How Token Health Scoring Works
The Health Score is our read on how solid a token looks, on a scale from 0 to 100. We build it from public on-chain data, holder counts, and price history, and recalculate it every day. A higher score means a token looks healthier on what we measure. It's our own estimate, not a guarantee and not financial advice.
- Weighted inputs
- 0–100 scale
- Editorial impact
Overview
We score every token on four pillars, Community, Liquidity, Project, and Interest, and blend them into one number from 0 to 100.
The score draws on public on-chain data, holder counts, and price history. We recalculate it once a day.
A holder count is a count of addresses, summed across every chain we can resolve the token on. Addresses, not people: one holder is counted once on each chain they hold on, so the figure is a ceiling on the number of people behind it, not a headcount.
What we score
- Community
- Many addresses holding the token, and no small group owning most of the supply, score well. Ten wallets holding nearly everything lowers the score, since those few can move the price or sell out at will. Holder growth is the third input, and it declines to answer across a change in which chains we count: a one-chain reading and an every-chain reading are not the same measurement, and the gap between them is not growth. For the week after such a change the input drops out and the other two are reweighted, rather than growth being scored as flat. New tokens get a small floor so a thin early crowd doesn't sink them on day one.
- Liquidity
- Trading pools hold the money that lets you buy or sell without moving the price. We check how much sits in those pools, whether daily volume fits that depth, and whether the token's valuation holds up against it. Deep pools with steady, proportionate volume score well. Thin pools, or volume that dwarfs them, score poorly.
- Project
- This pillar looks at the project behind the token, how long it has existed and how openly it operates. Older projects with a real website and active social channels score higher. A safety check runs here too. A token flagged as a honeypot, one you can buy but can't sell, scores zero on this pillar.
- Interest
- Rising trading volume and real price movement lift this pillar's score. We measure how large a price move is, not which way it goes: a sharp drop shows the same active trading as a sharp rise. A token that's gone quiet and flat scores low. Larger, more established tokens score a little higher too.
The formula & weights
We blend the four pillars into the overall score at fixed weights: Community 30%, Liquidity 30%, Project 25%, Interest 15%. Community and Liquidity carry the most weight. A wide holder base and real trading depth matter most for whether a token has staying power.
- Community
- 30% of the total score. Holder count 40%, Concentration 35%, Growth 25%. An input with nothing comparable to measure drops out and the remaining weights are shared out, so a missing reading is never scored as an average one.
- Liquidity
- 30% of the total score. Pool depth 50%, Volume vs depth 30%, Valuation 20%.
- Project
- 25% of the total score. Age 35%, Presence 45%, Safety 20%.
- Interest
- 15% of the total score. Volume trend 40%, Price movement 35%, Market size 25%.
How to read the score
The score tells you how solid a token looks today. It doesn't predict price. A token that scores high looks well-held, liquid, and established on what we measure. A low score is a reason to look closer before you trust the token.
You'll find the full breakdown, the overall score plus a meter for each pillar, on the token's own page. The tokens list and each game's related-tokens table show the score without the breakdown.
Caveats
We treat the score as evidence, not a verdict. It's an automated read of public data, and it has limits.
- Scores update daily: we recalculate every token from the latest holder counts, prices, and project details.
- Holder growth can decline to answer: a holder count sums addresses across every chain we can resolve, and when that set of chains changes there is a week with no comparable pair to read a change from. Growth drops out of the Community pillar for that week instead of reporting the change in our own coverage as growth in the token.
- Quiet tokens lose their score: a token with no real trading for two weeks stops getting one. There's nothing current left to measure.
- Thin data means no score: a token with too little data to judge fairly shows no score rather than a misleading one.
- It's our own estimate: built from public data, and we can be wrong. Weigh it alongside your own research.
