BK AK-Cluster Bombs
A multi-profile volume-distribution and price-clustering framework for mapping where market activity has concentrated across sessions, swings, anchors, and structural legs.
Acknowledgment
All glory and gratitude to G-d.
The “AK” in BK AK-Cluster Bombs honors my mentor, A.K.—the man whose discipline, patience, judgment, and respect for clean execution continue to influence every serious indicator I build.
BK AK-Cluster Bombs was developed by Killa_B as a unified volume-distribution framework designed to organize multiple profile perspectives without forcing the trader to run several disconnected indicators.
What Cluster Bombs Is
BK AK-Cluster Bombs maps where trading volume has accumulated across several independently defined market structures.
The framework can calculate and compare:
Current RTH Point of Control and Value Area
Prior RTH POC, VAH, VAL, high, and low
Time-anchored or pivot-anchored profiles
RTH-open anchored POC
Recent swing POCs
Clusters above and below current price
K-means volume clusters
Low-volume areas between clusters
Completed pivot-to-pivot leg clusters
The current developing pivot leg
Native per-bar footprint POC
Session cumulative footprint delta
Gann-octave range references
Level overlap and confluence
Heuristic level-strength ratings
Profile histograms and Value Area fills
The indicator follows one connected analytical sequence:
Define the market segment → distribute volume across price → identify POC and Value Area → separate concentration zones → compare overlapping levels → control what remains visible
It does not predict that a level must hold, reject, attract price, or produce a profitable trade.
Volume-Profile Foundation
A volume profile distributes analyzed volume across price bins.
From that distribution, Cluster Bombs calculates:
POC: the price bin containing the greatest analyzed volume
VAH: the upper boundary of the selected Value Area
VAL: the lower boundary of the selected Value Area
Cluster POC: the highest-volume bin inside an individual cluster
Cluster boundaries: the upper and lower price limits assigned to that cluster
LVN: a comparatively low-volume area between stronger concentrations
These are calculated references based on the selected data, lookback, bin count, source mode, session definition, and anchor.
They are not direct observations of resting orders or guaranteed support and resistance.
RTH Session Profile
The RTH engine calculates a profile for the configured regular trading session.
The default settings use:
Timezone: America/New_York
Session: 09:30–16:00
The current session can display:
POC
VAH
VAL
Optional profile histogram
Optional Value Area fill
Optional session-start marker
The prior completed session can display:
Prior POC
Prior VAH
Prior VAL
Prior high
Prior low
Prior session-start marker
The session definition and timezone must match the instrument being analyzed. An incorrect session or timezone will produce a profile that does not correspond with the intended trading period.
Anchored Profile
The Anchored Profile begins from a user-selected structural point and continues through the current market.
Available anchor types include:
Exact time
Most recently confirmed pivot low
Most recently confirmed pivot high
Time-anchor behavior can be configured as:
Containing bar
Next bar open
Previous bar close
When Exact Start Inside Bar is enabled with a valid Lower-TF source, the script can begin the calculation from intrabar data rather than assigning the entire chart bar to the profile.
Pivot anchors require right-side confirmation bars. The anchor therefore becomes known only after the configured pivot-right length has elapsed.
The Anchored Profile can display:
Anchored POC
Anchored VAH
Anchored VAL
Histogram
Value Area fill
Anchor marker
RTH-Open Anchored POC
The RTH-Open POC measures the developing highest-volume price beginning at the configured RTH session opening.
It provides a separate reference from the complete-session POC.
Early in the session, this value may move rapidly because the analyzed sample is still small.
Swing POC
The Swing POC module profiles recent structural swing segments and retains the selected number of completed swings.
This allows the trader to compare where volume concentrated during different directional legs.
A swing POC can remain relevant after the swing ends, but its continued relevance must be judged from current price behavior rather than assumed from historical volume alone.
Split Clusters Above and Below Price
The Split Cluster engine divides the selected profile at the current market price.
It separately analyzes:
Volume concentrations above price
Volume concentrations below price
The scope can use:
Current RTH session
Full configured lookback
Each side can display:
Cluster POCs
Cluster boundaries
Cluster-zone boxes
LVNs between clusters
Volume-sensitive color intensity
Because the split is based on the current close, the membership of the above-price and below-price groups changes as price moves.
Above-price clusters are potential overhead volume references.
Below-price clusters are potential underlying volume references.
They should not be described as guaranteed resistance or support.
K-Means Volume Clustering
Cluster Bombs applies one-dimensional K-means clustering to populated profile price bins.
The algorithm:
Identifies the active profile range.
Initializes price-based cluster centers.
Assigns each populated bin to its nearest center.
Recalculates each center using volume-weighted price.
Repeats the process for the configured iterations.
Orders the resulting clusters from lowest to highest.
Finds the highest-volume price bin within each cluster.
K can be selected manually from two through five.
The Auto setting estimates a reasonable cluster count from significant profile peaks. It should be interpreted as a heuristic estimate—not a mathematically proven optimum.
Cluster scope can use:
RTH profile
Swing profile
Both
The module can display cluster POCs, boundaries, zones, LVNs, and a session-structure classification.
Pivot-to-Pivot Clusters
The Pivot-to-Pivot engine profiles volume across confirmed structural legs.
It can display:
A selected number of completed legs
The current developing leg
Cluster POCs
Cluster boundaries
Cluster zones
LVNs
Leg direction
Retracement position inside the leg
Completed legs depend on confirmed pivots.
The current leg can change until the next opposite pivot confirms.
The retracement percentage is measured from the completed or developing leg range and compared with nearby common Fibonacci references for context.
Volume-Source Modes
Cluster Bombs supports three source modes.
Footprint
When native footprint information is available, the script reads each bar’s buy and sell volume and calculates a buy-volume ratio.
That ratio is used to divide the bar’s profile contribution into positive and negative volume.
The bar’s total volume is still distributed across the profile bins according to the portion of its price range overlapping each bin.
The script separately provides:
Native per-bar footprint POC
RTH cumulative footprint delta
Footprint status
This is footprint-informed profiling, not a reconstruction of TradingView’s complete native price-row footprint history.
Lower TF
Lower-TF mode requests intrabar OHLCV information from the selected lower timeframe.
The chart timeframe must be greater than the selected lower timeframe, and the script must receive valid intrabar arrays.
This mode provides greater price-distribution granularity than chart-bar mode, but it remains dependent on available lower-timeframe history and the configured safety limit.
Chart
Chart mode uses the current chart’s OHLCV bars.
Each bar’s volume is allocated across the profile bins intersected by that bar’s range.
Positive and negative classification is based on candle direction.
This is an approximation. It is not bid-versus-ask order flow.
Native Footprint Overlay
When the Footprint source is active and native data is available, the script can display the current bar’s native footprint POC.
This line updates as each new footprint bar develops.
Convergence between the per-bar POC and a broader session or anchored POC means that different calculation scopes are referencing a similar price. It does not guarantee a reaction.
Native footprint access depends on plan and instrument support. The script contains an on-chart warning when the requested footprint data is unavailable.
Session-Structure Classification
The profile clusters can be summarized as:
Balance: one volume cluster contains the configured dominant share
Split: major clusters are separated by at least the configured ATR threshold
Normal: neither Balance nor Split conditions qualify
This is an internal classification based on cluster distribution.
It is not a standard exchange-defined Market Profile day type and does not independently determine how the session should be traded.
Low-Volume Areas
LVNs are identified where volume is comparatively low relative to nearby profile concentrations.
They can indicate areas where the analyzed market previously spent less activity.
Price may move through an LVN quickly, but it can also stop, reverse, or build new volume there.
An LVN is a historical distribution measurement—not proof of a future liquidity vacuum or stop hunt.
Confluence Detection
The confluence engine compares calculated levels within a configurable price tolerance.
Potential contributors include:
RTH POC and Value Area
Prior-session levels
Anchored levels
Swing POCs
Cluster POCs
Pivot-leg POCs
Split-cluster levels
Native footprint POC
The displayed confluence count shows how many tracked references are located near the same price.
It does not represent a probability of success.
Strength Rating
The one-to-five-star rating is a heuristic score based on factors such as:
Cluster volume share
Number of overlapping levels
Proximity to a Gann-octave reference when applicable
It is not a validated win rate, prediction, or probability.
Gann-Octave Reference
The Gann-octave function divides the active analyzed range into eighths and identifies the nearest division to a calculated level.
This provides a range-position reference such as:
0/8
1/8
2/8
4/8
6/8
8/8
It is a nearest-eighth comparison inside the active profile range. It is not a complete Square of Nine or traditional Gann-angle engine.
Level Manager
Because the script can generate many simultaneous profile and cluster levels, the Level Manager limits visual density.
It can:
Preserve the current RTH POC
Preserve current RTH VAH and VAL
Preserve prior-session POC and Value Area
Limit the number of additional displayed levels
Hide associated boundaries and zones when a level is suppressed
This manages chart visibility only. It does not change the underlying profile calculations.
How to Use BK AK-Cluster Bombs
1. Configure the correct session
Set the timezone and RTH session for the instrument.
For U.S. index futures, the default 09:30–16:00 New York session may be appropriate for cash-session analysis.
For other instruments, adjust the session rather than assuming the default is correct.
2. Begin with the core RTH levels
Start with:
Current RTH POC
Current VAH
Current VAL
Prior POC
Prior VAH
Prior VAL
Leave the additional cluster modules off until the primary profile is understood.
3. Select the volume source
Use:
Footprint when supported and footprint-informed buy/sell separation is needed
Lower TF when more granular intrabar distribution is required
Chart when simplicity and broader compatibility are preferred
Do not treat the three modes as mathematically identical.
4. Set the profile resolution
Adjust:
Lookback
Number of bins
Value Area percentage
Lower-TF source
Footprint ticks per row
More bins create finer resolution but can increase noise and computational load.
Fewer bins create broader, smoother structures.
5. Add the Anchored Profile
Anchor from a meaningful event such as:
Confirmed swing high
Confirmed swing low
Session event
Breakout origin
Major news candle
Manually selected time
Use the anchored POC and Value Area to compare longer-duration acceptance with current-session structure.
6. Add K-means clusters
Enable cluster POCs first.
Then add:
Boundaries
Zones
LVNs
Session classification
Avoid enabling every visual element simultaneously.
7. Compare independent profile scopes
The strongest analytical information comes from agreement between different scopes, for example:
RTH POC near Anchored POC
Prior POC near a Swing POC
Cluster POC near VAH or VAL
Pivot-leg POC near the native footprint POC
Multiple levels produced by the same profile should not be mistaken for independent confirmation.
8. Use Split Clusters for current-location context
Review the nearest concentration above price and the nearest concentration below price.
As price moves, these groups recalculate.
Treat them as current-location references, not permanently fixed support and resistance.
9. Review LVNs between major clusters
An LVN separating two strong clusters can help define the transition between established volume areas.
Observe whether price:
Moves quickly through it
Rejects it
Begins building new volume
Accepts above or below it
10. Apply the Level Manager
Lock the core session levels and limit additional levels to the number that can be read clearly.
A cleaner chart is generally more useful than displaying every calculated object.
11. Confirm with price behavior
Before acting around any profile level, evaluate:
Approach speed
Candle closes
Acceptance or rejection
Volume development
Delta context
Market structure
Defined invalidation
No profile level should be used as an automatic entry.
Recommended Starting Configuration
RTH Profile: On
RTH Histogram: Off
Prior POC/VA: On
Anchored Profile: Off initially
Swing POC: On
Split Clusters: On
K-Means Clustering: On
Cluster Scope: RTH
Auto-Estimate K: On
Cluster POCs: On
Cluster Boundaries: On
Cluster Zones: Off
LVNs: On
Pivot-to-Pivot Clusters: Off initially
Confluence Detection: On
Strength Rating: On
Level Manager: On
Maximum Additional Levels: 5
Label Style: Compact
What Is Original
Volume profiles, POC, Value Area, K-means, pivots, ATR, LVNs, and Gann eighth divisions are established analytical concepts.
The original BK contribution is the specific architecture connecting:
Current and prior RTH profiles
Time and pivot anchored profiles
RTH-open developing POC
Swing-segment POCs
Dynamic above/below-price clustering
RTH and swing K-means profiles
Completed and developing pivot-leg clusters
Footprint-informed volume separation
Native per-bar footprint POC
Lower-timeframe exact anchor handling
Session cumulative footprint delta
Cross-profile confluence
Heuristic strength grading
Gann-octave proximity
Dynamic label construction
Profile histograms
Level-ranking and clutter management
The value of Cluster Bombs is not one individual line.
It is the ability to compare where volume concentrated across several independent market segments inside one governed framework.
Realtime Behavior and Limitations
Current-session POC, VAH, and VAL change as new volume is added.
Anchored profiles change until their ending point is fixed.
Pivot anchors confirm only after the configured right-side bars.
The current pivot leg changes until a new pivot confirms.
Split clusters change when current price changes.
K-means centers can move when profile data changes.
Auto K is a heuristic estimate.
Lower-TF history can be limited.
Footprint data may be unavailable.
Footprint-informed profiles are not native row-by-row footprint reconstructions.
Chart mode uses candle-direction classification.
Confluence counts are not probabilities.
Strength stars are heuristic.
LVNs do not guarantee fast movement.
POCs do not guarantee attraction or rejection.
Drawing limits can require older or lower-ranked objects to be hidden.
Risk Disclosure
BK AK-Cluster Bombs is provided for analytical and educational purposes.
It does not provide financial advice, guarantee performance, identify institutional intent, or predict how price must react to a profile level.
Users remain responsible for their own analysis, entries, exits, position sizing, order execution, and account risk.
Map the volume. Separate the clusters. Compare the profiles. Confirm the reaction.
A multi-profile volume-distribution and price-clustering framework for mapping where market activity has concentrated across sessions, swings, anchors, and structural legs.
The “AK” in BK AK-Cluster Bombs honors my mentor, A.K.—the man whose discipline, patience, judgment, and respect for clean execution continue to influence every serious indicator I build.
BK AK-Cluster Bombs was developed by Killa_B as a unified volume-distribution framework designed to organize multiple profile perspectives without forcing the trader to run several disconnected indicators.
BK AK-Cluster Bombs maps where trading volume has accumulated across several independently defined market structures.
The framework can calculate and compare:
Current RTH Point of Control and Value Area
Time-anchored or pivot-anchored profiles
The indicator follows one connected analytical sequence:
Define the market segment → distribute volume across price → identify POC and Value Area → separate concentration zones → compare overlapping levels → control what remains visible
It does not predict that a level must hold, reject, attract price, or produce a profitable trade.
A volume profile distributes analyzed volume across price bins.
From that distribution, Cluster Bombs calculates:
POC: the price bin containing the greatest analyzed volume
VAH: the upper boundary of the selected Value Area
VAL: the lower boundary of the selected Value Area
Cluster POC: the highest-volume bin inside an individual cluster
Cluster boundaries: the upper and lower price limits assigned to that cluster
LVN: a comparatively low-volume area between stronger concentrations
These are calculated references based on the selected data, lookback, bin count, source mode, session definition, and anchor.
They are not direct observations of resting orders or guaranteed support and resistance.
The RTH engine calculates a profile for the configured regular trading session.
The prior completed session can display:
The session definition and timezone must match the instrument being analyzed. An incorrect session or timezone will produce a profile that does not correspond with the intended trading period.
The Anchored Profile begins from a user-selected structural point and continues through the current market.
Time-anchor behavior can be configured as:
When Exact Start Inside Bar is enabled with a valid Lower-TF source, the script can begin the calculation from intrabar data rather than assigning the entire chart bar to the profile.
Pivot anchors require right-side confirmation bars. The anchor therefore becomes known only after the configured pivot-right length has elapsed.
The RTH-Open POC measures the developing highest-volume price beginning at the configured RTH session opening.
It provides a separate reference from the complete-session POC.
Early in the session, this value may move rapidly because the analyzed sample is still small.
The Swing POC module profiles recent structural swing segments and retains the selected number of completed swings.
This allows the trader to compare where volume concentrated during different directional legs.
A swing POC can remain relevant after the swing ends, but its continued relevance must be judged from current price behavior rather than assumed from historical volume alone.
The Split Cluster engine divides the selected profile at the current market price.
Because the split is based on the current close, the membership of the above-price and below-price groups changes as price moves.
Above-price clusters are potential overhead volume references.
Below-price clusters are potential underlying volume references.
They should not be described as guaranteed resistance or support.
Cluster Bombs applies one-dimensional K-means clustering to populated profile price bins.
Initializes price-based cluster centers.
Assigns each populated bin to its nearest center.
Recalculates each center using volume-weighted price.
Repeats the process for the configured iterations.
Orders the resulting clusters from lowest to highest.
Finds the highest-volume price bin within each cluster.
K can be selected manually from two through five.
The Auto setting estimates a reasonable cluster count from significant profile peaks. It should be interpreted as a heuristic estimate—not a mathematically proven optimum.
The module can display cluster POCs, boundaries, zones, LVNs, and a session-structure classification.
The Pivot-to-Pivot engine profiles volume across confirmed structural legs.
Completed legs depend on confirmed pivots.
The current leg can change until the next opposite pivot confirms.
The retracement percentage is measured from the completed or developing leg range and compared with nearby common Fibonacci references for context.
Cluster Bombs supports three source modes.
When native footprint information is available, the script reads each bar’s buy and sell volume and calculates a buy-volume ratio.
That ratio is used to divide the bar’s profile contribution into positive and negative volume.
The bar’s total volume is still distributed across the profile bins according to the portion of its price range overlapping each bin.
This is footprint-informed profiling, not a reconstruction of TradingView’s complete native price-row footprint history.
Lower-TF mode requests intrabar OHLCV information from the selected lower timeframe.
The chart timeframe must be greater than the selected lower timeframe, and the script must receive valid intrabar arrays.
This mode provides greater price-distribution granularity than chart-bar mode, but it remains dependent on available lower-timeframe history and the configured safety limit.
Chart mode uses the current chart’s OHLCV bars.
Each bar’s volume is allocated across the profile bins intersected by that bar’s range.
Positive and negative classification is based on candle direction.
This is an approximation. It is not bid-versus-ask order flow.
When the Footprint source is active and native data is available, the script can display the current bar’s native footprint POC.
This line updates as each new footprint bar develops.
Convergence between the per-bar POC and a broader session or anchored POC means that different calculation scopes are referencing a similar price. It does not guarantee a reaction.
Native footprint access depends on plan and instrument support. The script contains an on-chart warning when the requested footprint data is unavailable.
The profile clusters can be summarized as:
Balance: one volume cluster contains the configured dominant share
Split: major clusters are separated by at least the configured ATR threshold
Normal: neither Balance nor Split conditions qualify
This is an internal classification based on cluster distribution.
It is not a standard exchange-defined Market Profile day type and does not independently determine how the session should be traded.
LVNs are identified where volume is comparatively low relative to nearby profile concentrations.
They can indicate areas where the analyzed market previously spent less activity.
Price may move through an LVN quickly, but it can also stop, reverse, or build new volume there.
An LVN is a historical distribution measurement—not proof of a future liquidity vacuum or stop hunt.
The confluence engine compares calculated levels within a configurable price tolerance.
The displayed confluence count shows how many tracked references are located near the same price.
It does not represent a probability of success.
The one-to-five-star rating is a heuristic score based on factors such as:
Proximity to a Gann-octave reference when applicable
It is not a validated win rate, prediction, or probability.
The Gann-octave function divides the active analyzed range into eighths and identifies the nearest division to a calculated level.
This provides a range-position reference such as:
It is a nearest-eighth comparison inside the active profile range. It is not a complete Square of Nine or traditional Gann-angle engine.
Because the script can generate many simultaneous profile and cluster levels, the Level Manager limits visual density.
Preserve prior-session POC and Value Area
Limit the number of additional displayed levels
Hide associated boundaries and zones when a level is suppressed
This manages chart visibility only. It does not change the underlying profile calculations.
Set the timezone and RTH session for the instrument.
For U.S. index futures, the default 09:30–16:00 New York session may be appropriate for cash-session analysis.
For other instruments, adjust the session rather than assuming the default is correct.
Leave the additional cluster modules off until the primary profile is understood.
Footprint when supported and footprint-informed buy/sell separation is needed
Lower TF when more granular intrabar distribution is required
Chart when simplicity and broader compatibility are preferred
Do not treat the three modes as mathematically identical.
More bins create finer resolution but can increase noise and computational load.
Fewer bins create broader, smoother structures.
Use the anchored POC and Value Area to compare longer-duration acceptance with current-session structure.
Avoid enabling every visual element simultaneously.
The strongest analytical information comes from agreement between different scopes, for example:
Pivot-leg POC near the native footprint POC
Multiple levels produced by the same profile should not be mistaken for independent confirmation.
8. Use Split Clusters for current-location context
Review the nearest concentration above price and the nearest concentration below price.
As price moves, these groups recalculate.
Treat them as current-location references, not permanently fixed support and resistance.
An LVN separating two strong clusters can help define the transition between established volume areas.
Lock the core session levels and limit additional levels to the number that can be read clearly.
A cleaner chart is generally more useful than displaying every calculated object.
Before acting around any profile level, evaluate:
No profile level should be used as an automatic entry.
Volume profiles, POC, Value Area, K-means, pivots, ATR, LVNs, and Gann eighth divisions are established analytical concepts.
The original BK contribution is the specific architecture connecting:
Completed and developing pivot-leg clusters
The value of Cluster Bombs is not one individual line.
It is the ability to compare where volume concentrated across several independent market segments inside one governed framework.
Current-session POC, VAH, and VAL change as new volume is added.
Anchored profiles change until their ending point is fixed.
Pivot anchors confirm only after the configured right-side bars.
The current pivot leg changes until a new pivot confirms.
Split clusters change when current price changes.
K-means centers can move when profile data changes.
Footprint-informed profiles are not native row-by-row footprint reconstructions.
Chart mode uses candle-direction classification.
Confluence counts are not probabilities.
POCs do not guarantee attraction or rejection.
Drawing limits can require older or lower-ranked objects to be hidden.
BK AK-Cluster Bombs is provided for analytical and educational purposes.
It does not provide financial advice, guarantee performance, identify institutional intent, or predict how price must react to a profile level.
Users remain responsible for their own analysis, entries, exits, position sizing, order execution, and account risk.
Map the volume. Separate the clusters. Compare the profiles. Confirm the reaction.
Open-source script
In true TradingView spirit, the creator of this script has made it open-source, so that traders can review and verify its functionality. Kudos to the author! While you can use it for free, remember that republishing the code is subject to our House Rules.
Disclaimer
The information and publications are not meant to be, and do not constitute, financial, investment, trading, or other types of advice or recommendations supplied or endorsed by TradingView. Read more in the Terms of Use.