TL;DR LPs are losing out to MEV extraction and LVR in today’s DeFi landscape. But several new AMM designs such as HOT are pioneering ways to protect LPs by reducing LVR. These approaches rely on offchain systems such as oracles and auctions and they offer a preview into how LPing could become more sustainable in DeFi in the future.

Key Takeaways:

• Due to the impact of MEV extraction and LVR, LPs are getting burned on early-generation AMMs.

• LVR is a metric for quantifying losses resulting from arbitrage-related MEV. When arbitrageurs target price discrepancies in AMM pools, LPs suffer from LVR.

• AMMs can use offchain systems such as oracles and auctions to directly reduce LVR, offering benefits to LPs.

• Arrakis is building onchain solutions that vertically integrate into the MEV supply chain to empower LPs. The HOT AMM, which reduces LVR through a request-for-quote system, is the first of these solutions.

AMMs have been at the heart of DeFi since it began. Uniswap and other similar AMMs were successful because they took a radically new approach to facilitate asset trading, using algorithms to replace centralized intermediaries. Uniswap’s liquidity pools are decentralized and permissionless so anyone can participate. But they’re designed for two types of user: LPs and swappers.

Arbitrageurs are also key actors on AMMs. They provide an important service: by balancing out price discrepancies, they ensure pools have updated prices. However, they’re also responsible for the biggest portion of MEV extraction. Arbitrage accounts for about ¾ of the $1 billion worth of MEV leaked since the Merge.

When MEV gets leaked to arbitrage, LPs carry the cost. Arbitrageurs step in when LPs offer stale prices, profiting from the price difference on another pool or CEX. In doing so, they correct the discrepancy and eat into the LPs’ fees. The cost of LPing is commonly referred to as Loss-Versus-Rebalancing (LVR), a metric that represents the “loss” of holding an LP position “versus rebalancing” at the updated CEX price.

In today’s DeFi landscape, passive LPs are losing out to sophisticated actors that dominate the MEV supply chain and the problem is getting worse. However, some innovative AMM designs have emerged with ways to protect LPs. These solutions typically take one of two approaches to solve the problem: LVR reduction and mitigation.

For the first edition of a two-part series, we explore how AMMs are directly reducing LVR below. These solutions rely on offchain systems such as oracles and batch auctions. Our findings point us to a conclusion that strengthens our conviction in DeFi’s future: the AMM space is entering a renaissance and next-generation solutions that protect LPs will be central to the space’s evolution.

Reducing LVR to Solve the MEV Crisis

Due to MEV extraction and LVR, LPs are getting burned on AMMs today. To protect themselves against MEV and arbitrage-related losses, LPs frequently hedge against their position by farming rewards. LVR refers to losses accrued from a position that cannot be hedged—it’s an opportunity cost that comes from providing liquidity to an AMM.

We can think of LVR as the cost of LPing. Arbitrageurs are responsible for this cost because they search pools for opportunities to extract MEV.

As arbitrage accounts for most MEV extraction (and LVR), and AMMs need liquidity from LPs, many teams have worked on new designs focused on reducing its impact. The current solutions that aim to reduce LVR rely on offchain systems. This is because arbitrage occurs due to price discrepancies between trading venues or pools. Notably, some offchain systems can provide information faster than DEX prices update, minimizing the window of opportunity for arbitrageurs.

Price Oracles

Uniswap uses the bonding curve formula x*y=k to determine the prices of assets in any given pool, where k is an invariant and x and y represent the amount of each type of token in the pool. LPs add two tokens to a pool and arbitrageurs rebalance pools when they have an opportunity to profit.

While x*y=k has served early-generation AMMs well, new solutions are exploring how price oracles could be used to update asset prices. These designs have proposed using oracles to build dynamic AMMs or make a quote based on the current market price.

Using Oracles to Build Dynamic AMMs

Instead of using x*y=k, dynamic AMMs propose using price oracles to dynamically define the price curve. Philippe Bergaut et al. explore this idea in depth in the research paper Dynamic Automated Market Makers for Decentralized Cryptocurrency Exchange. They argue that using oracles can eliminate arbitrage as pool prices would update continuously with the market price. To date, this approach has not been widely explored in DeFi.

Using Oracles to Make Quotes Based on Market Rates

AMMs can alternatively use oracles to feed in market prices from trading venues like CEXs. With this design, the AMM makes a quote for a bid or ask based on the oracle’s price updates. This is similar to how market makers facilitate OTC trading in TradFi. As the oracle receives price updates from CEXs faster than AMM pools can update, this would theoretically minimize arbitrage opportunities, thereby reducing the cost of LPing.

Such an AMM design would ideally use multiple oracles to get information on variants like drift and volatility—not only market prices. Bhaskar Krishnamachari, Qi Feng, and Eugenio Grippo discuss this approach in Automated Market Makers: Mean-Variance Analysis of LPs Payoffs and Design of Pricing Functions.

Whether using oracles for dynamic AMMs, quotes for bids, or another design, they pose some risks. First, they can be manipulated. They can also suffer from delays, leaving LPs exposed to greater losses in more volatile pairs. Still, given that oracles can help AMMs update prices faster than Ethereum L1’s 12-second block times, LPs could benefit from them if AMMs use them in a manner that circumvents the risks.

In today’s landscape, LPs are not profitable due to MEV extraction. Oracle-based approaches to AMM design offer a way to reduce the cost of LPing by eliminating stale prices, albeit with some risk.

Batch Trading and Order Flow Auctions

Order Flow Auctions (OFAs) have been reducing the impact of MEV extraction since Flashbots implemented them to redirect arbitrage profits to users. OFAs use bidding systems where actors in the MEV supply chain compete for the right to make a transaction. When the auction closes, a portion of the winning bid goes back to the party that generated the MEV.

Some new AMM designs have explored batching trades through auctions to tax arbitrageurs when they extract LVR. With batch trading, many transactions are settled at once within a set time parameter, minimizing arbitrage and eliminating front running. In such designs, arbitrageurs compete based on price rather than speed.

If we think of arbitrageurs as DoorDash drivers, batching prioritizes the driver who can deliver the meal at the best price within a specified cut-off window (not the driver who can deliver fastest).

CoW AMM is an innovator in MEV protection. CoW AMM asks solvers to bid in batch auctions for the right to rebalance pools. The highest bidder gets to arb the pool and the proceeds go to LPs, eliminating LVR. As CoW is an intent-based protocol, swappers also have less MEV risk because it gets passed onto solvers.

Ångstrom by Sorella Labs is another forthcoming AMM that implements batch auctions. Ångstrom uses auctions to find the best settlement price for swappers. Arbitrageurs bid across Uniswap V4 pools for the right to arbitrage a pool and the best price gets submitted to the hook. Swappers get to execute their trade close to the market price and, as with CoW, they’re protected from frontrunning.

In such designs, the auction winners are sophisticated entities that bribe builders to get their transactions included at the top of the block. Auctions use economic incentives to ensure that builders get rewarded for reducing MEV, arbitrageurs get rewarded for keeping prices updated, and LPs get protected from LVR.

Request-for-Quote and the HOT AMM

Request-for-Quote (RFQ) systems introduce signed quotes for the right to access liquidity in a pool. This approach is similar to batched OFAs in that it uses auctions to find a market price for liquidity. In HOT, the MEV-aware AMM jointly developed by Arrakis and Valantis Labs, solvers receive a guaranteed quote to fulfill users’ intents via The Arrakis Quoting Service. This service uses offchain strategies to inform the pricing of each signed quote, meaning solvers get fair price quotes that closely match onchain prices. This results in more healthy volume and fees for LPs.

HOT is a liquidity pool built on Valantis and it forms part of the system’s modular architecture. As DeFi’s first modular DEX, HOT is blazing the trail in the AMM design space with a focus on reducing MEV and LVR.

HOT’s intent-centric design is key to making LPs more profitable. The system also implements dynamic fees for swappers, where fees increase until the price resets with a new solver update. As dynamic fees increase with time, they disincentivize arbitrageurs.

When more swaps are executed, fees stay lower for longer. Arbitrageurs need low fees to be profitable, but as prices have less time to deviate from the market price, there are fewer arbitrage opportunities.

In short, dynamic fees keep prices in the pool in line with CEX prices, and this protects LPs from LVR. When fees are low, prices stay close to the market price, and when they’re high, arbitrageurs can not profit.

HOT is the first solution from Arrakis that focuses on reducing LVR and MEV to protect LPs. Arrakis plans to vertically integrate into the MEV supply chain to make onchain markets healthier and HOT is the first step towards that goal.

Why AMMs Are Evolving to Reduce LVR

For as long as DeFi has existed, MEV extraction has impacted users. Arbitrage accounts for the bulk of MEV today, creating a cost to LPing known as LVR. LPs keep losing out to arbitrageurs and DeFi needs a fix.

To solve this issue, several new AMM designs have proposed directly reducing the cost of LPing. Whether through price oracles, batched auctions, or signed quotes, these approaches rely on offchain sequencing to reduce LVR.

MEV gets extracted when actors in the MEV supply chain engage in offchain brokering. While there are trade-offs to introducing offchain systems, AMMs are increasingly finding ways to use them to drive better outcomes for the users their systems were originally designed for.

Additionally, MEV researchers have recently begun to explore how Application Specific Sequencing (ASS) could reduce LVR. ASS offers applications a way to reduce MEV by setting their own rules for ordering transactions, though this space is still in its infancy today.

While MEV research is still in its early stages, many next-generation AMMs are taking innovative approaches to reduce LVR and protect LPs from arbitrage-related MEV. HOT is one such solution that proves the AMM space is evolving and DeFi is set to benefit.

We believe that DeFi needs sustainable onchain markets. To achieve sustainability, AMMs need to attract LPs, and LPs need to be incentivized to provide liquidity. By implementing an RFQ system to reduce LVR, HOT aims to make LPs more profitable.

Our research into LVR reduction cements our belief that the AMM space is entering a renaissance and MEV protection is set to play a key role in DeFi’s future.

We explore this idea further in part two of this blog series, where we discuss how next-generation AMMs are using MEV recapture mechanisms to mitigate LVR. Arrakis is positioned to be a key player in this story and we’ll explain how in our next post. Keep an eye on our blog to learn more and follow us for more exciting updates dropping soon.

Editor’s note: This piece was inspired by an in-depth research report compiled by Ari Rodriguez and Daniel Contreras at Arrakis Finance. Thanks to both for their feedback and editing contributions to this piece. Read the full report here.

References

• ALGEBRA Ecosystem: Decentralized exchange [Algebra Finance]

• am-AMM: An Auction-Managed Automated Market Maker [Austin Adams et al. in arXiv preprint arXiv:2403.03367]

• Ambient Finance Documentation [Ambient Finance]

• Ångstrom Documentation [Sorella Labs, private]

• Announcing Oval: Earn Protocol Revenue by Capturing Oracle MEV [Hart Lambur for UMA]

• Application Specific Sequencing: Justin Drake, Apriori, Robert Miller, Stephane Gosselin Hart Lambur [Flashbots]

• Are Retail LPs Going Extinct? Exploring the MEV Supply Chain [@dreamsofdefi for Arrakis Finance]

• Automated Market Makers: Mean-Variance Analysis of LPs Payoffs and Design of Pricing Functions [Philippe Bergault et al. in arXiv:2212.00336]

• Automated Market Making and Loss-Versus-Rebalancing [Jason Milionis et al. in arXiv preprint arXiv:2208.06046]

• CoW Documentation [CoW]

• Diamond: A Uniswap V4 Hook minimizing LVR [Arrakis Finance, Conor McMenamin and Uniswap Foundation]

• Dynamic Fees for Automated Market Makers: Liquidity, Volatility, and Collected Fees [atiselsts.eth]

• Dynamic Automated Market Makers for Decentralized Cryptocurrency Exchange [Bhaskar Krishnamachari, Qi Feng, and Eugenio Grippo for University of Southern California]

• HOT, the MEV-Aware AMM Built to Empower LPs, Is Live [@dreamsofdefi for Arrakis Finance]

• Hybrid Order Type: A New MEV Aware AMM Design [Arrakis Finance and Valantis Labs]

• Introducing the Arrakis DVMM: DeFi’s First MEV-Aware Onchain Market Maker [@dreamsofdefi for Arrakis Finance]

• JOE v2.1 Liquidity Book [Trader Joe]

• LVR from Theory to Practice: A Survey of Next Generation DEX Designs [Ari Rodriguez and Daniel Contreras for Arrakis Finance]

• MEV-Share: programmably private orderflow to share MEV with users [Bert Miller for Flashbots]

• Uniswap Documentation [Uniswap]