What happens after a Telegram channel says a token is "trending"?
Short answer: Across 300 Solana tokens listed by a large public trending channel, buying 0 to 15 minutes after the post and selling an hour later averaged -23% after costs, and 54% of the tokens were down 90% or more within 24 hours.
Trending channels post a new token every few minutes, plus a stream of “up 1,700% since listing” posts. In ten days we counted 529 new listings and 848 brag posts. Brag posts only cover tokens that went up. We measured every listing instead.
The test
- Source: a large public Telegram trending channel (464,000 subscribers), read through its public web preview. 1,377 posts from 2026-10-11 to 2026-10-21.
- Sample: 300 randomly chosen first listings that were at least 24 hours old.
- Entry: the open of the first 15-minute candle after the post, so 0 to 15 minutes late, like a human reading Telegram.
- Costs: 3% round trip (DEX fee plus slippage) and network fees on a 100 USD position. No price-impact model, so real results on thin pools are worse.
Results
| Rule | Win rate | Median | Average | 95% range of the average |
|---|---|---|---|---|
| Hold 1 hour | 23% | -25% | -23% | -31% to -14% |
| Hold 24 hours | 11% | -94% | -42% | -68% to +1% |
| Take profit +30%, stop -20% | 24% | -23% | -10% | -12% to -7% |
| Take profit +100%, stop -50%, 72 hours | 20% | -53% | -18% | -26% to -11% |
- 54% of the tokens were at -90% or worse 24 hours after the post.
- 69% touched -90% within 72 hours.
- The median best possible exit within 24 hours was +46%, but that peak usually came before or right at the post. Readers arrive after the move.
2023 looked the same
On the same channel’s Ethereum listings from June 2023 to April 2024 (11,806 unique pairs), on-chain reserves today show 68% of the pairs with under 1,000 USD of liquidity and only 1.6% above 50,000 USD. In a random sample of 129 tokens priced from on-chain swaps, a 24-hour hold had a median of -86%. A few lottery winners pulled the average up; none of them could have been sold at size.
Honeypots were rare, collapses were common
In the 2023 sample we found no token where people could buy but nobody could sell. The losses come from rugs, dumps by early holders, and simply buying after the pump. Honeypot checkers do not protect against any of those.
Verdict
FAIL. Every rule we tried lost money on average after costs. Our read: the profit in this market goes to the people who are earlier than the channel: deployers, snipers and insiders. Not financial advice.