Highlights:

(i) Classify Uniswap & CPMM-AMMs versus oracle-based AMMs.

(ii) CPMM suites new tokens, but exposes to high impermanent loss, arbitrage and sandwich attacks.

(iii) Half of Uniswap V3’s LPs lose due to high IL while suffer complex LP management. On the entire Ethereum, LPs have lost $745M since the Merge.

(iv) Oracle-based AMMs outperform CPMM over CEX-listed assets, bridged cryptos, particularly tokenized real-world assets.

Uniswap revolutionized the decentralized exchange (DEX) landscape by introducing a novel automated market maker (AMM) model. This model, pioneered by Uniswap, leverages liquidity pairs (aka pools) of two tokens to facilitate peer-to-peer trading without the need for a centralized order book. Prior to Uniswap, decentralized exchanges like IDEX, Ether Delta, and FolkDelta relied on off-chain limit order books (LOBs), which suffered from slow speeds, poor user experience, and low liquidity. While Bancor and Kyber attempted to address these issues with their peer-to-peer trading models based on token reserves, they ultimately faced challenges with liquidity exhaustion and complex asset management. For a more comprehensive overview of the evolution of DEXes, please refer to [1, 2].

Uniswap pioneered a revolutionary spot trading model, Automated Market Making (AMM), in November 2018. Its core innovation, the constant product function $x\times y = k$, disrupted traditional price discovery mechanisms like LOBs. Uniswap's model democratized liquidity provision, allowing anyone to create and contribute to liquidity pools of any pair of tokens. This innovation fueled the DeFi boom of 2020-2021. Other AMMs, such as Kyber, Sushiswap, and Pancakeswap, adopted the constant product model, solidifying it as a cornerstone of DeFi. Uniswap and CPMM-based AMMs offer several key advantages:

• Permissionless listing: anyone can create a liquid pool pairing two any tokens.

• Crowdsourcing liquidity: everyone can contribute to any liquid pool.

• Seamless trade: users easily swap for their wished tokens without KYC or intermediaries, provided unbounded liquidity.

• Non-custodial security: all trades are securely executed onchain by smartcontracts, eliminating the need of centralized intermediaries.

Balancer expanded the CPMM model to support multi-asset liquidity pools, using the equation $x_1 \times x_2 \times … \times x_n=k$. Curve introduced a constant sum market making (CSMM) model, $x+y=k$, tailored for stablecoin pairs (e.g. DAI, USDT, etc).

Despite Uniswap and CPMM-based AMMs have achieved significant success, they face challenges such as low capital efficiency, high slippage, and impermanent loss (IL). Uniswap V3 addressed these issues by introducing liquidity concentration, a modification of the original CPMM model. However, as research by Vladislav Goncharov [3] indicates that nearly half* of Uniswap V3's liquidity providers still experience losses due to high IL and out-of-range positions*. Moreover, managing concentrated liquidity positions can be complex and requires specialized financial knowledge and professional skills, making it less accessible to average users.

Impermanent loss (IL) is an inherent risk faced by market makers (MMs) across all financial markets. MMs on CEXs and LOB-based DEXs have active control over their liquidity provision strategies, enabling them to mitigate IL. Unfortunately, LPs on CPMM-based AMMs have limited influence over pool prices. This passive role exposes LPs to price deviations between AMMs and external markets. Arbitrageurs can exploit these discrepancies, profiting at the expense of LPs. For instance, if the external market price for ETH is 2500 USD, a CEX market maker can quickly adjust their bids and asks around 2500 to maintain a competitive price, in real-time manner. However, on a CPMM-based AMM, the pool price is determined solely by the token reserve ratio (e.g. 1 ETH : 2510 USD), making it impossible for LPs to adjust the price to match external market conditions. Moreover, price on AMMs is discretely updated block by block, not continuously as CEXes, causing gaps in price moves. This leaves the opportunity for arbitrageurs to profit, ultimately reducing the gains of LPs. According to Sorella Labs, LPs on Ethereum have lost $750M for arbitrageurs since the Merge (see real time tracking on Sorella website).

Obviously, oracle-based AMM models offer a significant advantage by eliminating arbitrage opportunities, including round-trip and sandwich attacks. Research by Álvaro Cartea and financial scientists at the University of Oxford [4] demonstrated that LPs on CPMM-AMMs are often providing liquidity at a loss. To address this issue, they proposed alternative models that leverage oracle prices. The paper of Philippe Bergault and financial mathematicians in Paris [5] introduced an AMM model combining external price, showing that even with lagged oracle data and adverse market conditions, optimized oracle-based mechanisms outperform traditional CPMM-AMMs. Additionally, S. Jiang et al [6], B. Krishnamachari et al [7], and D.J. Im [8] have further improved AMM design using oracle prices. These advancements aim to protect the interests of LPs and average traders by mitigating the risks associated with traditional AMM models.

Oracle-based AMMs rely on accurate price quote from external markets. To assess the dominance of external markets, we can compare trading volumes. As of September 19, 2024, the total ETH trading volume (24H) on all DEXes was less than $1.2 billion (source: DefiLlama), while the volume on the top 50 CEXes was approximately $11.8 billion (source: CoinGecko). This significant disparity, nearly a 10x difference, highlights the dominance of CEXes in terms of trading volume.

In summary, Uniswap and CPMM-AMMs, with their independent price discovery mechanisms, offer a unique advantage to provide liquidity for new tokens that may not be available on centralized exchanges. However, their passive market-making nature, susceptibility to arbitrage and sandwich attacks, and potential for huge LP losses, particularly with popular and highly liquid assets, present limitations. Oracle-based AMM models emerge as a promising alternative and complementary solution. By incorporating external price feeds, these models can address the shortcomings of traditional AMMs, potentially offering a more efficient and profitable liquidity provision mechanism (read examples how oracle-based AMMs outperform Uniswap). This implies that oracle-based AMMs can be the best solution to trade tokenized real world assets onchain, whose volume on CEXes is dominant.

All emerging L1 and L2 ecosystems rely on bridged assets (e.g., wETH, wBTC, wDAI, ckUSDT, USDC) and their native tokensfor liquidity onchain. Oracle-based AMMs offer a simple and capital-efficient solution for swapping native tokens against bridged assets (e.g., STRK/ETH, STRK/USDC, SUI/wETH, SUI/USDC, ICP/ckUSDT). Unlike CPMM-AMMs, which are best suited for CEX-unlisted tokens, oracle-based AMMs excel in trading CEX-listed crypto-assets. Dodo and Lifinity are notable examples of successful oracle-based AMMs. However, Dodo's capital efficiency is limited, with a maximum of 10x compared to Uniswap V2, and Lifinity restricts liquidity provision to a strictly selected group of LPs. BrownFi addresses these limitations by introducing an innovative oracle-based AMM with 100x higher capital efficiency than Uniswap V2, flexible market making, a user-friendly interface, and open access for all LPs. BrownFi's mathematical model, inspired by the statistical properties of Limit Order Books, enables these advanced features.

Remember that DEXes cannot replace CEXes, CPMM cannot replace limit order-book. They all live together to provide multiple market making and trading options under various circumstances. Although CPMM-based AMMs (e.g. Uniswap) are a popular solution for onchain liquidity in general, oracle-based AMMs (e.g. BrownFi) play the role of a complementary block in DeFi. Specifically, regarding average users farming onchain liquidity of CEX-listed assets with well-liquid, BrownFi AMM fits better, and is potential to outperform Uniswap V3.

Website, Twitter, Telegram, blog

[1] Daniel Phillips, 2023, The Evolution of Decentralized Exchanges

[2] Thuat Do et al, 2023, Derivable: A Novel Derivatives Pricing Markets based on a Family of Asymptotic Power Curves

[3] Vladislav Goncharov, 2023, Hackernoon, Half of Uniswap v3 Users Lose Money — Here's Why

[4] Álvaro Cartea, et al, University of Oxford - Institute of Quantitative Finance, 2023, Automated Market Makers Designs beyond Constant Functions

[5] Philippe Bergault et al, 2024, Automated market makers: mean-variance analysis of LPs payoffs and design of pricing functions

[6] S. Jiang, J. Chen, F. Li, H. Geng and H. Chi, "DCAMM: Dynamic Curve-Based Automated Market Maker," GLOBECOM 2023 - 2023 IEEE Global Communications Conference, Kuala Lumpur, Malaysia, 2023, pp. 4491-4496, doi: 10.1109/GLOBECOM54140.2023.10437668.

[7] Bhaskar Krishnamachari, Qi Feng, 2021, Eugenio Grippo Viterbi School of Engineering, University of Southern California, Dynamic Automated Market Makers for Decentralized Cryptocurrency Exchange

[8] Daniel Jiwoong Im, Alexander Kondratskiy, Vincent Harvey, Hsuan-Wei Fu, 2024, UAMM: Price-oracle based Automated Market Maker

Highlights:

(i) Classify Uniswap & CPMM-AMMs versus oracle-based AMMs.

(ii) CPMM suites new tokens, but exposes to high impermanent loss, arbitrage and sandwich attacks.

(iii) Half of Uniswap V3’s LPs lose due to high IL while suffer complex LP management. On the entire Ethereum, LPs have lost $745M since the Merge.

(iv) Oracle-based AMMs outperform CPMM over CEX-listed assets, bridged cryptos, particularly tokenized real-world assets.

Uniswap revolutionized the decentralized exchange (DEX) landscape by introducing a novel automated market maker (AMM) model. This model, pioneered by Uniswap, leverages liquidity pairs (aka pools) of two tokens to facilitate peer-to-peer trading without the need for a centralized order book. Prior to Uniswap, decentralized exchanges like IDEX, Ether Delta, and FolkDelta relied on off-chain limit order books (LOBs), which suffered from slow speeds, poor user experience, and low liquidity. While Bancor and Kyber attempted to address these issues with their peer-to-peer trading models based on token reserves, they ultimately faced challenges with liquidity exhaustion and complex asset management. For a more comprehensive overview of the evolution of DEXes, please refer to [1, 2].

Uniswap pioneered a revolutionary spot trading model, Automated Market Making (AMM), in November 2018. Its core innovation, the constant product function $x\times y = k$, disrupted traditional price discovery mechanisms like LOBs. Uniswap's model democratized liquidity provision, allowing anyone to create and contribute to liquidity pools of any pair of tokens. This innovation fueled the DeFi boom of 2020-2021. Other AMMs, such as Kyber, Sushiswap, and Pancakeswap, adopted the constant product model, solidifying it as a cornerstone of DeFi. Uniswap and CPMM-based AMMs offer several key advantages:

• Permissionless listing: anyone can create a liquid pool pairing two any tokens.

• Crowdsourcing liquidity: everyone can contribute to any liquid pool.

• Seamless trade: users easily swap for their wished tokens without KYC or intermediaries, provided unbounded liquidity.

• Non-custodial security: all trades are securely executed onchain by smartcontracts, eliminating the need of centralized intermediaries.

Balancer expanded the CPMM model to support multi-asset liquidity pools, using the equation $x_1 \times x_2 \times … \times x_n=k$. Curve introduced a constant sum market making (CSMM) model, $x+y=k$, tailored for stablecoin pairs (e.g. DAI, USDT, etc).

Despite Uniswap and CPMM-based AMMs have achieved significant success, they face challenges such as low capital efficiency, high slippage, and impermanent loss (IL). Uniswap V3 addressed these issues by introducing liquidity concentration, a modification of the original CPMM model. However, as research by Vladislav Goncharov [3] indicates that nearly half* of Uniswap V3's liquidity providers still experience losses due to high IL and out-of-range positions*. Moreover, managing concentrated liquidity positions can be complex and requires specialized financial knowledge and professional skills, making it less accessible to average users.

Impermanent loss (IL) is an inherent risk faced by market makers (MMs) across all financial markets. MMs on CEXs and LOB-based DEXs have active control over their liquidity provision strategies, enabling them to mitigate IL. Unfortunately, LPs on CPMM-based AMMs have limited influence over pool prices. This passive role exposes LPs to price deviations between AMMs and external markets. Arbitrageurs can exploit these discrepancies, profiting at the expense of LPs. For instance, if the external market price for ETH is 2500 USD, a CEX market maker can quickly adjust their bids and asks around 2500 to maintain a competitive price, in real-time manner. However, on a CPMM-based AMM, the pool price is determined solely by the token reserve ratio (e.g. 1 ETH : 2510 USD), making it impossible for LPs to adjust the price to match external market conditions. Moreover, price on AMMs is discretely updated block by block, not continuously as CEXes, causing gaps in price moves. This leaves the opportunity for arbitrageurs to profit, ultimately reducing the gains of LPs. According to Sorella Labs, LPs on Ethereum have lost $750M for arbitrageurs since the Merge (see real time tracking on Sorella website).

Obviously, oracle-based AMM models offer a significant advantage by eliminating arbitrage opportunities, including round-trip and sandwich attacks. Research by Álvaro Cartea and financial scientists at the University of Oxford [4] demonstrated that LPs on CPMM-AMMs are often providing liquidity at a loss. To address this issue, they proposed alternative models that leverage oracle prices. The paper of Philippe Bergault and financial mathematicians in Paris [5] introduced an AMM model combining external price, showing that even with lagged oracle data and adverse market conditions, optimized oracle-based mechanisms outperform traditional CPMM-AMMs. Additionally, S. Jiang et al [6], B. Krishnamachari et al [7], and D.J. Im [8] have further improved AMM design using oracle prices. These advancements aim to protect the interests of LPs and average traders by mitigating the risks associated with traditional AMM models.

Oracle-based AMMs rely on accurate price quote from external markets. To assess the dominance of external markets, we can compare trading volumes. As of September 19, 2024, the total ETH trading volume (24H) on all DEXes was less than $1.2 billion (source: DefiLlama), while the volume on the top 50 CEXes was approximately $11.8 billion (source: CoinGecko). This significant disparity, nearly a 10x difference, highlights the dominance of CEXes in terms of trading volume.

In summary, Uniswap and CPMM-AMMs, with their independent price discovery mechanisms, offer a unique advantage to provide liquidity for new tokens that may not be available on centralized exchanges. However, their passive market-making nature, susceptibility to arbitrage and sandwich attacks, and potential for huge LP losses, particularly with popular and highly liquid assets, present limitations. Oracle-based AMM models emerge as a promising alternative and complementary solution. By incorporating external price feeds, these models can address the shortcomings of traditional AMMs, potentially offering a more efficient and profitable liquidity provision mechanism (read examples how oracle-based AMMs outperform Uniswap). This implies that oracle-based AMMs can be the best solution to trade tokenized real world assets onchain, whose volume on CEXes is dominant.

All emerging L1 and L2 ecosystems rely on bridged assets (e.g., wETH, wBTC, wDAI, ckUSDT, USDC) and their native tokensfor liquidity onchain. Oracle-based AMMs offer a simple and capital-efficient solution for swapping native tokens against bridged assets (e.g., STRK/ETH, STRK/USDC, SUI/wETH, SUI/USDC, ICP/ckUSDT). Unlike CPMM-AMMs, which are best suited for CEX-unlisted tokens, oracle-based AMMs excel in trading CEX-listed crypto-assets. Dodo and Lifinity are notable examples of successful oracle-based AMMs. However, Dodo's capital efficiency is limited, with a maximum of 10x compared to Uniswap V2, and Lifinity restricts liquidity provision to a strictly selected group of LPs. BrownFi addresses these limitations by introducing an innovative oracle-based AMM with 100x higher capital efficiency than Uniswap V2, flexible market making, a user-friendly interface, and open access for all LPs. BrownFi's mathematical model, inspired by the statistical properties of Limit Order Books, enables these advanced features.

Remember that DEXes cannot replace CEXes, CPMM cannot replace limit order-book. They all live together to provide multiple market making and trading options under various circumstances. Although CPMM-based AMMs (e.g. Uniswap) are a popular solution for onchain liquidity in general, oracle-based AMMs (e.g. BrownFi) play the role of a complementary block in DeFi. Specifically, regarding average users farming onchain liquidity of CEX-listed assets with well-liquid, BrownFi AMM fits better, and is potential to outperform Uniswap V3.

Website, Twitter, Telegram, blog

[1] Daniel Phillips, 2023, The Evolution of Decentralized Exchanges

[2] Thuat Do et al, 2023, Derivable: A Novel Derivatives Pricing Markets based on a Family of Asymptotic Power Curves

[3] Vladislav Goncharov, 2023, Hackernoon, Half of Uniswap v3 Users Lose Money — Here's Why

[4] Álvaro Cartea, et al, University of Oxford - Institute of Quantitative Finance, 2023, Automated Market Makers Designs beyond Constant Functions

[5] Philippe Bergault et al, 2024, Automated market makers: mean-variance analysis of LPs payoffs and design of pricing functions

[6] S. Jiang, J. Chen, F. Li, H. Geng and H. Chi, "DCAMM: Dynamic Curve-Based Automated Market Maker," GLOBECOM 2023 - 2023 IEEE Global Communications Conference, Kuala Lumpur, Malaysia, 2023, pp. 4491-4496, doi: 10.1109/GLOBECOM54140.2023.10437668.

[7] Bhaskar Krishnamachari, Qi Feng, 2021, Eugenio Grippo Viterbi School of Engineering, University of Southern California, Dynamic Automated Market Makers for Decentralized Cryptocurrency Exchange

[8] Daniel Jiwoong Im, Alexander Kondratskiy, Vincent Harvey, Hsuan-Wei Fu, 2024, UAMM: Price-oracle based Automated Market Maker