The Transaction Graph

A weighted, directed graph capturing economic relationships, used via snapshot-relative diffusion to allocate incentives and resist Sybil attacks.

The transaction graph is a weighted, directed graph that captures economic relationships between participants. Each completed interaction adds or updates an edge, and the protocol uses the resulting connectivity (via snapshot-relative diffusion) to allocate incentives and resist Sybil manipulation.

Graph Structure

Let the global transaction graph at epoch be:

  • : participants (buyers, producers, agents, domains, etc.)
  • : directed edges representing completed interactions
  • : nonnegative edge weights

In a marketplace setting:

  • buyer purchasing from producer adds edge
  • an optional reverse edge can represent fulfillment confirmation, dispute outcomes, or service-proof acknowledgements

Edge weights

Each completed transaction produces an edge weight:

Where:

  • amount is the economic value (price, fee base, etc.)
  • quality accounts for dispute outcomes, refunds, chargebacks, delivery SLAs, etc.
  • proof_factor is derived from attached service proofs and identity proofs

The protocol constrains weights to prevent pathological abuse (per-edge min/max, per-transaction caps, epoch caps, and/or decay).

Key Features

1. Dynamic Adjustments

The transaction graph dynamically adjusts based on participant interactions. As transactions occur, edge weights are updated, causing changes in connectivity and node influence. This creates a self-optimizing system where token distributions reflect the evolving state of the network.

2. Connectivity as a Measure of Value

The graph not only captures transaction volume but also connectivity:

  • Nodes with more connections to well-connected nodes are considered more influential.
  • This approach ensures that participants contributing to network growth through broad connectivity earn higher rewards, not just participants with high transaction volumes with a single counterparty.

3. Sybil Resistance

The graph’s structure inherently resists manipulation through Sybil attacks:

  • Sybil nodes (fake users) typically form isolated clusters without strong connections to real nodes.
  • The graph's weighting system prioritizes connections that enhance network-wide connectivity, making it difficult for isolated Sybil nodes to earn high rewards.

Next Steps

The transaction graph sets the foundation for diffusion and verification: