Copycat Token Warning: 83% Were Dead Within 24 Hours

By the MemeAssist Research Desk · Published 2026-08-25 · Updated 2026-08-25 · 6 min read

In a frozen August 6–24, 2026 production cohort, 101 of 122 Solana tokens flagged as copycats were dead within 24 hours: 82.8%. The same-window rate among 3,308 resolved tokens without that flag was 14.8%. This is a strong association, not proof the label caused the outcome; 261 flagged checks had no definitive outcome and were excluded.

What counted as a copycat warning

This study is not about every token that reuses a ticker. The production warning looked for a fresh, non-established Solana token whose normalized name and symbol matched an older token with a materially deeper liquidity pool. Missing creation times did not count as evidence, and established tokens were exempted so the original could not be mistaken for its imitator. The flag was written at analysis time, before the next day's outcome was known.

We froze the window at August 6–24, 2026, the dates on which this signal appeared in the outcome tracker, and compared flagged and unflagged analyses from exactly the same period.

The result: 101 of 122 resolved warnings ended dead

Analysis-time cohort	Resolved n	Dead within 24h	Survived	Median return (survivors)
Copycat warning	122	101 (82.8%)	21	0.0%
No copycat warning	3,308	491 (14.8%)	2,817	−3.1%

The gap is 68.0 percentage points. It is large enough to make the warning useful, but it should not be read as an isolated causal effect. Copycats are often younger, smaller and thinner than the tokens they imitate. Those characteristics can also raise death risk, and this descriptive study does not match or randomize on them. For the broader size baseline, see our 24-hour survival study.

Why the survivor median is not the headline

The 21 survivors had a median 24-hour return of 0.0%; 11 finished above their analysis price and three lost at least half. But selecting only survivors removes the 101 worst outcomes. We therefore report death first and survivor returns separately. A vanished market is not a flat return and is never silently converted to 0%.

The missing-data limitation is unusually important here

Another 261 copycat-flagged analyses ended no_data: the tracker could not establish either a valid +24-hour candle or a confirmed vanished market under its pool-identity and quote-quality safeguards. They are excluded from both numerator and denominator. That is more no_data than resolved data in the flagged group, so 82.8% is the death rate among definitive outcomes, not among every warning shown.

For a deliberately pessimistic sensitivity check in the opposite direction, if every one of those 261 unknowns had survived, confirmed deaths would still represent 101 of 383 terminal flagged checks (26.4%). That is not an estimated death rate; it is simply a lower-bound scenario showing how much the unresolved rows can matter.

Methodology and limitations

Frozen cohort: production analyses from August 6 through August 24, 2026 (UTC). A resolved outcome is either done, meaning a validated price candle near analysis +24 hours, or dead, meaning the token's market had vanished from the tracked venue. For the flagged cohort, n = 122 resolved: 101 dead and 21 done. The same-window unflagged comparison has n = 3,308: 491 dead and 2,817 done. Death-rate denominators include both done and dead rows. Return statistics use done rows only. No_data rows (261 flagged; 1,394 unflagged) are excluded rather than guessed; pending rows are outside every figure. Each reported cohort exceeds the 30-resolved-observation minimum.

This is observational screening data, not a randomized experiment. Repeat analyses can represent the same token at different moments; the unit is an analysis-time observation, not a claim about unique projects. Signal logic also evolved during production, including a conservative established-token exemption. Results describe this frozen historical implementation and may not transport unchanged to a future market regime. No token, wallet or user-level records were inspected or published.

How to use the finding

A copycat warning is a reason to stop, not an instruction to short or a guarantee of a rug. Confirm that you have the intended mint address, compare the older market's age and liquidity, then run the ordinary checks for sellability, holder concentration and pool depth. The companion sell-route outcome study shows what happened when an actual exit quote could not be found.

Frequently asked questions

How risky are Solana copycat tokens?

Among 122 definitive 24-hour outcomes flagged by our production copycat check from August 6–24, 2026, 101 were dead: 82.8%. The same-window unflagged comparison was 14.8%, but this is an association and copycats may also be smaller or thinner.

What did this study call a copycat?

A fresh, non-established Solana token whose name and ticker matched an older token with a materially deeper pool. Ticker reuse by itself was not enough, and missing creation-time evidence did not count.

Were missing outcomes counted as rugs?

No. The 261 flagged no_data rows were excluded, not called dead. The 82.8% figure uses only 122 definitive outcomes: 101 confirmed dead markets and 21 validated 24-hour prices.

Does the study prove that copying a token causes it to fail?

No. It is descriptive production data. Copycats can differ in age, size, liquidity and other risks, and the study did not randomize or match those factors.