Julian's Notes

Analyzing MEV Transactions

This document is old and more of a scratchpad. Notes below may be incorrect.


  • USV2: Uniswap V2
  • SS: ShibaSwap



  1. https://etherscan.io/tx/<tx_hash>/advanced
  2. https://etherscan.io/txsInternal?a=<pair_contract_address>&&m=advanced
  3. Go to page w/ block in question, look at transactions around the <tx_hash> of interest
  4. https://etherscan.io/tx/<tx_hash>#statechange (can also see miner direct payouts; or if a miner made this transaction)

Reverse-Engineering Bots


  1. Bot smart contract: 0x78A55B9b3BBEffB36A43D9905F654d2769dC55e8 (#0x78A)
  2. EOA that triggers the withdrawals: 0xF46b0BDfB89B704D09AcF3AE722a49db0C7efe8B (#0xF46)
    • This contract triggers the #0x78A bot contract to send withdrawals to a subset (3 + 1) EOA tx issuer accounts (example)
    • Withdrawals happens every ~1 day
  3. EOA tx issuers: 0x3F6349708CB72750fFAF7F69fae94718b983de39, 0x55f6b9cb4CcF714BEc345a95DF3d1460Bc9d35Cb, ... (many of them)
    • Their ETH balances never go above 2 ETH (hard threshold?). Usually stays above 1 ETH.
  4. Strategy




  • Misclassification. This is a 0x-based trade that filled the order via "the best price" under the hood across multiple exchanges (see also matcha.xyz).

"Wavelet" Arbitrage

An arbitrage opportunity created due to a "big" trade. These opportunities cascade through exchanged until the price equilizes.


  • #0x78a --> #0xd0d SS(WETH --> SAITAMA) --> #0x78a --> #0x9cb USV2(SAITAMA --> WETH) --> #0x78a
  • $17 in gas fees, $330 in revenue
  • Seems a rookie mistake to route back to #0x78a instead of going directly to USV2
  • SS pair was proceeded by a $6.7k trade
  • It was the first trade on USV2
  • Odd: state change says #0x002 (a miner) was paid 0.007 ETH, but that's not shown in the decompilation (?)


  • A clean arbitrage of $1375 USD in profit. Didn't pay any gas or eth for this, but it ends with 3 suicides. Perhaps gas reclaimed from that(?) (double check this)


  • Miner finding the arbitrage and paying themselves. As a result no tx fees. Profit of about $50 USD. (double check this)


  • Transaction from 0xdfee68a9adb981cd08699891a11cabe10f25ec44 team
  • #0xdfee --> #0x9cbf UNIV2(WETH --> SAITAMA) --> #0xd0dc SHIBASWAP(SAITAMA --> WETH) --> #0xdfee
  • $40 revenue, $8 tx fee
  • Burned gas tokens to keep fees low
    • Does pay miner directly, so gas token burning may actually be to reduce slots used in the tx to encourage miner to include it at no loss
  • Opportunity was created from a $124 trade right before on the uniswap pair (only 2 transactions on that pair in the block, ShibaSwap had 1 transaction in that block)


  • Another transaction from 0xdfee68a9adb981cd08699891a11cabe10f25ec44 team
  • #0xdfe ($1455) --> UNIV2 GALA --> #0x000 --> UNIV3 GALA --> #0xdfe ($1546)
  • $100 revenue, "$0" tx fee due to using gas tokens
  • This was the first UNIV2 GALA trade on the block, subsequent transactions are getReserves checks from other bots(?); this bot got in first
  • Opportunity looks like it was opened from the first trade on this block of UNIV3 GALA ($5k worth of GALA --> ETH), this bot was the second trade that came in just after. Subsequent transactions are getReserves checks from other bots(?); this bot got in first


  • Split profits w/ miner (miner took 48 ETH, they took 5 ETH); they took exactly 10%, and gave 90% to miner
  • 70 ETH -(SS)-> 197k YGG -(UV3)-> USDC -(UV3)-> 123 ETH



  • A sandwich attack, made ~0.033 ($95 CAD), gave a 0.01 ETH to miner. Misreported on MEV Explorer. Oddly combines a swap between Sushi and Uniswap as well.


  • #0x000 --> #0xa5e US V2 (WETH --> STARL) --> #0x2a5 SS (STARL --> WETH) --> #0x000
  • $11000 revenue, w/ $50 to miner
  • Looks like a standard bundle

Referred in

Analyzing MEV Transactions