Fascinating talk between @mevintern and @hasufl
Senpai gets surgical on the top MEV strategies today and also lays out a hollistic framework of a Searcher’s workflow.
Learned a ton – couldn’t walk away without giving back so I’m sharing my notes with CT
🧵..
Senpai gets surgical on the top MEV strategies today and also lays out a hollistic framework of a Searcher’s workflow.
Learned a ton – couldn’t walk away without giving back so I’m sharing my notes with CT
🧵..
https://twitter.com/hasufl/status/1417251801024405505
What actors exist in the MEV space today?
1) Miners - block producers in ETH today, validators in ETH2
2) Searchers - parties that express transaction ordering preferences
3) Users - end users of ETH
4) Dapps
5) Protocol developers - create the base architecture of the game
1) Miners - block producers in ETH today, validators in ETH2
2) Searchers - parties that express transaction ordering preferences
3) Users - end users of ETH
4) Dapps
5) Protocol developers - create the base architecture of the game
What is the role of a Searcher?
Is technical
Has a tight grasp of DeFi/ETH
Can pull data from chain
Can parse the data in his own system
Send transaction bundles to be mined
Is technical
Has a tight grasp of DeFi/ETH
Can pull data from chain
Can parse the data in his own system
Send transaction bundles to be mined
What is a bundle and what guarantees does it give to you or the users who submit it?
Bundle is a group of tx that a searcher submits to a miner
Bundle has to be executed in the ORDER submitted
This provides granular organization of txs for arbitrage maximization
Bundle is a group of tx that a searcher submits to a miner
Bundle has to be executed in the ORDER submitted
This provides granular organization of txs for arbitrage maximization
Bundles submitted as all txs included or not at all
Bundles are basically a social contract between flashbots relay, searcher and miner
Bundles are basically a social contract between flashbots relay, searcher and miner
Miners are sensitive of Searchers directly interacting with them:
No guarantee miner gets paid for bundles they accept
Miner has to expend computation resources to determine if bundle is useful
Miner could be spammed with shit bundles
Relationship must be established first.
No guarantee miner gets paid for bundles they accept
Miner has to expend computation resources to determine if bundle is useful
Miner could be spammed with shit bundles
Relationship must be established first.
PGA vs Flashbots:
* PGA battle -> low latency view of the mempool instead of writing the most gas-efficient contracts
* Competitive, with only 1 person that can capture the arb – every other bot will lead to failed/cancelled tx taking up block space and congesting the network
* PGA battle -> low latency view of the mempool instead of writing the most gas-efficient contracts
* Competitive, with only 1 person that can capture the arb – every other bot will lead to failed/cancelled tx taking up block space and congesting the network
Flashbots battle -> only the winner of the auction will land on chain.
In practice should lead to less congestion and high gas fees for users.
Flashbots turned the competition from having a tight view of the mempool to operating from a standpoint of gas-efficiency optimization
In practice should lead to less congestion and high gas fees for users.
Flashbots turned the competition from having a tight view of the mempool to operating from a standpoint of gas-efficiency optimization
Can we categorize MEV?
There are broadly 3 strategies:
1) Arbitrage – Senpai’s specialization
2) Liquidations
3) Sandwiches
With an extremely long-tail of MEV strategies that don’t fit here --> typically very specialized contracts
There are broadly 3 strategies:
1) Arbitrage – Senpai’s specialization
2) Liquidations
3) Sandwiches
With an extremely long-tail of MEV strategies that don’t fit here --> typically very specialized contracts
How many Searchers compete and fail?
Many people try – some may land a bundle
Has immediately become an extremely competitive game
Top heavy – like 10 teams that run 80% of MEV
Many people try – some may land a bundle
Has immediately become an extremely competitive game
Top heavy – like 10 teams that run 80% of MEV
On gas tokens
Destroying a contract refunds the gas spent on deploying it.
When prices low – mint gas contracts
When price high – destroy the MEV bundle to refund the gas
Net effect is to reduce the gas-price in the flashbots gas option
Destroying a contract refunds the gas spent on deploying it.
When prices low – mint gas contracts
When price high – destroy the MEV bundle to refund the gas
Net effect is to reduce the gas-price in the flashbots gas option
Any unique edges/advantages for Searchers?
One example: solidity has a bunch of guard-rails by default. Programming at assembly level languages allows searchers to sometimes for-go the guard-rails in the contracts they write to optimize for higher gas efficiency
One example: solidity has a bunch of guard-rails by default. Programming at assembly level languages allows searchers to sometimes for-go the guard-rails in the contracts they write to optimize for higher gas efficiency
Strategy Sustainability
Bundles r sent thru the flashbots relay to a miner, not thru public mempool
Although the miner and flashbots API can see your tx, atleast private from public mempool and hungry sandwich attackers/malicious threats
Bundles r sent thru the flashbots relay to a miner, not thru public mempool
Although the miner and flashbots API can see your tx, atleast private from public mempool and hungry sandwich attackers/malicious threats
Strategy Secrecy
Flashbots has an API – if you don’t send your bundle in the flashbots API – would get harder for competition to filter random long-tail MEV opportunities.
Keep high value strategies out of the public Flashbots API and the mempool by using other private relays
Flashbots has an API – if you don’t send your bundle in the flashbots API – would get harder for competition to filter random long-tail MEV opportunities.
Keep high value strategies out of the public Flashbots API and the mempool by using other private relays
Sandwich Attacks on UNI V3
The pools higher capital efficiency means a lot more funds are required to move the LP's curve and capitalize on the sandwich
Sandwich bots can get wrecked so they don't want to take a huge risk – this prices out a bunch of sandwich arbs.
The pools higher capital efficiency means a lot more funds are required to move the LP's curve and capitalize on the sandwich
Sandwich bots can get wrecked so they don't want to take a huge risk – this prices out a bunch of sandwich arbs.
Tip on gas optimization!
Initial storing of an ERC20 on a wallet requires more gas than subsequent interactions with it. If you’re running arb bots – having atleast 1 of a token in your wallet will cost you less gas on your arb tx than if you had to pay for the initial storage
Initial storing of an ERC20 on a wallet requires more gas than subsequent interactions with it. If you’re running arb bots – having atleast 1 of a token in your wallet will cost you less gas on your arb tx than if you had to pay for the initial storage
There were many more gems and certain topics such as competing bundles, miner's simulations of bundle profits, and other more technical concepts only @mevintern could say best!
MUST LISTEN interview for anybody wanting to up their understanding of MEV!
MUST LISTEN interview for anybody wanting to up their understanding of MEV!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
