What Separates a 10x Meme Coin from a Rug Pull: Patterns from On-Chain Data
--
I run MemeScout, a Solana meme coin scanner that watches new tokens the moment they hit the chain, scores them, and surfaces the ones with a real shot at running. After watching tens of thousands of launches go by, one pattern keeps showing up: a 10x meme coin and a rug pull look almost identical in their first hour — except for a handful of on-chain signals that quietly decide everything.
This post is about those signals. None of them are magic, and none of them guarantee a win. But every time we’ve gone back through the data, the same set of variables keeps separating the runners from the rugs.
The setup: why “vibes” lose money on Solana
Most people pick meme coins the same way: a screenshot in a group chat, a green candle on DexScreener, a “presale” link in a Telegram. By the time those signals fire, the entry market cap is already 3–10x what it was at launch. Worse, half of what makes the chart green is bot wash trading — wallets buying from themselves to manufacture momentum.
The fix isn’t a secret indicator. It’s reading the on-chain data Solana already publishes: authorities, holder distribution, liquidity behavior, and how quickly real (non-bot) buys are accelerating. The hard part is doing that fast enough to matter, and at scale.
Signal 1: Mint and freeze authority
These are the kill switches built into every SPL token. If a token’s mint authority is not revoked, the dev can print unlimited new supply at any time — diluting holders to zero. If freeze authority is not revoked, the dev can freeze your wallet so you literally cannot sell.
Both flags are visible on chain at the moment of launch via Helius DAS. You don’t need an indexer to wait three minutes to tell you. We pull them at first detection, and tokens with either flag still active take a stackable penalty on their radar score — up to −50 points combined with metadata mutability and top-wallet concentration.
If you take nothing else from this post: never buy a token where mint or freeze authority is still live. The fact that “the dev says they’ll revoke later” is not a feature.
Signal 2: Top-3 wallet concentration
Look at the top 3 non-LP holders. If they own 25%+ of supply, that’s a coordinated-dump risk — those three wallets can liquidate the chart in one block. Our radar penalizes tokens with concentrated top holders, because the data is unambiguous: clean distributions outperform concentrated ones.
This isn’t about whales being evil. It’s about the fact that a coordinated wallet cluster can exit before you can refresh your screen, and there is no recourse.
Signal 3: Sell pressure at discovery
A signal we added recently: if a token’s sells exceed buys by 2x in the discovery window, we hard-block it. The reasoning is mechanical — if the order flow is already net-negative when we first see it, you are buying into a wall of exits, not momentum.
The same logic applies to volume composition. We track first-detection volume, buys, and sells for every radar hit. The 5x runners almost always have a buy/sell ratio at or above 2.5 in their first 5 minutes. Heavy early selling is one of the cleanest leading indicators of a rug or a flatline.
Signal 4: The radar score floor (and why we keep raising it)
When I first launched the radar, the minimum radar score was 30. I raised it to 36 after a backtest of one day’s worth of calls: tokens scoring 30–35 accounted for 14 of 18 flatlines. Tokens scoring ≥36 hit the 1.25x threshold 78.6% of the time, versus 59.3% for the lower band. The score was the strongest single separator we could find.
A month later we did a bigger analysis and raised the floor again, this time to 42, while bumping the minimum market cap from $25K to $50K. That single change cut roughly 60% of low-quality calls — the “garbage” tier that was technically passing but never actually moved.
The takeaway for traders: a high-quality filter is more valuable than a high-volume one. You’re better off seeing 8 strong calls a day than 80 mediocre ones. The math on opportunity cost is brutal once you account for slippage, fees, and the emotional drain of bag-holding flatlines.
Signal 5: Freshness is overrated
This one surprised us. Conventional wisdom says newer = better, because you want to be early. The data says the opposite.
Tokens detected within 30 minutes of launch hit 5x at roughly 1.1%. Tokens that survived past the 2-hour mark hit 5x at 13.6% — more than 10x the rate. Why? Because the first hour is dominated by bot churn, snipers, and dev wallets cycling supply. Tokens that are still alive and well at hour 3 have already passed a brutal natural filter.
So our radar now penalizes freshness instead of rewarding it. Under 30m: −20 score. 31–60m: −15. 61–120m: −10. 121–180m: −5. After 180m, no penalty. You can still get in early, but you should know what “early” actually costs.
What the data still won’t tell you
On-chain analysis is necessary, not sufficient. Even with all of the above, only about 32% of clean radar calls hit 2x, and roughly 6% hit 5x. Meme coin trading is still a positive-expectation game only if your sizing, exits, and discipline are right.
What on-chain data does buy you is a much higher floor: fewer outright rugs, fewer flatlines you should never have touched, and more of your bankroll preserved for the calls that actually run. That’s the entire game.
How to use this
If you want to actually filter Solana meme coins this way, you have three options. You can pull Helius DAS, RugCheck, and DexScreener yourself, score everything in real time, and rebuild the bot-detection stack. You can find a tool that does it. Or you can let me do it for you.
That’s what MemeScout is. Real-time radar, on-chain rug checks at detection time, bot-filtered Smart Money leaderboard, and the same scoring rules I just walked through — applied to every new Solana meme coin the moment it appears.
If a daily summary is enough, the Daily Digest emails one curated batch per day, free. Sign up at memescout.io/digest.
If you want the live radar, that’s at memescout.io.
Either way, please don’t ape into mint-authority-live tokens. We’re all trying to make it.
— Nate
