Delayed Copy Trading Strategy Yields Losses on Solana
A study conducted in late August 2026 shows that following large wallet clusters into new tokens results in significant losses.
Updated on Oct. 20, 2026 in Investing
Live Poll
Would you use automated copy-trading software to manage your cryptocurrency investments?
Research from late August 2026 reveals that delayed copy trading on Solana platforms Pump.fun and Stonk often leads to substantial financial declines. An automated test of 93 tokens resulted in an average loss of 41% per token, highlighting the risks of chasing large wallet clusters.
Why it matters
These findings suggest that retail investors attempting to mirror the activity of large wallet clusters two minutes after their initial purchases are likely to underperform. The data illustrates how quickly liquidity shifts in new token markets, making delayed entry strategies ineffective.
The delayed copy-trading test saw a median decline of 95% across 93 tokens. A neural network model, which analyzed 771 picks over eight days, similarly failed to gain traction, with 55% of its selections triggering a stop loss within 24 hours.
The players
Pump.fun
This is a prominent decentralized platform built on the Solana blockchain that facilitates the creation and trading of memecoins.
Stonk
Stonk is a trading platform operating within the Solana ecosystem that allows users to engage in speculative token transactions.
Solana
Solana is a high-performance blockchain known for its fast transaction speeds and its expansive ecosystem of decentralized finance applications.
The details
The analysis utilized neural network modeling, APIs, and virtual private servers to track 7.7 million wallets and 1.7 million liquidity pools. Researchers measured 22.8 million first-buy transactions to assess whether following identified large-wallet groups could produce viable returns.
Timeline
The review of new-token trading on Pump.fun and Stonk began in late August 2026.
Researchers measured 22.8 million first buys over a period of two weeks.
A neural network model tested 771 token picks over an eight-day period.
Market Dynamics
This study follows a pattern set by the high-frequency trading latency arbitrage model, where delayed participants are consistently outpaced by earlier actors. It highlights the structural difficulty for retail traders to compete against automated systems in decentralized environments.
Retail investors should be aware that attempting to copy high-volume traders on decentralized platforms often exposes them to significant losses due to transaction timing. Relying on automated signals or following large wallet clusters after the initial buy typically results in diminished portfolio performance.
The takeaway
The data demonstrates that even sophisticated neural network models face substantial hurdles when attempting to mirror the activity of early market participants in volatile token markets. Investors are better served by developing proprietary strategies rather than relying on delayed entry tactics.
Further reading
For more on the risks associated with automated strategies, visit our Investing section.
Live Poll
Would you use automated copy-trading software to manage your cryptocurrency investments?
