2.4 Node Reputation System

Reputation is the cumulative reflection of a node's long-term behavioral record. A new node begins its lifecycle with a neutral initial reputation score. When a node consistently contributes successfully to the network and its work outputs remain consistent across random verifications, its reputation score gradually increases in subsequent cycles. The longer a node maintains honest, reliable behavior, the faster and higher its reputation may grow, forming a positive feedback loop.

Nodes with low reputation scores face heightened scrutiny from the network—their produced work is subjected to a higher probability of being spot-checked, potentially reaching the point where every task is verified. Conversely, "established" nodes with high reputation scores enjoy a trust dividend, facing less frequent checks, thereby reducing their operational overhead. Consequently, the net actual reward a node ultimately receives depends not only on the volume of work completed but also on the proportion of its work that gets verified by others (which is influenced by reputation). This is a built-in insurance mechanism: high-reputation nodes gain higher net rewards due to lower verification costs; new or low-reputation nodes build trust by enduring higher verification intensity.

If a node is found producing malicious work or work that fails statistical tests during block production or random verification, it will lose all expected rewards for that period, and its reputation score will be reset to zero, returning it to the highest level of scrutiny. In each block production cycle, all nodes participating in consensus also need to collectively complete a resource-intensive but easily verifiable "consensus task." This task is similar to hash computation in Bitcoin's Proof of Work; finding a specific solution may require a vast number of attempts, but once found, any node can instantly verify its correctness. Upon each consensus completion, nodes update their voting weights used for governance and validation based on the valid computational proofs contributed during the most recent computational sprint, replacing the weight records from the old cycle. These updated weights are then used to assess whether the network has reached consensus on specific block candidates. The system does not allocate voting weight based solely on the amount of staked cryptocurrency but rather based on the effective computational workload each node executed during the sprint, essentially adopting and evolving the "one CPU, one vote" principle outlined in the Bitcoin whitepaper. The distribution of voting weight in the proposed system is comparable to the hashrate distribution observed in standard PoW systems, where voting weight is directly linked to expended computational effort. However, unlike traditional PoW systems—where all computational energy is "wasted" on solving hash puzzles—Origins Network's method channels this computing power towards tasks with real-world productive outputs, such as AI model training and inference. The allocated voting weights remain valid for a full governance cycle until a new computational sprint is completed, determining the next set of weights.