Deterrence under partial identity linkage

Determine how partial linkage among identities changes the cluster-level deterrent in NFT-based reward mechanisms, including whether the deterrence term retains the form T L(a) F and how much deterrent value survives imperfect attribution.

Background

The model assumes the strongest possible linkage mechanism: if any identity in a cluster is detected during the vesting window, the entire cluster is forfeited and the farmer pays a penalty. This produces the deterrence term T L(a) F. The authors identify partial linkage—where only some identities can be connected through devices, funding, or behavioural signals—as a practically important but unresolved extension.

Under partial linkage, detection of one identity may not expose the entire operation, and the deterrent may no longer have the multiplicative form derived for complete cluster detection. Characterizing the resulting deterrence is important because linkage strength is itself a design variable in NFT-based reward programmes.

References

We conjecture, but do not derive, that partial linkage would generally weaken the cluster-level deterrent and move the model towards the account-level benchmark. Depending on how linkage is structured, the deterrent need not retain the form $T\,L(a)\,F$ at all; establishing what it becomes would show how much deterrent value survives imperfect attribution, and matters directly for practice, since linkage strength is itself a design variable.

— NFT-Based Reward Mechanisms: Sybil Farming, Vesting, and Stochastic Verification  (2609.13064 - Javarone et al., 11 Sep 2026) in Section 6, “Limitations and future work,” subsection “Cluster detection”