Methodology: where every number comes from
Where SOLLMG's data comes from, what it covers, and how launches, graduations, survival rate, serial deployers and risk signals are counted, updated and limited.
Where the data comes from
The site has three kinds of data, each from a different source:
- Live monitoring (the new-token feed, hot list, crash tracker, smart money, daily report and token check): all from public Solana on-chain records. Our collector subscribes to the pump.fun program’s on-chain events and parses and counts every launch, every buy or sell on the bonding curve and every curve completion.
- Token library and rankings: the list of tokens comes from CoinGecko’s Solana ecosystem category. Prices and 30-day changes come from CoinGecko; pool liquidity and 24-hour trades come from DexScreener. Mint authority, freeze authority, on-chain supply and the largest holder accounts are read directly from Solana.
- pump.fun graduate archive: graduation records come from our own live collection. Post-graduation liquidity, trades and prices come from DexScreener.
We do not use any data that requires a login.
Coverage
Live monitoring covers only the pump.fun bonding-curve stage: launches, buys and sells, and curve completion (graduation). Other launchpads such as LetsBonk, and trading in the pool after a token graduates, are not part of the live figures. Post-graduation market data appears only in the token library and graduate archive, from the third-party sources listed above. New launchpads will be noted in the version history on this page when they are added.
How each metric is calculated
| Metric | How it is calculated |
|---|---|
| New launches | Successful pump.fun token creations in the period. Failed transactions are not counted |
| Graduations | Tokens whose curve completed in the period (all sellable supply sold). After completion, a migration transaction moves the token into a pool |
| Graduation rate | Graduations ÷ new launches in the same period. Both use the same period, but a token that graduates was not necessarily launched in that period, so this is only a rough comparison |
| 24-hour / 7-day survival rate | Tokens are grouped by launch time. When a group reaches 24 hours / 7 days, each token is checked for recent trades. A token with trades counts as alive |
| Serial deployers | Counted by the creator address in the launch transaction: how many new tokens each address launched in the past 24 hours |
| Hot list | Bonding-curve volume over the last 5 minutes / 1 hour (buys plus sells, in SOL). Tokens with fewer than 3 trades are excluded |
| Crash tracker | Tokens whose curve once held 8 SOL or more and whose current price is 90% or more below the curve’s peak price |
| Smart money profit | Realized profit and loss when an address sells during the curve stage, using the average-cost method, over a rolling 24 hours. Addresses with fewer than 5 closed positions are excluded |
The five risk signals for new tokens
| Signal | Trigger |
|---|---|
| Serial deployer | The creator launched 5 or more tokens in 24 hours |
| Large developer first buy | The developer bought 2 SOL or more at launch |
| Suspected bundled buys | 4 or more different addresses (excluding the developer) bought in the same block as the launch |
| Concentrated holdings | Once there are 10 or more trades, the top 10 addresses hold 50% or more of circulating supply (tokens already bought out of the curve and still held by addresses), estimated from the buys and sells we collected |
| Developer has sold | The developer sold half or more of the amount they bought |
A token with 2 or more signals, or with “Developer has sold”, is rated High risk. One signal is rated Caution. No signals is rated None detected. That only means nothing has fired yet. It does not mean the token is safe.
Token library check rules
The token library first separates tokens by asset type: meme tokens, project tokens, bridged or multichain assets, stablecoins, liquid staking tokens and tokenized assets. For the last four types, keeping a mint or freeze authority is usually by design, so it is shown as a note rather than a risk.
| Check | When it counts as a risk |
|---|---|
| Mint authority | A meme or project token has not renounced its mint authority |
| Freeze authority | A meme or project token has a freeze authority set |
| Concentrated holdings | The ten largest accounts of a meme or project token hold 60% or more of supply |
| Thin liquidity | Pool liquidity of a meme or project token is under 0.3% of its on-chain market cap (price × supply on Solana) |
| 30-day crash | The price fell 60% or more over the past 30 days |
| Low activity | Fewer than 20 trades in 24 hours. No conclusion is drawn on days when pool data couldn’t be refreshed |
Grading: a meme token that keeps a mint or freeze authority is rated High risk outright. Otherwise the fired checks are counted: 3 or more is High risk, 1 or 2 is Caution, and 0 is None detected. A meme token without either authority is also rated High risk once 3 other checks fire. For non-meme tokens that keep both a mint and a freeze authority, the two count as a single item (both are “retained authority”), so the number of checks shown as fired on the page can be one higher than the count used for grading. When the on-chain authorities couldn’t be read, the page says so instead of treating them as renounced. Rankings are sorted from the same data.
Graduate archive threshold
A token gets an archive page once it has been graduated for 24 hours and, at the daily check, had 100 or more trades in the past 24 hours and $10,000 or more in liquidity. Pages already published are kept even if the token later falls below the threshold; they show the current status and when the market data was observed.
Timing and updates
- All times are Beijing time (UTC+8). “Past 24 hours” means the 24 hours leading up to the time the report was produced.
- Live leaderboards refresh every 5 seconds. Daily reports are summarized by the hour, and each report has its own number in the format
SLG-YYYYMMDD-HH. - Right after the collector starts or restarts, 24-hour figures are labeled “collected for X hours”. Figures covering less than 24 hours cannot be compared directly with a full day.
- The weekly report covers Monday 00:00 to Sunday 24:00.
How verdict labels are set
The verdict in the right-hand column of the report table is a reading aid, not investment advice. When new launches change by more than 3% versus the previous 24 hours, they are labeled “Elevated” or “Down”. A graduation rate below 2% is labeled “Low”. A single address launching 500 or more tokens in 24 hours is labeled “Abnormal”.
Known limits
- Public nodes occasionally disconnect. Events missed during an outage are not estimated or backfilled, and the affected periods are marked on the page.
- A small number of tokens may launch through non-standard paths and may not be detected. We keep checking and record any rule changes here.
- We report statistics only. We do not judge whether any specific token is good or bad.
Methodology version history
- v1.0 (October 2026): first version. Covers launches, trades and graduations in the pump.fun bonding-curve stage; the five risk signals; and three live leaderboards for volume, crashes and profitable wallets.
If you find an error in the data, email [email protected] with the page, the time and the figure you saw.