By the On-Chain Research Desk. Field study window: 2026-04-10 to 2026-05-10. Sample frame anchored on the public whale_hunter Banana Gun Bot dashboard on Dune, supplemented with Cielo public leaderboards and Nansen Smart Money labels. Methodology constraints, exclusions, and survivorship caveats disclosed in full below. This is structured observation, not a statistically rigorous sample.
Methodology and What This Study Cannot Do
The 100-wallet sample is anchored on the public top-25 leaderboard from the whale_hunter Banana Gun Bot dashboard on Dune, queried 2026-05-10, and extended with mid-tier copier wallets surfaced through Cielo public leaderboards and Nansen Smart Money labels over the 30-day window of 2026-04-10 to 2026-05-10. The Dune leaderboard ranks by lifetime volume, not 30-day return, so we treat those rows as a population frame rather than a return panel. Outcomes are read through transaction-level data, not self-reported PnL screenshots. Two limits to flag up front. The study cannot recover wallets that blew up and dropped off public dashboards during the window, and it cannot reconstruct the full memecoin tail for unindexed wallets. Both are forms of survivorship bias, and both push the visible distribution upward relative to the unobserved truth. Bucket percentages in the table further down are directional readings of the sample, not survey-grade percentages, and the FAQ at the end of the piece summarizes the practical takeaways.
The Headline Distribution
Across the 100-wallet sample, the median copier landed close to break-even after fees and slippage over the 30-day window, the bottom decile finished meaningfully negative, and the top decile produced the multi-hundred-percent screenshots that dominate copy-trade marketing. A useful anchor sits in the underlying Dune snapshot. Across the top 20 wallets ranked by lifetime volume on the whale_hunter dashboard, sell volume exceeded buy volume on 17 of 20 wallets, the three exceptions all on Solana, and the aggregate sell-to-buy ratio across those 20 wallets read 1.0746 on the Dune snapshot. That asymmetry is what you would expect from a population filtered for survival: the visible top has been monetizing, and the wallets that have not are below the visible cut. Anyone reading “average copier returns” without that asymmetry is reading a number that already excludes the losers.
We routed copy entries through Banana Gun Pro for the Ethereum and Base subset of the sample, where Base Flashblock execution narrows the gap between target trade and copier landing to roughly 200 milliseconds. Solana entries used the same wallet-tracking workflow without the Flashblock floor, which widened the win-rate spread between same-block and next-block copies and pulled outcomes lower in the distribution. Execution latency moves a copier within an archetype, not across archetypes.
Who Made Money and What They Did Differently
The wallets that finished positive shared three patterns the losers did not. They concentrated capital in a small number of names rather than rotating across hundreds, they exited on rules rather than vibes, and they avoided KOL-signal copy flows. The Dune anchor illustrates the concentration pattern. Rank 4 logged $23,390,373 in lifetime volume across only 53 unique pairs over 16 active days, while rank 16 logged $16,335,771 across 19,310 unique pairs over 582 active days. Same order of magnitude in dollars moved, two orders of magnitude apart in dispersion. Checking concentration claims against the on-chain volume data exposed by public bot dashboards makes the asymmetry legible: lifetime volume tells you a wallet is active, not that copying every one of its trades was profitable.
In our 30-day window, the concentration-style wallets landed predominantly in the upper half of the return distribution, the high-dispersion rotators in the lower half, and the bottom decile was dominated by KOL-signal flows. Exit discipline did most of the remaining work. Wallets that mirrored entries but applied their own take-profit and stop-loss thresholds consistently outperformed pure-mirror copies of the same source, because the public top-tier targets often distribute into late copier volume on the way out.
The Three Archetypes and Their Distinct Distributions
The 100 wallets split into three archetypes that produce three different return shapes. Smart-money copying targets addresses that analytics platforms tag as consistently profitable across multiple tokens over weeks. The distribution here is the tightest, with most copiers landing in a band around the target’s reported PnL minus latency drag and fees, and this archetype filled most of the upper half of our 30-day distribution. Auto-mirror copying triggers an immediate copy on every target transaction, and the shape depends entirely on whether the target wallet is genuinely consistent or simply lucky on a handful of names. The Dune leaderboard captures both flavors. Rank 1 logged 9,548 trades across 1,986 unique pairs over 288 days, which reads as durable rotation, while rank 8 logged 8,802 trades across only 461 pairs in 89 days, which reads as a short, intense sprint.
Auto-mirror copies of the first profile sat near the median of our distribution. Auto-mirror copies of the second clustered in the bottom quartile because the sprint ended inside the window and copiers caught the unwind. KOL signal-following produced the widest distribution and the deepest losses, because the wallets behind public calls often distribute into copier volume themselves and pure mirror logic cannot fix that conflict. Archetype mix moves the copier far more than execution venue or position size, and the result holds whether you look at the smart-money slice in isolation or the auto-mirror cohort that dominated the middle band of the sample.
Results Distribution by Bucket
The table below summarizes how the 100-wallet sample sorted by 30-day outcome, dominant archetype, and the behavioral signal that defined each tier. Read it as a structured observation set, not as a forecast for any specific copier.
- Top decile, multi-hundred-percent. Roughly 10 percent of the sample. Dominant archetype: smart-money concentration. Top behavioral signal: low pair count combined with rule-based exits.
- Upper quartile excluding the top decile. Roughly 15 percent. Dominant archetype: smart-money plus selective auto-mirror. Top behavioral signal: target screened for multi-week consistency.
- Middle band, the median bucket. Roughly 45 percent. Dominant archetype: auto-mirror across mixed source quality. Top behavioral signal: net break-even after fees and slippage.
- Lower quartile excluding the bottom decile. Roughly 20 percent. Dominant archetype: auto-mirror of short-sprint targets. Top behavioral signal: caught the unwind inside the window.
- Bottom decile, meaningfully negative. Roughly 10 percent. Dominant archetype: KOL-signal copy flows. Top behavioral signal: source wallet distributing into copier volume.
- Out-of-sample tail not captured here. Wallets that dropped off public dashboards during the window. The visible sample under-represents this tier by construction, which pushes the true distribution lower than the buckets above report.
Bucket shares are directional readings of the sample. They exclude wallets that dropped off public dashboards during the window, which pushes the unobserved distribution lower than the table reports.
Frequently Asked Questions
What return should a typical copy trader expect over 30 days?
Close to break-even after fees and slippage at the median, meaningfully negative at the bottom decile, multi-hundred-percent at the top decile. The expectation is a distribution, not a single number. Where a specific copier lands is driven mostly by archetype choice, concentration discipline, and exit rules.
Why are the losers missing from public copy-trade leaderboards?
Public leaderboards rank wallets by visible activity over time. Wallets that blow up tend to stop trading from the same address and drop off the visible cut within days. The result is a population frame that systematically over-represents survivors and under-represents the realistic loss distribution facing a new copier.
Which archetype produced the worst outcomes in the 30-day window?
KOL signal-following dominated the bottom decile. Entry timing variance is large and the wallets behind public calls frequently distribute into copier volume on the way out, which forces pure-mirror copiers into the exit at the wrong moment. No execution-layer optimization removes that structural conflict.
Does execution speed matter more than wallet selection?
No. Execution latency, including the roughly 200-millisecond Base Flashblock floor, tightens outcomes within an archetype. Archetype mix and exit discipline move outcomes across the distribution. A faster route applied to the wrong target still ends in the bottom quartile.
