Key Learnings
- 01In this cohort, the observed 24-hour death rate was 2.0% for the 30–44% top-10 band (n = 559), 7.4% for 45–64% (n = 486), and 70.6% for 65%+ (n = 513).
- 02By largest-holder band, observed 24-hour death rates were 4.6% at 8–14% (n = 546), 2.1% at 15–29% (n = 512), and 60.4% at 30%+ (n = 636).
- 03Among resolved survivors in the displayed bands, median 24-hour returns ranged from −1.9% to +0.1%; survivor-only medians do not describe tokens classified as dead.
- 04No confidence intervals or significance tests were calculated, so differences remain subject to sampling uncertainty and should not be treated as validated thresholds.
What this cohort can—and cannot—show
This is a descriptive study of 3,779 resolved 24-hour outcomes for Solana tokens submitted to or discovered by MemeAssist from July 22 through August 24, 2026. It is a non-random convenience cohort, not a representative sample of all Solana tokens. Holder-concentration flags were recorded when each token was analyzed and joined to its subsequent 24-hour status.
The results describe associations within this exact cohort and observation window. They do not show that concentration caused an outcome, establish a universal boundary, or prescribe whether to trade. Token age, liquidity, creator behavior, market regime, and other measured or unmeasured factors may differ across bands. The service may also have analyzed the same token more than once; this extract does not support identifying or adjusting for repeated analyses.
Observed outcomes by top-10 concentration band
Tokens were grouped by the share of supply held by the top 10 non-pool wallets at analysis time. The displayed denominators contain resolved outcomes only:
| Top-10 holders' share | Tokens | Dead within 24h | Median 24h return (survivors) |
|---|---|---|---|
| 30–44% | 559 | 2.0% | −1.9% |
| 45–64% | 486 | 7.4% | +0.1% |
| 65%+ | 513 | 70.6% | −0.2% |
The observed rate was higher in each successive displayed band and differed substantially between the 45–64% and 65%+ groups. This comparison is descriptive. The bands were pre-existing signal categories rather than an estimated dose-response curve, and no adjustment for confounders was performed.
Observed outcomes by largest-holder band
A separate grouping used the largest individual holder's share at analysis time:
| Largest single holder | Tokens | Dead within 24h | Median 24h return (survivors) |
|---|---|---|---|
| 8–14% of supply | 546 | 4.6% | −1.0% |
| 15–29% of supply | 512 | 2.1% | −0.6% |
| 30%+ of supply | 636 | 60.4% | −1.1% |
The pattern was not monotonic across the two lower displayed bands: the observed rate was 4.6% at 8–14% and 2.1% at 15–29%. The 30%+ group had a higher observed rate in this cohort. These figures do not identify why the groups differed and should not be converted into a hard cutoff.
"In this non-random cohort, observed 24-hour death rates differed sharply across concentration bands; the data establish association, not causation or a universal cutoff."
How to read the survivor-return column
Median returns among survivors ranged from −1.9% to +0.1% across the displayed concentration bands. These conditional summaries exclude tokens classified as dead, so they cannot be interpreted as total cohort performance or evidence that concentration has no relationship with returns. No confidence intervals or statistical significance tests were calculated; apparent differences or similarities may reflect sample uncertainty.
Inclusion, outcomes, and missing data
The 3,779-record cohort included submitted or discovered tokens analyzed during July 22–August 24, 2026 for which analysis-time risk signals were captured and a 24-hour outcome resolved. Each band analysis then included only records carrying that specific concentration category; resolved records outside the displayed categories were excluded from those rows. “Dead” denotes a market that stopped quoting entirely within 24 hours. The rate in each row is dead / (done + dead), where done is the resolved non-dead outcome. Median return was calculated only among done survivors.
Records marked no_data, pending, or otherwise unresolved were excluded from both numerator and denominator. The aggregate extract retained for this article did not retain counts for those excluded statuses, so we cannot report how many records were unavailable overall or by band. This missingness may be informative: for example, outcome availability may relate to token age, liquidity, market disappearance, or data-provider coverage. Excluding unavailable outcomes can therefore bias the reported rates in an unknown direction.
The published rows are signal-specific and must not be added together: a token can appear in both a top-10 row and a largest-holder row. No public row-level export accompanies this analysis, and the underlying holder or market sources were not manually verified token by token. The findings are consequently not independently reproducible from this page.
Limits on interpretation and generalizability
The cohort reflects what users submitted and what MemeAssist discovered during one short market period. Selection into analysis, changing market conditions, and the confounders above limit generalizability to other tokens, platforms, or dates. There was no causal design, multivariable adjustment, confidence interval, significance test, sensitivity analysis, or correction for repeated analyses. The band counts are substantial enough to report descriptively, but every percentage remains an estimate from a finite, selected sample.
Holder concentration can be considered alongside liquidity, token age, creator history, wallet clusters, and current market conditions; these data do not support using any one band as a standalone decision rule. Our guide to reading top-holder concentration explains the metric without turning these cohort boundaries into universal thresholds.
Frequently asked questions
Did this study establish a universal holder-concentration cutoff?
No. It reports descriptive associations in a non-random cohort of tokens submitted to or discovered by MemeAssist from July 22 through August 24, 2026. The displayed bands were existing signal categories, not validated trading thresholds, and the analysis did not establish causation.
What were the observed top-10 concentration outcomes?
Among resolved outcomes in this cohort, 24-hour death rates were 2.0% for the 30–44% band (n = 559), 7.4% for 45–64% (n = 486), and 70.6% for 65%+ (n = 513). No confidence intervals or significance tests were calculated, and confounding was not adjusted for.
What did the largest-holder bands show?
Observed 24-hour death rates were 4.6% for 8–14% (n = 546), 2.1% for 15–29% (n = 512), and 60.4% for 30%+ (n = 636). The non-monotonic lower bands and possible confounding are reasons not to interpret 30% as a universal boundary.
How were unresolved and unavailable outcomes handled?
Rates use dead / (done + dead). Records marked no_data, pending, or otherwise unresolved were excluded. Their counts were not retained in the aggregate extract used for this article, so unavailable counts cannot be reported. Because availability may relate to token or market characteristics, that missingness may be informative and may bias the rates.
Can the findings be independently verified or generalized?
No public row-level export is available, and holder and market sources were not manually verified token by token. The selected, time-limited cohort may not represent other tokens or periods. Token age, liquidity, creator, market regime, and repeated analyses were not controlled, so the results should be treated as cohort-specific descriptive evidence.
Sources & further reading
Related guides
8 min read
The 24-Hour Survival Study
A 5,014-token Solana study reports 24-hour survival by market-cap band, with exact cohort rules, missing-data treatment and key limitations.
7 min read
Holder Concentration: the #1 Rug Signal
Why top-10 holder concentration is the most frequent danger flag in Solana meme coin analysis, what thresholds matter, and how to read a holder list like a pro.
6 min read
Insider Risk vs. 24-Hour Outcomes (Study)
A study of 2,259 Solana tokens compares analysis-time insider-risk scores with captured 24-hour prices, including exclusions and limitations.
6 min read
The Clean-Contract Safety Myth (Study)
Descriptive 24-hour outcomes for 3,744 clean-contract Solana tokens, with cohort selection, missing-data treatment, confounding and uncertainty disclosed.
Continue: Before You Buy
More evidence-led reading for this decision stage.
Run these checks automatically
Paste any Solana mint address into the MemeAssist analyzer for holder intelligence, authority checks, creator history and an AI health score in one report. New accounts get 1 free full token analysis with Trade Setup. Further analyses and other paid tools require account credit.
Try with 1 free full analysis