Token Overlap Engine
The Token Overlap Engine is Raphael Sentinel’s subsystem responsible for identifying relationships between wallets across multiple token ecosystems. While individual token analysis delivers insight into what is happening right now, the Token Overlap Engine uncovers what has happened before — and how historical wallet behavior influences current risk.
Its purpose is simple:
Detect whether the same wallets, clusters, or behavioral patterns reappear across different tokens — especially in high-risk or malicious contexts.
Overlap patterns are one of the strongest predictors of dev intent, swarm manipulation and coordinated scams. A token may appear clean in isolation, but overlap exposes hidden connections that completely change the risk profile.
1. Why Token Overlap Matters
Wallets rarely operate in isolation. Most malicious or coordinated actors behave repeatedly across many tokens, leaving behind:
repeated cluster structures
identical funding relationships
synchronized timing patterns
familiar exit signatures
recurring wallet ensembles (swarm sets)
This allows Sentinel to detect:
devs reusing wallet clusters
repeated farm launches
professionalized exit teams
fake “community” wallets recycled across launches
silent partner wallets funding multiple scams
A token with overlap to multiple high-risk archives becomes exponentially riskier.
2. Types of Overlap Detected
The engine identifies several categories of cross-token relationships:
1. Wallet Overlap
Simple intersection:
Wallet X appears in Token A and Token B.
Frequency and role are analyzed.
2. Behavioral Overlap
Even if wallets differ, patterns may be the same:
same timing sequences
same trade cadence
same snipe/dump intervals
If two sets behave identically → likely same operator.
3. Cluster Overlap
Entire clusters reappear across launches:
identical subgraph structure
similar funding origins
same dev-controlled pivots
This is the strongest form of overlap.
4. Funding Overlap
Shared sources of funding:
same root wallet
same funding hop chain
same multi-hop patterns
Funding overlap often exposes hidden dev wallets.
5. Exit Behavior Overlap
Dev clusters tend to “exit the same way”:
same time windows
same asymmetry patterns
similar dump curves
This helps detect serial farm developers.
3. Overlap Scoring Model
The Token Overlap Engine assigns each token an Overlap Risk Score, based on:
A. Volume of Overlap
How many wallets or clusters reappear?
B. Quality of Overlap
Are they high-risk?
rug actors
farm devs
bot swarms
wash traders
C. Overlap Role
Did overlapping wallets act as:
early buyers?
exit wallets?
LP providers?
dev wallets?
D. Behavioral Similarity Score
How closely do timing/momentum patterns match previous tokens?
E. Cross-Token Intensity
Number of overlapping tokens per wallet.
If a wallet appears across 10+ high-risk tokens → it becomes heavily weighted in risk calculations.
4. Integration With Other Systems
With WIM (Wallet Intelligence Model)
Overlap contributes to each wallet’s long-term reputation.
With Risk Engine
Overlap elevates:
Dev Cluster Score
Swarm Detection Score
Holder Integrity Risk
Historical Behavior Risk
With Cluster Detection
Overlap may reveal hidden clusters not visible in a single-token graph.
With Behavioral Pipeline (Pump.fun)
If a dev cluster has historically launched multiple farm tokens → risk skyrockets.
5. Cross-Token Dev Pattern Detection
One of the most powerful outcomes of the Token Overlap Engine is the ability to detect serial dev behavior.
Patterns include:
launching many bonding-curve tokens using similar wallet clusters
repeating entry/dump sequences
creating fake community wallets
duplicating behavior from previous launches
When Sentinel identifies a dev with a history of exploitative patterns, all new tokens associated with that cluster receive an immediate risk elevation.
6. Backward & Forward Influence
The Overlap Engine updates both:
Backward
Older tokens gain new risk intelligence when new overlaps appear.
Forward
New tokens inherit risk from historical overlaps.
This creates a dynamic, self-improving intelligence system.
7. Continuous Learning
Every new token processed strengthens overlap detection:
more wallet relationships
more cluster fingerprints
more behavioral templates
more cross-token correlation data
Sentinel’s understanding of malicious actors becomes deeper and more accurate over time.
Last updated