Skip to main contentPump Guard uses a transparent, metric-based scoring framework to surface risk and integrity signals. Each metric maps on-chain patterns to a normalized score, then aggregates into a composite risk score.
Scoring Principles
- Metric signs:
- [+] Positive metric: starts at 0%, improves toward 100%
- [-] Negative metric: starts at 100%, decreases toward 0%
- Normalization: all metrics mapped to 0–100 for comparability
- Weights: configurable per profile (retail, pro, platform)
- Output: per‑metric scores + aggregate score with badges
Example Metrics
[+] Total Native Balance (Top-50 Holders)
- Idea: If a token has migrated to pump AMM, healthy holders usually keep meaningful native SOL balances.
- Observation: Scam/farming tokens often have top wallets with very low native SOL (e.g., 40–100 SOL total across top-50), while healthier sets exceed 1,000+ SOL total.
- Score: min 0, max 100 (internally capped to prevent outliers). Also exposed as a raw value cap (e.g., 5,000) for dashboards.
[-] Fresh Wallets Concentration
- Idea: Recently created wallets among top holders increase manipulation risk.
- Rule of thumb: If the sum of ownership by 7 day old wallets in top‑50 exceeds ~5%, risk is elevated.
- Score: min 0, max 100; decreases as the fresh-wallet share grows.
[-] Zero-SOL Holders (Top-50)
- Idea: Bots often spend all SOL when entering, leaving zero native SOL as a red flag.
- Method: Count wallets with native SOL == 0 among top‑50; multiply by a factor (e.g., ×2) to derive penalty.
- Score: penalty toward 0; best case is zero zero-SOL wallets.
Interpretability
- Each token page shows:
- Metric cards with [+]/[-] badges
- Raw values and normalized scores
- Short rationale and thresholds
- Overall risk badge (Low / Medium / High)
Disclaimer: Pump Guard is not financial advice. It provides data-driven signals to improve decision-making.
