VELLIN
A bot that snipes new tokens on Robinhood Chain at one fixed stake, thousands of times — and publishes the whole book: every position, every fee, every loss.
These docs describe what the bot does, how the numbers on the app are rebuilt from the chain, and what the token would and would not be. Where something is not built, this documentation says so rather than describing it in the present tense.
The state of each part
| Part | State |
|---|---|
| The sniping bot | Historical snapshot. Last recorded trade 2026-09-27 15:30 UTC. Wallet at snapshot 0xb6f8b683dd999593964e073c11788e49a56c43de |
| The reconstruction pipeline | Built. Read-only, rebuilds the full history from public RPC |
| The app | Published. A snapshot, not a feed. No wallet connection, no account, nothing to sign |
| Capital link between token and desk | Not built. The desk trades its own money |
| Payout rule and contract | Not built. No rule decided, no contract deployed |
| The proposed token | Official updates on @acreslol |
The desk has used 3 wallets
In the observed period, the operator moved between wallets; each new wallet started trading before the previous went quiet. They are the same desk — the sniping contract behind every wallet is the same bytecode, differing only in the owner address compiled into it, which is how each successor was identified (how the book is rebuilt).
| Wallet | Period | Realised | Closed | Win rate | State |
|---|---|---|---|---|---|
0xddf8cbf85ebdae4ca5c6a5d2291c310216d26c14 | 2026-08-31 → 2026-09-11 | +29.78 ETH net of recorded lost-race gas | 5 326 | 18.7% | Stopped at snapshot |
0xb770cf25c27358d303e49b40fab0c6a88a7988f4 | 2026-09-11 → 2026-09-19 | +15.93 ETH before lost-race gas | 2 202 | 22.3% | Stopped at snapshot |
0xb6f8b683dd999593964e073c11788e49a56c43de | 2026-09-19 → 2026-09-27 | +4.01 ETH before lost-race gas | 1 591 | 17.2% | Trading at snapshot |
Everything on this site adds those books together. Money adds up; the stake size and the entry speed do not, so those are reported per wallet.
The record, in one table
Everything below is measured over 26.86 days of trading, 2026-08-31 to 2026-09-27, at block 74,052,077.
| Realised, before the gas on lost races | +52.41 ETH |
| Realised, net of that gas | unknown ETH — measured only where the list of reverted transactions exists |
| Closed positions | 9,119 — 1,764 up, 7,216 down (19.3% win rate) |
| Profit factor | 2.46 |
| Transactions sent | 309,358, of which 171,591 are not in this snapshot |
| Of those read, reverted | unknown |
| Gas burned on reverted attempts | unknown ETH |
| Distinct tokens bought | 9,221 |
| Still with a live pool | 253 (2.7%) |
| Median entry | 15 blocks after a token's first transfer |
The last two lines are the shape of this strategy: 97.3% of what the bot buys stops trading, and the profit comes from the minority that does not.
Where to go if you want in
Follow @acreslol for official updates.
How to read these docs
Start with what sniping means here if you want the strategy. Start with how the book is rebuilt if you want to check the numbers. Start with the token mechanism if you want to know what would be on offer, and read the ceiling immediately after — it is the part that decides whether the idea scales.
Honesty rules these docs follow
- Where the data cannot support a number, the app and these docs print
unknowninstead of estimating. - Nothing that is not built is described in the present tense.
- Losses, reverted attempts and dead tokens are published with the same prominence as profit.
- Every number here is generated from the snapshot, not typed in by hand.