Methodology
Every score on SMARTZ is explainable — here is exactly how it is built
Smart money score
A 0–100 composite. Realised profitability (35%), entry timing versus token launch (25%), consistency of outcomes across positions (20%), conviction measured by hold time and position sizing (10%), and trade quality versus market average (10%). Every component is percentile-ranked against all active wallets, so a score is always relative to the current population, not an absolute dollar figure.
Wallet categories
Smart Money is a score of 55 or higher. Whales are ranked by absolute position size and volume. Early Snipers consistently enter within minutes of launch. Conviction wallets hold significantly longer than the median. Exit Liquidity wallets consistently buy after smart money has begun selling. Suspicious wallets appear inside funding clusters with near-identical entry timing and sizing.
Token score
Weighted from the count and quality of smart money buyers, net smart money flow, buyer-to-seller ratio, holder distribution, liquidity depth, bonding curve progress and launch age. A token bought by five high-scoring wallets ranks above one with higher raw volume from unscored wallets.
Risk score
Raised by concentrated holder distribution, detected wallet clusters sharing a funding source, unusually fast sell pressure from early buyers, and thin liquidity relative to market cap. A high risk score is a warning, not a verdict.
Position and PnL accounting
Positions are tracked per wallet and token using weighted average cost basis. Selling realises PnL against that basis; the remaining balance is marked to the latest indexed price for unrealised PnL. Win rate counts closed positions with positive realised PnL. Hold time is measured from first entry to final exit.
Data sources and freshness
Trades, launches and graduations are indexed from Robinhood Chain (chain ID 4663) covering the Pons V2 launch factory, router and meme hook contracts. Ingestion is incremental by block, so each run only reads new blocks. The header shows how recently the last run completed, and marks the dataset as demo when live indexing is not yet enabled.
Limitations
Scores describe historical on-chain behaviour. They are not predictions, financial advice or a guarantee of future results. Wallets can be split across addresses, and a single operator may control several clusters. Always corroborate before acting.