Wallbot

Settings reference / Walls

Walls settings

How walls are detected and confirmed, and spoof detection. 16 settings on this tab.

Generated from the code. Defaults, ranges, choices, units and the live confirmation rule on this page are read from the settings schema each time the docs are generated, so they always match the bot.

Summary

SettingInternal nameDefault
Wall size vs. typical levelwall_k8 x
Minimum wall size: modewall_min_modedepth
Minimum wall sizewall_min_usd250000 USDT
Minimum wall size: share of 24h volumewall_min_vol_pct0.1 %
Minimum wall size: share of book depthwall_min_depth_pct10 %
Minimum wall size: hybrid floorwall_min_floor_usd50000 USDT
Wall core sharewall_core_pct90 %
Closest wall distancewall_min_dist_pct0.3 %
Farthest wall distancewall_max_dist_pct3 %
Zone widthwall_bucket_bps5 bps
Must persist forwall_persist_s45 s
Vanished belowwall_vanish_ratio0.35
Pulled if cancelled share ≥spoof_cancel_share0.7
Wall fade gracewall_fade_grace_s10 s
Wall rules: grace after startwall_start_grace_mult2 x wall_persist_s
Wall baseline bandwall_baseline_band_pct5 %

All settings

Wall size vs. typical level

wall_k

Default
8 x
Range
2 to 100 x
Type
Number
Unit
x
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: A price zone counts as a wall when it is this many times bigger than the median zone nearby.

What it does

How much bigger than a typical price zone a zone must be to count as a wall. The bot groups the book into zones (see wall_bucket_bps), takes the median zone size within 5% of price on that side, and a zone becomes a wall only if it is at least this many times the median (and also at least wall_min_usd).

Example

The median bid zone near price holds 40,000 USDT. At 8x a bid zone needs 320,000 USDT to be a wall. A 300,000 USDT zone does not qualify, even though it is above the 250,000 USDT floor.

When to change it

Raise it if the bot sees too many ordinary clumps as walls. Lower it on coins where real walls are rare and you want more candidates.

Higher, lower, on or off

Higher: fewer, more outstanding walls. Lower: more walls, including weaker ones. A tracked wall also 'fades' when it no longer meets this test for a short grace time.

Risk notes

Too low means trading in front of walls that are not really special. Too high can mean almost no trades.

Minimum wall size: mode

wall_min_mode I ACCEPT in live

Default
depth
Choices
fixed, volume, depth, hybrid, depth_floor
Type
Choice
Live confirmation
Yes: any change in live mode asks you to type I ACCEPT.

In the dashboard: How the absolute size floor of a wall is set (the 'wall size vs. typical level' test applies in every mode). Fixed: the same USDT amount on every coin (Minimum wall size). Volume: a share of the coin's 24h volume. Depth: a share of the coin's order book within 1% of price (both sides), re-measured every 10 s. Hybrid: the larger of a small fixed floor and the volume share. Depth + floor: the larger of the fixed floor (Minimum wall size: hybrid floor) and the depth share, so a wall must pass both. Until a coin's volume or depth is known, the fixed minimum is used.

What it does

How the absolute size floor of a wall is set. A zone must always pass the wall_k test (several times the typical zone nearby); this mode decides the second test. **fixed**: the same USDT amount on every coin (wall_min_usd). **volume**: a share of the coin's 24h trading volume (wall_min_vol_pct). **depth**: a share of the coin's order book within 1% of price, bids and asks together (wall_min_depth_pct), re-measured every 10 seconds. **hybrid**: the larger of a small fixed floor (wall_min_floor_usd) and the volume share. Until a coin's volume (after the first scan) or depth (after the first book) is known, wall_min_usd is used. Depth + floor (depth_floor): the larger of wall_min_floor_usd and the depth share, so a wall must pass both the fixed floor and the share of book depth.

Example

Depth mode at 10%: a coin with 2,600,000 USDT of orders within 1% of price needs a wall of at least 260,000 USDT; a small coin with 150,000 USDT within 1% needs only 15,000 USDT. With fixed 250,000 USDT the small coin would almost never show a wall, and the big one would accept walls that are small for its book.

When to change it

Fixed suits a list of similar, liquid coins. Depth or volume suit mixed lists (big coins and small movers), because the floor follows each coin's size. New installs start in depth mode; installs that existed before keep fixed until you change it.

Higher, lower, on or off

Depth: the floor follows the book right now, so it is high on deep coins and low on thin ones. Volume: it follows the day's trading, which changes slowly. Hybrid: like volume, but never below the fixed floor.

Risk notes

Moderate. Lower floors on thin coins mean smaller walls, which are easier to pull or eat through. In depth mode a sudden thin book also lowers the floor for a while (until the next 10 s measurement).

Minimum wall size

wall_min_usd

Default
250000 USDT
Range
1000 to 1e+10 USDT
Type
Number
Unit
USDT
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: Fixed mode: absolute floor, so thin coins don't produce tiny 'walls'. In the other modes it is the fallback while the coin's volume or depth is not known yet.

What it does

An absolute size floor for a wall, in USDT, used in the fixed wall_min_mode (and in the other modes while the coin's volume or depth is not known yet). A zone must pass both this and the wall_k test. It stops thin coins, where the median zone is tiny, from producing small 'walls'.

Example

On a thin coin the median zone is 5,000 USDT, so 8x would be only 40,000 USDT. With a 250,000 USDT floor, a zone still needs 250,000 USDT to count.

When to change it

Raise it to only trade really large walls. Lower it if you trade smaller coins and see no walls at all.

Higher, lower, on or off

Higher: only big walls count, fewer setups. Lower: smaller walls count on thin coins.

Risk notes

Small walls are easier to pull or eat through, so lowering this accepts weaker protection.

Minimum wall size: share of 24h volume

wall_min_vol_pct I ACCEPT in live used only with wall_min_mode volume / hybrid

Default
0.1 %
Range
0.001 to 10 %
Type
Number
Unit
%
Live confirmation
Yes: lowering it in live mode asks you to type I ACCEPT.

In the dashboard: Volume and hybrid modes: a wall must hold at least this share of the coin's 24h quote volume.

What it does

Volume and hybrid modes: the share of the coin's 24h quote volume (USDT) a wall must hold. The volume comes from the scanner (24h ticker of the coin's signal book) and is refreshed at every scan.

Example

At 0.1%, a coin trading 300,000,000 USDT a day needs walls of at least 300,000 USDT; one trading 20,000,000 needs 20,000.

When to change it

Raise it if small coins show too many weak walls; lower it if big coins show none.

Higher, lower, on or off

Higher: fewer, bigger walls on every coin. Lower: more walls.

Risk notes

Volume can be high for a day after a spike while the book stays thin, so the floor may be too high for a while (and the reverse).

Minimum wall size: share of book depth

wall_min_depth_pct I ACCEPT in live used only with wall_min_mode depth / depth_floor

Default
10 %
Range
1 to 100 %
Type
Number
Unit
%
Live confirmation
Yes: lowering it in live mode asks you to type I ACCEPT.

In the dashboard: Depth mode: a wall must hold at least this share of the coin's order book within 1% of price (bids and asks together).

What it does

Depth mode: the share of the coin's order book within 1% of price (bids and asks added together, in USDT) a wall must hold. Measured from the live book every 10 seconds.

Example

At 10% (the default), with 1,000,000 USDT of orders within 1% of price, a wall needs at least 100,000 USDT. On recorded days, 10% was the only share of 10/20/30/40% that left walls on a very deep coin (ZEC, ~19M within 1%: ~13 walls an hour; none at 20%) while small coins still had walls.

When to change it

Raise it to keep only walls that dominate the nearby book; lower it to see more walls. Above ~20% the deepest coins show no walls at all.

Higher, lower, on or off

Higher: fewer, more dominant walls. Lower: more walls, including ones that are a small part of the book.

Risk notes

Too low accepts walls that are ordinary for the coin; too high leaves almost no walls.

Minimum wall size: hybrid floor

wall_min_floor_usd I ACCEPT in live used only with wall_min_mode hybrid / depth_floor

Default
50000 USDT
Range
0 to 1e+10 USDT
Type
Number
Unit
USDT
Live confirmation
Yes: lowering it in live mode asks you to type I ACCEPT.

In the dashboard: Hybrid mode: the small fixed floor; the minimum is the larger of this and the volume share. Depth + floor mode: the fixed floor a wall must reach as well as the depth share.

What it does

Hybrid mode: a small fixed floor in USDT. The minimum is the larger of this and the volume share (wall_min_vol_pct).

Example

Floor 50,000 and 0.1% of volume: a coin trading 10,000,000 a day needs 50,000 (the floor); one trading 500,000,000 needs 500,000.

When to change it

Use it so very small coins never count tiny clumps as walls.

Higher, lower, on or off

Higher: thin coins need bigger walls. It has no effect on coins whose volume share is above the floor.

Risk notes

Low.

Wall core share

wall_core_pct

Default
90 %
Range
50 to 100 %
Type
Number
Unit
%
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: Wall hug only. The wall's edges are taken from the fewest largest price levels in the zone that together hold this share of its size, so small stray orders at the zone edges are ignored.

What it does

Sets where a wall's edges are. Inside a wall's zone, the bot takes the fewest largest price levels that together hold this share of the zone's size, and uses their highest and lowest prices as the wall's near and far edges. Small stray orders at the zone edges are ignored.

Example

A zone holds 1,000,000 USDT: 600,000 at one price, 320,000 at the next and 80,000 spread over three other prices. At 90% the first two levels (920,000) form the core, so the edges are those two prices only.

When to change it

Lower it if small orders at the zone edges keep pushing the entry or stop to odd places. Raise it toward 100 to use the whole zone.

Higher, lower, on or off

Higher: edges include more of the zone, so the wall looks wider. Lower: edges hug the biggest levels, so the wall looks narrower.

Risk notes

Used by the wall hug planner for entry and stop placement, and for the edges used by stacks and the take-profit in front of the counter wall. Changes move those prices slightly.

Closest wall distance

wall_min_dist_pct

Default
0.3 %
Range
0 to 10 %
Type
Number
Unit
%
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: Ignore walls closer than this to price (no room to place an entry).

What it does

Walls closer to price than this are not picked up as new walls, because there is no room to place an entry in front of them. The distance is only checked when a wall is first discovered; a tracked wall that price approaches stays tracked.

Example

Price 60,000 USDT. With 0.3%, a new bid wall whose top is above 59,820 is ignored. One that starts at 59,700 (0.5% away) is tracked.

When to change it

Lower it to catch walls that appear close to price. Raise it if entries end up too close to price to be useful.

Higher, lower, on or off

Higher: only walls farther away are tracked. Lower: walls close to price are tracked too.

Risk notes

Must stay below wall_max_dist_pct (the dashboard refuses otherwise). Also feeds the wall band used by min_atr_to_wall_band and prefer_near_4h_level.

Farthest wall distance

wall_max_dist_pct

Default
3 %
Range
0.1 to 20 %
Type
Number
Unit
%
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: Ignore walls farther than this from price.

What it does

Walls farther from price than this are not picked up as new walls. It also limits the take-profit: a counter wall farther than this from price is not used as a target. The order-book view on the dashboard shows a little beyond this distance.

Example

Price 60,000 USDT. With 3%, a bid wall whose top is below 58,200 is ignored, and an ask wall above 61,800 is not used for the take-profit.

When to change it

Raise it to consider walls farther away (they are touched less often). Lower it to focus on walls near price.

Higher, lower, on or off

Higher: more distant walls tracked and allowed as targets. Lower: only nearby walls.

Risk notes

Far walls are rarely reached, so entries rest for a long time. Must stay above wall_min_dist_pct.

Zone width

wall_bucket_bps

Default
5 bps
Range
1 to 100 bps
Type
Number
Unit
bps
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: Adjacent price levels are grouped into zones this wide, because walls are often spread across several ticks.

What it does

The width of the price zones the bot groups the order book into before looking for walls. Walls are often spread over several ticks, so neighbouring levels are added up into one zone. The width is rounded up to whole ticks.

Example

Price 60,000 USDT with a 0.1 USDT tick: 5 bps is 30 USDT, so all bids from 59,700 to 59,730 are added up as one zone.

When to change it

Widen it on coins where walls are spread over many ticks. Narrow it to get tighter wall edges.

Higher, lower, on or off

Higher: wider zones, bigger sums, fewer but broader walls. Lower: narrower zones, walls must sit on fewer levels.

Risk notes

Changing it changes every wall's size and edges. Zones are recomputed when the price moves more than 20%, which also clears the tracked walls.

Must persist for

wall_persist_s

Default
45 s
Range
1 to 3600 s
Type
Whole number
Unit
s
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: A wall has to stay put this long before the bot trusts it. This is the main anti-spoofing filter.

What it does

How long a wall must stay before the bot trusts it (marks it confirmed). Only confirmed walls are used for entries, stacks and take-profit targets. This is the main anti-spoofing filter.

Example

A 400,000 USDT bid wall appears at 10:00:00. With 45 s it is confirmed at 10:00:45 if it is still there. If it is pulled at 10:00:30, it is never traded.

When to change it

Raise it on coins with many short-lived fake walls. Lower it to react faster to new walls.

Higher, lower, on or off

Higher: fewer walls get confirmed, slower reaction, fewer spoofs traded. Lower: faster, but more spoofed walls get through.

Risk notes

Too low trades walls that are placed only to be pulled. Too high can miss walls that matter only briefly.

Vanished below

wall_vanish_ratio

Default
0.35
Range
0.05 to 0.95
Type
Number
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: A wall is considered gone when it shrinks below this fraction of its peak.

What it does

A tracked wall is considered gone when it shrinks below this fraction of its peak size. The bot then decides if it was pulled (mostly cancelled, see spoof_cancel_share) or absorbed (mostly eaten by trades).

Example

A wall peaked at 1,000,000 USDT. With 0.35 it is gone once it drops below 350,000 USDT. At 400,000 it is still tracked (if it still stands out from the book).

When to change it

Raise it to react earlier to a shrinking wall. Lower it to give walls more room to shrink before acting.

Higher, lower, on or off

Higher: walls are declared gone sooner, more cancels and exits. Lower: the bot waits for a bigger drop.

Risk notes

A pulled wall cancels a pending entry and can close an open trade at market, so this directly changes how often that happens.

Pulled if cancelled share ≥

spoof_cancel_share

Default
0.7
Range
0.3 to 1
Type
Number
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: When a wall shrinks, the share removed by cancellation (not by trades) above which it is treated as pulled / spoofed.

What it does

When a wall vanishes, the bot looks at how its size dropped since its peak: removed by trades at the wall's prices, or removed by cancellation. If the cancelled share is at or above this value, the wall is treated as pulled (a likely spoof); otherwise as absorbed.

Example

A wall dropped by 700,000 USDT: 560,000 cancelled and 140,000 traded. Cancelled share 80%, at or above 0.70, so it is pulled. If 300,000 were cancelled and 400,000 traded (43%), it is absorbed.

When to change it

Raise it if you want only clear cancellations to count as spoofs. Lower it to treat more drops as spoofs.

Higher, lower, on or off

Higher: fewer walls labelled pulled, more labelled absorbed. Lower: more labelled pulled.

Risk notes

Pending entries are cancelled either way. For open trades, pulled can mean a re-anchor or a market exit, while absorbed leaves the stop protecting the trade (with the absorbed handling). It also feeds the spoof statistics used by max_spoof_rate.

Wall fade grace

wall_fade_grace_s

Default
10 s
Range
0 to 600 s
Type
Number
Unit
s
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: A wall that shrinks below the vanish ratio slowly (no big cancel) counts as faded only after this long.

What it does

A tracked wall that no longer stands out from the book (it falls below the size test made by wall_k and wall_min_usd) is only dropped as "faded" after it has stayed too small for this long. A sudden drop below the vanish ratio is handled separately, as pulled or absorbed.

Example

Grace 10 s: a bid wall shrinks slowly until it no longer qualifies at 12:00:00. If it grows back by 12:00:08 it is kept. If it is still too small at 12:00:11, it is marked faded and any entry built on it is handled as a lost wall.

When to change it

Raise it if walls flicker in and out on busy coins and you lose good setups to brief dips. Lower it if you want slowly melting walls dropped faster.

Higher, lower, on or off

Higher: walls survive longer short dips, fewer false "faded" events, but a genuinely melting wall is trusted longer. Lower: faster reaction to fading walls, more churn of entries on noisy books. 0: a wall fades on the first check where it is too small.

Risk notes

Medium. A long grace can keep an entry resting in front of a wall that is slowly disappearing.

Wall rules: grace after start

wall_start_grace_mult

Default
2 x wall_persist_s
Range
0 to 20 x wall_persist_s
Type
Number
Unit
x wall_persist_s
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: After a start (or when a coin's walls are rebuilt from scratch) no wall can be confirmed for wall_persist_s. For this many times wall_persist_s, rules that act because NO confirmed wall is seen wait: breakeven when unprotected, exiting a fill whose wall is not tracked, exiting after a pull of a wall not yet confirmed since the start for lack of a backup wall, and auto-managing a new manual position. Positions keep their exchange stops meanwhile. 0 = off.

What it does

A start-up grace for the rules that react to NOT seeing a confirmed wall. After Wallbot starts (or a coin's wall tracker starts again from scratch, e.g. when the coin re-enters the watch list or its price moved 20%), a wall needs wall_persist_s of persistence before it counts as confirmed, so for that time every position looks unprotected. During the grace (this value x wall_persist_s from the coin's first book update) these rules wait: breakeven when unprotected (protection_by_wall_between / breakeven_when_unprotected), the exit of a fill whose wall is not tracked yet, the exit after a pulled or faded wall that was not yet confirmed since the start, for lack of a confirmed backup wall (the position keeps its stop instead; once a wall has confirmed since the start, any backup wall present then had the same time, so the normal rule applies), and auto-managing a new manual position (retried after the grace).

Example

wall_persist_s 45, value 2: for 90 s after a restart a resumed short in profit keeps its stop instead of moving it to breakeven only because no wall has been confirmed yet. After 90 s the normal check runs with real confirmed walls.

When to change it

Leave it on. 0 restores the old behaviour (the rules act immediately after a start).

Higher, lower, on or off

Only the first seconds after a start or a tracker reset. Exchange stops, trailing, take profit and real wall events (a confirmed wall being pulled) work as always.

Risk notes

During the grace a position relies on its exchange stop (as it does while Wallbot is stopped).

Wall baseline band

wall_baseline_band_pct

Default
5 %
Range
0.5 to 50 %
Type
Number
Unit
%
Live confirmation
No. Changes apply in every mode without the typed confirmation.

In the dashboard: Zones within this % of the mid form the 'typical level' baseline that walls are compared with.

What it does

Sets which part of the order book is used to work out the "typical level" size. Only price zones within this distance of the mid are summed, the median of those zones is the baseline, and a wall must be wall_k times that baseline. Zones outside the band are not looked at at all, so walls beyond it are never seen.

Example

Band 5% with BTC at 100,000: zones from 95,000 to 105,000 form the baseline. If the median zone holds 40,000 USDT and wall_k is 8, a wall needs at least 320,000 USDT (and at least wall_min_usd).

When to change it

Rarely. Widen it if you raise "Farthest wall distance" (wall_max_dist_pct) close to or past 5%. Narrow it to compare walls only with the liquidity close to price.

Higher, lower, on or off

Higher: the baseline includes thinner far-away levels (often a lower median, so more walls qualify) and farther walls can be seen; slightly more work per book update. Lower: the baseline reflects the dense levels near price (often a higher median, fewer walls), and walls outside the band are invisible.

Risk notes

Medium. Keep it wider than wall_max_dist_pct, or walls in the gap between the two can never be found.