of the three moves: everyone who counted on it is left without footing at once.
Reading a sell wall as a bearish signal
A sell order means exactly one thing: somebody is willing to sell at that price. It does not follow that price will not go there — only that it will have to be pushed through. And if it is pushed through, the move up becomes STRONGER, not weaker: the large seller is gone, and above them the book is thinner because orders were piling up under an obvious level. The fastest impulses start exactly where a large order stood a second ago. A wall is a place where something gets decided, not a ready answer about which way.
Age: freshly placed versus standing for half an hour
A live measurement of the map at one moment: out of 60 walls, 31 were older than 3 minutes, 10 older than an hour, and the oldest one (on silver) had stood for 116 minutes. The difference between a fresh wall and a standing one is the difference between an intention and a deed: an order that lives into its third minute has already survived several approaches of the market and was not pulled. One caveat, without which the number lies: age counts from the moment we first saw an order at that price. Move it by a single tick and for us it is a new wall. So a young wall is sometimes a relocated old one, and erring low on age is safer here than erring high.
Six order books instead of one
The same order in money on a large venue and on a small one is two different events. Where turnover is smaller, less volume pushes the wall through, and price there cannot drift far from the rest of the market: arbitrageurs take the difference. Hence the practical reason we always show the exchange next to the level: without it you do not know where to place your own order. And the reverse case — a wall standing SIMULTANEOUSLY on several venues at one price — is rare and weighs far more than a single large one: it cannot be coordinated, so independent interests met at that price.
Putting your stop just behind the wall
The spot right behind a large order looks protected — you have to break through first. For exactly that reason everyone else puts their stops there too, and you get a cluster of forced orders in a predictable place. If the wall is pushed through, your stop fires at the worst possible point: where everyone who thought the same is closing at once. A stop belongs where your own rule puts it — beyond the edge of the level's zone — while the wall only explains why price is unlikely to get there for free.
What the book does not show
The book shows only limit orders that have been placed. An iceberg shows a small part of its size: the visible hundred gets bought and the next hundred appears, which from outside looks like a wall that never melts. Stop orders are not in the book at all: the exchange holds them and turns them into market orders the moment price touches. That is why price sometimes passes a level and accelerates sharply — a chain fired that was never on the map. Our map honestly shows what stands; it does not show what will happen.
The screener has a density column, and its header has a Standing > 3 min button. It keeps only the walls that have survived several approaches of the market. Turn it on and compare the map before and after: roughly half usually disappears. That gap is the difference between what stands and what merely flickered.
Order book wallsTwenty walls and three numbers
Take twenty walls in a row from the map and write down three numbers for each: share of daily turnover, distance to price, and age. After 15 minutes note what became of it — standing, eaten, or pulled. Sort the twenty cases into three piles and see which of the three numbers separated standing from pulled best. For most people it turns out to be age rather than size — and that will be your own measurement, not somebody else's claim.
A coin with $40 million daily turnover has a $2 million buy order sitting in the book. Is that a lot?
That is 5% of daily turnover, about 72 minutes of ordinary trading — a notable wall the market will take more than an hour to work through. The same order on bitcoin with $5.98 billion of turnover is half a minute, which on the scale of a day means it is not there at all. One sum, two different events: the meaning appears only after dividing by turnover.
Price came to a sell wall, ate it whole in two minutes and went higher. What was that?
An event stronger than the wall itself. The large seller who stood in the way is gone, and above them the book is thinner — orders had been piling up under the obvious level. On top of that, everyone who shorted off the wall is now closing with buys. That is why an eaten wall often turns out to be the start of an impulse rather than its end, and its removal deserves as much attention as its appearance.