Robinhood Chain: Heavy Volume, Limited Directional Flow
Robinhood Chain's most-traded tokens showed substantial activity without much directional flow in many cases. Across 19 tokens in the 24 hours ending 2026-09-27, reported DEX volume was $206,279,008. SellTape measured $109,754,237 of selling and $104,837,711 of buying directly from pool transfers. The distinction is not simply between real traders and bots. High turnover, same-wallet round trips, liquidity operations and incomplete attribution each change what the headline volume tells a trader.
Key findings
- 13 of 19 tokens traded more than 20 times their pool liquidity.
- Buys and sells closely matched for most tokens, leaving small net flows.
- Round trips accounted for 30.4% of selling; six tokens exceeded 60%, excluding a wrapped-BTC artefact.
- Classified arbitrage bots accounted for 4.9% of selling, concentrated in three tokens.
- We excluded 11,999 liquidity transactions that tools could otherwise misread as trades.
Method: Read pool transfers, then separate their purpose
We selected the 19 most-traded tokens on Robinhood Chain by GeckoTerminal's reported 24-hour DEX volume. For each token, we read every transfer into and out of every DEX pool directly from the chain, covering Uniswap-style V2, V3 and V4 pools across approximately 860,000 blocks.
Tokens moving into a pool count as selling; tokens moving out count as buying. We measure net flow per transaction, so a route through multiple pools nets out rather than counting each intermediate movement as separate demand.
Before attributing sales, we removed every transaction that added or removed liquidity. We then assigned each sell to the wallet that sent the transaction and classified its activity. Categories include holders, arbitrage bots, unstakers, protocol or creator wallets selling minted supply, and activity we could not attribute within lookup limits.
Arbitrage classification uses transaction frequency, near-zero balances and buying and selling across pools. These are behavioural heuristics, not verified identities.
The volume comparison is a sanity check, but not an exact reconciliation. Measured sells of $109,754,237 plus measured buys of $104,837,711 exceed reported volume of $206,279,008. They are on a similar aggregate scale, but the supplied data do not establish why they differ. Reported volume and measured flows therefore remain separate series throughout this report.
Turnover: Activity is not executable depth
13 of the 19 tokens traded more than 20 times their pool liquidity during the window. That measures how much reported activity passed through relative to the quoted liquidity base. It does not measure how much a trader could execute without moving the price.
Reported volume relative to GeckoTerminal pool liquidity for the sampled tokens.
Token C recorded $16,140,949 in reported volume against $183,163 of pool liquidity. Token I recorded $11,925,382 against $26,661. Both illustrate why a large volume figure can coexist with a much smaller liquidity figure.
The most extreme liquidity readings require additional caution. Token H showed $12,782,482 in reported volume and $2 of pool liquidity; Token S showed $4,448,657 and $38. Those are supplied GeckoTerminal readings, not evidence that the full window's trading occurred against those exact liquidity amounts.
Pool liquidity can change, and a quoted total does not describe liquidity distribution around the current price. For execution, inspect the relevant pools and available depth rather than treating historical turnover as capacity.
Matched flows: Large gross activity, small net changes
For most sampled tokens, measured buying and selling were closely balanced. This limits the case for reading their gross volume as sustained directional demand.
Measured buys and sells after excluding liquidity transactions and netting routes within each transaction.
Token C had $8,353,717 of selling and $8,419,741 of buying, with reported net flow of $66,024. Token D had $8,537,728 of selling and $8,580,553 of buying, leaving $42,825 net.
Token A was a clear contrast. Selling of $14,067,704 exceeded buying of $8,004,082, producing net flow of -$6,063,622. Token R, a tokenized stock, also showed a directional imbalance: $1,077,808 sold against $464,522 bought, with net flow of -$613,286.
Matched totals do not establish that the same participants were on both sides. Independent buyers and sellers can balance in aggregate. Nor do small net flows imply stable prices. The useful distinction is narrower: high gross activity need not represent a large net transfer of tokens out of pools.
Round trips: Same-wallet buying matched substantial selling
For each wallet, round-trip volume is the portion of its selling matched by its own buying within the same window. Volume-weighted across the sample, that share was 30.4% of selling.
Same-wallet round-trip shares of selling; the wrapped-BTC line is a single-wallet artefact.
Six tokens exceeded 60% after excluding Token L, the wrapped-BTC line. Token C and Token I each registered 68.7%; Token K registered 66.3%, Token M 77.4%, Token O 61.1% and Token P 72.0%.
Token C had 1,588 detected round-trip wallets, the highest count in the sample. Token L instead showed 100.0% with a single detected wallet. That line is an artefact of the single-wallet observation, not comparable evidence of broad participation.
Round-trip volume is not proof of wash trading or of intent by any project. Market making, arbitrage, inventory adjustment and ordinary traders reversing positions can all produce buying and selling by the same wallet.
The detector also has a blind spot: activity split across different wallets. Tokens D and E showed near-zero detected round-trip shares despite closely matched buys and sells. That is consistent with wallet rotation, but does not prove it. Aggregate balance alone cannot identify coordination.
Seller composition: Classified arbitrage is a limited explanation
Arbitrage bots accounted for 4.9% of measured selling and were concentrated in three tokens. On this classification, “it's just bots” does not explain the broader pattern.
Seller attribution shares, including unchecked holders and activity left unattributed.
Checked holders accounted for 66.8% of selling. Holders not checked for bots accounted for 12.3%, while 15.8% was not attributed. Creator-wallet selling registered 0.1%, and unstakers registered 0.0%.
The holder label should not be read as proof of human discretion or independent demand. In particular, the unchecked holder bucket remains unresolved for bot activity. The classified arbitrage share is not a ceiling on all automated trading.
Coverage also varied by token. Unattributed selling reached 59.2% for Token J, 41.6% for Token B, a tokenized stock, and 37.0% for Token A. Conclusions about who sold those tokens deserve less confidence than their measured pool-flow totals.
Liquidity operations: Transfers that were not trades
We excluded 11,999 transactions that added or removed liquidity. Without separating these operations, transfer-based tools can label tokens deposited into pools as sales and tokens withdrawn as purchases.
That error affects more than volume. It can also assign apparent selling to liquidity providers who were funding a pool rather than executing a swap.
The exclusion applies to entire transactions containing liquidity additions or removals. Accordingly, this report describes the retained transaction set, not every possible swap leg embedded in mixed-purpose transactions.
Limitations
This is a single 24-hour snapshot. It cannot establish persistence, normal activity or a project's longer-term trading pattern.
Attribution remains incomplete: 15.8% of selling was not attributed, and wallet classifications are heuristic. Same-wallet matching misses activity distributed across wallets and cannot establish common ownership.
Pool liquidity comes from GeckoTerminal rather than a reconstruction of executable depth throughout the window. Reported volume also does not reconcile exactly with measured buys plus sells. These constraints matter most when comparing tokens with extreme turnover or uneven attribution coverage.
What this means if you trade these tokens
For traders, separate activity from direction and execution. Check gross volume, net pool flow and live depth independently. None substitutes for the others.
For market makers, balanced flows can still involve substantial inventory and execution risk. Same-wallet matching describes participation, not the spread available to capture or the cost of adverse selection.
For listing teams, high turnover warrants closer review, not an automatic rejection. Examine repeat participation, attribution gaps and liquidity operations before presenting volume as evidence of independent demand.
Run the same analysis on any Robinhood Chain or Ethereum token at t.me/SellTapeBot. SellTape is an information service, not financial advice.