Curriculum·J302 Swap Execution: AMMs, MEV, and DEX Architectures·about 31 min
Execution discipline on-chain
By the end of this lesson you can
- →Run the pre-swap checklist that the preceding four lessons produce
- →Explain why separate wallets do not protect against a repeated configuration error
- →Compute the annual cost of a default setting against the tail it permits
- →Record realized against quoted price so that execution becomes measurable
Junior · enrolled learners
This lesson opens with Six swaps, one setting, one day.
- What happened
- Following the sandwich attack of 12 March 2025 that turned a 220,764 USDC swap into about $5,271, the researcher known as DeFiac, citing internal tools, reported that the same unknown entity had been sandwiched on six separate USDC to USDT swaps that day across different wallets: 732,583 USDC in, 18,636 USDT out, about $714,000 lost, every swap traded with 100 percent slippage allowed, per DeFiac's post as quoted by ForkLog. Cointelegraph reported two of the other wallets sandwiched for $138,838 and $128,003 a few minutes earlier in the same pool. The wallets were separate. The transactions were separate. What was not separate was the configuration each swap was submitted with, which per J302-04 is the field a searcher reads to decide whether a transaction is worth acting on. Two observers raised the possibility that the swaps were deliberate: DeFiac called it either someone burning a lot of money or an odd attempt at laundering it, and DefiLlama's 0xngmi suggested that swaps this bad could be money laundering, so the reading below treats the configuration as the lesson and does not assume a careless victim.
- The decision point
- If the swaps were an operator's mistake, the lesson is about compartments. Using different wallets is compartmentalization, and per S209-03 compartmentalization only works when a failure inside one cannot reach the others. Here the failure was not in a wallet. It was in a setting that traveled with the operator to every wallet they used, so six compartments produced six independent instances of one error rather than containing it. A compartment that does not isolate the actual failure mode is a filing system.
- Recorded loss
- $714,000
What you will be able to answer
- →Why did separate wallets not help?
- →What is the pre-swap checklist?
- →What shape does execution cost have?
- →What makes execution measurable?
Orientation and Year One are open: anyone can read them without an account. From Year Two onward the lessons are for enrolled learners, because progress through the later years only means anything if it is tracked against a record.
It is free. We do not sell the list and there is nothing to buy at the end of it.
Sources and review
- https://cointelegraph.com/news/crypto-trader-loses-least-215000-stablecoins-from-sandwich-attack
- https://www.cryptotimes.io/2025/03/13/crypto-trader-loses-215k-in-mev-sandwich-attack/
- https://forklog.com/en/sandwich-attack-victim-swapped-732583-for-18636/
- https://blog.uniswap.org/minimize-slippage-on-swaps
- https://collective.flashbots.net/t/post-mortem-april-3rd-2023-mev-boost-relay-incident-and-related-timing-issue/1540
Confidence medium·Volatility medium·Reviewed 2026-08-06·Owner unassigned
Contested
The six-in-one-day attribution is a researcher's on-chain analysis linking wallets by behavior, citing internal tools, not a confirmed account of one person's activity. Marked medium confidence. The six-swap totals (732,583 USDC for 18,636 USDT, traded with 100 percent slippage allowed) are DeFiac's, quoted by ForkLog. Cointelegraph's $138,838 and $128,003 for two other wallets do not say whether they are swap sizes or losses, so this lesson does not add them to a total. An earlier version summed three figures to $482,341; that sum has been removed.
Whether these were losses at all is contested. DefiLlama's 0xngmi suggested that some very bad swaps could be money laundering, routed privately to a cooperating bot, and DeFiac called it either someone burning a lot of money or an odd attempt at laundering. Cointelegraph also reported that the swaps did not come from Uniswap's own front end, which has MEV protection and default slippage settings. The structural point about a configuration traveling with an operator holds either way.
Per P10 the cost figures in the worked example are illustrative of a trading frequency, not a prediction about any reader's results, and the lesson makes no claim that better execution produces profit.