Traders run dozens of checks before they buy a memecoin. We collected the 40 most common ones from 283 posts on X and 128 Reddit threads, then tested every one we could measure against what actually happened to the tokens on Robinhood Chain, from 10 August to 15 September 2026. Most checks are noise. A few are not. Here are the ones that hold up, and the one almost nobody runs.
What "rug" means here
A token can die two ways. It gets dumped: the wallets that bought first sell together, into the buyers who came after them, a coordinated exit. Or it simply fades: no buyers, no sellers, dead within 48 hours. Both lose you money. The checks below are about the first, because it is the one you can see coming.
How common it is depends on where the token is. Measured at the usual entry points, 45.6% of bonding-curve tokens got dumped that way, 36.1% of tokens between $150k and $1M, and 33.1% of tokens above $1M.
Check 1: who bought in the first 2 blocks
Snipers buy a token in its first 2 blocks, usually with bots. A lot of them is not a sign of hype. It is a sign that nobody else is coming.
Split tokens into five groups by the share of buyers who were snipers. In the group with the fewest snipers, 23.3% of tokens reached 5x. In the group with the most, 0.8% did. Heavy sniping marks tokens that go nowhere.
Check 2: bundled blocks
A bundled block is a block where 3 or more new wallets bought together. It measures coordinated attention, and attention cuts both ways: in the group of tokens with the most bundled blocks, 19.6% reached 5x and 36.5% got dumped. In the group with the fewest, 2.0% and 6.5%. Bundles bring the dump faster than the win. Never read them alone.
Check 3: where the buyers' money came from
Look at how much ETH the buyers' wallets were funded with over their lifetime. Tokens bought by wallets with a real funding history got dumped less, with no cost to the upside. Wallets funded with dust a few minutes ago are the pattern of a farm, not of traders.
Check 4: the same ticker, launched before
If the same ticker was launched earlier by someone else, the upside does not change but the downside grows. Copycat tickers are a small, consistent red flag.
Check 5: what the dev is doing now
Most traders check whether the dev has sold. The more useful question is whether the dev is selling right now. A dev who already sold before you arrive is priced in. A dev selling while you buy is the event that kills tokens.
Check 6: the family behind the early wallets
This is the check almost nobody runs, and the one that matters most. Rug operators do not use one wallet. They fund dozens, and they come back launch after launch.
When you link wallets that fund each other into families and follow each family across launches, the pattern is hard to miss. One 6-wallet family extracted 14 of its last 16 launches. One 40-wallet family extracted 4 of its last 5, at a median of -90%. When 5 of its wallets bought HUMANITY early, the token was down 98% within 24 hours.
A holder list shows you addresses. It does not tell you those addresses are family, or what that family did last time. That is the check to add.
What does not work
- Checking the top 10 holders' share after the fact. A dumped token ends with a hollowed-out top 10, so the number "predicts" the dump only because it was measured after it.
- Contract checks alone. Honeypot, sell tax, mint authority, locked liquidity: run them, they are table stakes. They are free and everywhere, and a token can pass all of them and still be dumped by its own early buyers.
- One number. No single check separates rugs from runners. The ones above work together, and each one should be read against the base rate of its market segment.
A 60-second routine
- Contract checks first: honeypot, tax, liquidity locked or burned.
- Snipers and bundled blocks: how much of the early buying was machines.
- Buyers' funding: real wallets or fresh dust.
- Dev: holding, selling now, and how many tokens they launched before.
- Families: are any early wallets linked to a family with a rug record?
FAQ
What is the most reliable sign of a memecoin rug?
Early wallets linked to a family that dumped its previous launches. One 40-wallet family on Robinhood Chain extracted 4 of its last 5 launches at a median of -90%.
Are snipers a bad sign?
A lot of them is. Among tokens with the highest share of first-2-block buyers, 0.8% reached 5x, against 23.3% for tokens with the fewest.
Is a contract check enough to avoid rugs?
No. Contract checks catch honeypots and taxes, but a token can pass them all and still be dumped by the wallets that bought it first.
How much of the market gets dumped?
On Robinhood Chain, measured at the usual entry points, 45.6% of bonding-curve tokens and about a third of tokens above $150k got dumped by their own early buyers.
Clément Hecquet
Co-founder and CEO of Otomato. Engineer turned product leader, building the portfolio-aware alerts layer for on-chain users.