Bitquery finds most Pump.fun wallet gains hard to copy

A Bitquery study found that only 26% of $752.5 million in large memecoin wallet gains came from conventional trades, limiting what copy traders could replicate.

A Bitquery study published Oct. 22 found that only 26% of $752.5 million in reported gains from large memecoin wallets came from conventional market trades. The findings cover 3,170 wallet-token positions with gains of at least $100,000 recorded between October 2025 and August 2026.

In 51% of the positions, wallets sold tokens received from other addresses rather than tokens they had visibly purchased. Another 13% showed patterns linked to potentially coordinated trading, while 10% could not be classified conclusively.

Copy-trading systems monitor wallet purchases and place similar orders for followers. A wallet that receives tokens through a transfer can later record a large sale without showing the market purchase needed to create an equivalent copy-trading order.

Bitquery identified 1,579 profitable positions in which most of the sold tokens had arrived through transfers. Of those, 1,104 were linked directly or indirectly to addresses associated with the token’s creator. The study noted that transfers can distort profit calculations when analytics platforms compare recorded purchases with sales instead of tracking the original acquisition cost across several wallets.

Sales of transferred tokens accounted for about seven in 10 large wallet wins in August. The monthly total had remained below 100 cases through July.

In one example involving the UOTF token, four temporary wallets each received nearly 8 million tokens before forwarding them to selling addresses in a coordinated transaction. One recipient later sold its tokens for $1.59 million without a corresponding recorded market purchase.

Transfers do not by themselves prove manipulation because investors can move assets between wallets they control. Bitquery also identified patterns it classified as wash loops, in which concentrated buying coincided with direct financial links between wallets. Fifteen of the 25 largest reported gains met the study’s criteria for those patterns.

In one Arena token example, a wallet spent $4,223 and later recorded about $4.92 million in sales. Another address that had previously received funds from the apparent winner carried out 99% of the buying during the holding period. The seller had transferred more than $1.4 million in Solana to that buyer during the previous two days.

GMGN, a platform that lets users monitor wallets and copy trades, labels some addresses “Didn’t buy” and “Sold > Bought.” The labels identify wallets that received tokens or sold more than their recorded purchases. They do not establish whether followers could have achieved the same gains.

GMGN’s copy-trading system mirrors market purchases. A transferred token does not create an equivalent buy order. Followers can also receive different results because of liquidity and market-capitalization filters, transaction-size limits, slippage and execution delays. GMGN charges a 1% handling fee per transaction, excluding network and priority fees.

Bitquery found conventional profitable activity in the sample, including a SOULGUY position that generated $3.3 million across 3,632 trades. Replicating it would require followers to complete a large number of transactions at similar prices.

The study did not measure copy traders’ actual performance. It also excluded 2,943 six-figure winning positions that closed after the August cutoff, leaving their transfer patterns outside the analysis.

Articles by this author