Methodology

One number. Five signals. No black box.

Every reading Lockstep posts — on the ticker or from /audit — comes from the same scoring function. Higher score means riskier by this composite's own definition of risk: concentrated, illiquid, and coordinated-looking. It is not a claim about future price.

25% weight — concentration

Top-10 holder share

What fraction of total supply the ten largest holders control. Risk scales linearly up to 60% concentration, where it caps at maximum risk for this signal.

A

score 0–14

25% weight — liquidity depth

Liquidity vs. market cap

Liquidity as a share of market cap. Below roughly 10% liquidity-to-mcap, risk rises fast — thin liquidity means a modest sell can move price a lot.

B

score 15–29

20% weight — bundler-owned

Coordinated-buy wallets

Share of supply held by wallets flagged as part of a bundled, same-block coordinated buy — a classic setup for orchestrated pumps and dumps.

C

score 30–49

15% weight — sniper-owned

First-block buyers

Share held by wallets that bought within the first blocks of launch — often bots positioned to exit into early retail volume.

D

score 50–69

15% weight — dev / creator holdings

What the creator still holds

Share of supply still sitting in the token's own creator wallet — the clearest single-wallet dump risk.

F

score 70–100

Missing data is excluded, never scored as safe

Not every chain or data source has all five fields. When a signal is missing, Lockstep drops it from the composite entirely and renormalizes the remaining weights over what's actually available — a missing bundler flag is never quietly treated as "0% bundler risk." Every reading that's missing a signal says so explicitly.

Honest limits

What this score is — and isn't.

It's a heuristic triage score over on-chain concentration, liquidity, and bundler/sniper signals only — hand-picked thresholds based on tokens this project has audited, not a statistically fit model.

It can't see off-chain rug vectors — upgradeable contracts, hidden mint authority, multisig/admin key risk, wash-traded volume, or a social-engineering exit scam. None of those show up in these five signals.

It's not a price prediction. A clean A-grade token can still go to zero on demand alone; a risky F-grade token can still pump. This measures manipulation/concentration risk, not where price goes next.

See it run on a real token.

Run /audit on Discord or Telegram, or open the Mini App to see live grades on a watchlist.