On this page
Start with the baseline
Every rug pull checklist says roughly the same things: look at the creator, the holders, the socials, the token's authorities. What a checklist rarely says is how much each check is worth. To find out, we took every pump.fun coin that traded between October 23 and 30, 2026 (153,941 coins with three or more holders in their first ten minutes) and compared how often coins with and without each sign went on to rug.
The baseline is 19.8%: one in five of these coins rugged. A warning sign is useful when coins that show it rug more often than that. A reassurance is useful when coins that show it rug less. Some of the best-known checks turn out to be neither.
The signs that came before rugs
Six signs stood out. Each one is visible within seconds of launch, before most buyers arrive.
- Coins showing the sign
- Coins without it
- The creator's earlier coin had already rugged12%31%
- The creator launched 20+ other coins in the past day12%30%
- No X account linked11%32%
- No website linked11%25%
- Ten wallets held over half the supply at launch19%82%
- Other wallets bought a quarter or more at launch19%45%
Source: Anceta. pump.fun coins that traded (three or more holders in ten minutes), October 23–30, 2026.
Show the numbersHide the numbers
| Sign | Coins with it | Rugged | Coins without it | Rugged |
|---|---|---|---|---|
| The creator's earlier coin had already rugged | 64,888 | 30.8% | 89,053 | 11.7% |
| The creator launched 20+ other coins in the past day | 47,236 | 29.7% | 68,858 | 12.0% |
| No X account linked | 65,866 | 31.7% | 88,075 | 10.9% |
| No website linked | 97,298 | 24.8% | 56,643 | 11.2% |
| Ten wallets held over half the supply at launch | 1,528 | 82.1% | 149,924 | 19.3% |
| Other wallets bought a quarter or more at launch | 6,062 | 45.3% | 145,390 | 18.9% |
The creator's record
This was the most telling sign a buyer can check before the first trade. Coins whose creator had an earlier coin that had already rugged went on to rug 30.8% of the time, against 11.7% for creators without one. They were not rare either: one launch in three came from such a creator.
Volume points the same way. A creator's first coin of the day rugged 12.0% of the time. Coins from wallets that had launched 20 or more other coins in the previous 24 hours rugged 29.7% of the time.
How to check it yourself: open the token in a Solana explorer, find the wallet that created it, and look at the other tokens that wallet created recently. Many launches in a short time, or earlier coins that collapsed within hours, is the pattern. Creators often move to fresh wallets, so a clean record is weaker evidence than a bad one.
The links, or the lack of them
A coin with no X account linked rugged 31.7% of the time; with one, 10.9%. No website: 24.8%; with one, 11.2%. The absence is the signal. The presence proves little: plenty of coins with both still rugged, and the next section shows that some of the most trusted details on those pages make no difference.
Who holds the supply
When the ten largest holders (leaving out the bonding curve and the creator) held more than half the supply at launch, 82.1% of those coins rugged. It is uncommon, 1,528 coins, about 1.0% of those that traded, but when it happens it nearly always ends the same way.
A large launch crowd is a milder version of the same thing. When other wallets bought a quarter or more of the supply in the first moments, 45.3% of coins rugged, against 18.9% for the rest. Wallets that get in that early are often part of the launch itself.
How to check it yourself: the holders tab of any explorer or trading terminal lists the largest wallets and their share. Leave out the bonding curve, which holds the unsold supply, and look at what the rest add up to.
Checks that didn't help
Some of the most repeated advice made no difference in this data, or pointed the wrong way.
- Coins with the detail
- Coins without it
- A Telegram link20%23%
- An X account that names the contract11%14%
Source: Anceta. pump.fun coins that traded (three or more holders in ten minutes), October 23–30, 2026.
Show the numbersHide the numbers
| Detail | Coins with it | Rugged | Coins without it | Rugged |
|---|---|---|---|---|
| Telegram link | 3,310 | 23.1% | 150,631 | 19.7% |
| X account names the contract (coins with an X account) | 1,573 | 14.0% | 85,125 | 10.7% |
A Telegram link was no comfort: coins with one rugged 23.1% of the time, slightly more than the 19.7% of coins without.
An X account that names the contract is often treated as proof that a coin is the real one. Among coins with an X account, the few whose account named the contract (1,573 coins) rugged 14.0% of the time, against 10.7% for the rest. Posting an address costs nothing.
Mint and freeze authority are on every checklist. In a random sample of 1,500 pump.fun coins that traded, every one had both revoked, and so did the 293 among them that rugged. The check passes for every coin, so it can't separate the ones that rug. What mint and freeze authority are, and what to check instead.
Why speed beats any checklist
The signs above are useful only if they are seen in time. Among the coins that rugged, 35.8% had already lost 80% of their peak value 30 seconds after launch, 59.2% after one minute and 76.8% after five.
- Rugs already 80% or more below the peak
Source: Anceta. pump.fun coins that traded (three or more holders in ten minutes), October 23–30, 2026.
Show the numbersHide the numbers
| Time after launch | Rugs already collapsed |
|---|---|
| 30 seconds | 35.8% |
| 1 minute | 59.2% |
| 2 minutes | 70.7% |
| 5 minutes | 76.8% |
| 10 minutes | 79.7% |
A careful manual check takes a few minutes, and by then most of these coins have already collapsed. The checks still matter, but they have to run before the coin is in front of you. That is what Anceta does with every launch, and RugUAV puts the result on the coin's row in Axiom, padre.gg and GMGN.
A short routine
If you check coins by hand, this order puts the strongest signs first.
- 01
Look at the creator
Its other recent tokens: how many, and whether any collapsed soon after launch. A creator with earlier rugs, or dozens of launches in a day, is the strongest warning in this data.
- 02
Look at who holds the supply
Leave out the bonding curve. If ten wallets hold most of the rest, or wallets that bought in the first moments hold a large share, treat that as a warning.
- 03
Look for the links
No X account and no website is a warning. Having them proves little, and an X post that names the contract proves nothing on its own.
- 04
Skip the checks every coin passes
On launchpad coins, mint and freeze authority are revoked across the board. They are worth a glance on other tokens, not a reason to trust these.
- 05
Mind the clock
Most rugs are over within the first minute. A check that comes later mostly confirms what has already happened.
How we tested
Coins. All 153,941 pump.fun coins that launched between October 23 and 30, 2026 (UTC) and had at least three holders within ten minutes, read from Solana by Anceta. The holder signs are measured in the coin's first seconds; the links are read from its metadata; the creator's record covers what Anceta saw from October 20.
Rugged. Anceta's rule (what counts as rugged), decided by October 31. Shares are rounded to one decimal.
Limits. These are patterns across many coins, not causes. A sign makes a rug more or less likely; it doesn't decide the fate of any one coin. Mint and freeze authority come from a random sample of 1,500 of these coins, read on October 31.
The full numbers behind these figures are in the October 2026 rug pull report.
Pattern analysis, not financial advice. Similar coins can behave differently.