Scalability tradeoffs for AMMs Explore new opportunities for Uniswap

Automated market makers (AMMs) are a relatively new concept, one largely driven by Uniswap. Most discussions about AMM tradeoffs usually come in two flavors, the first from seasoned traders who can't jump past the inefficiencies created by AMM pool trading, and the second from the crypto-native crowd trying to persuade others Go adopt decentralized financial applications and preach that lazy liquidity will eat the world. As a former algorithmic trader, I certainly subscribe to the first argument and have been tossing and turning on the topic. The answer, though, is naturally somewhere in between, and it's a discussion of the scalability tradeoffs of the core strategy.

AMM VS Order Book

Imagine that you have some expertise in trading, and you decide to start a market-making robot. The goal of a market maker is simple - buy low and sell high, and then earn the difference. It sounds very simple, post a best buy order, and then post a best sell order, your buy order is filled, the price fluctuates, and then someone fills your sell order, you get the price difference. And the reality is more bleak, first of all, you need to consider the fees of market making, you will definitely pay the fee with a smaller transaction volume, and in addition, somehow leaving a large inventory of assets that should be bought and sold, now The value of the inventory has dropped by 10%, causing you to lose any profit you made on the spread. Inventory is the key here, at the center of any market making activity is inventory management, how long do you keep your inventory? To do this requires a combination of powerful modeling and advanced techniques to manage. The so-called market-making robots are not the type of "just open and let go without monitoring". Doing so will put your funds at risk, tailing events will occur, and the cost will be high. The strict requirements of market making are reflected in the trading volume of market makers, and the volume is distributed in a steep power law. There are relatively few largest market makers in centralized cryptocurrency exchanges and traditional finance, and they dominate the market.

DeFi platform dAMM announced the completion of a $2 million token sale round: Jinse Finance reported that institutional DeFi lending platform dAMM Finance has completed a $2 million round of private token sales. Participants in the round include Prismatic, WOO Network, LedgerPrime, Fischer8, Concave, Berachain, and System 9, Inc. [2022/9/27 5:55:17]

Successful market makers have two unique characteristics. First, they have high Sharpe ratios. Metrics such as the Sharpe or Sortino ratios are not usually emphasized in cryptocurrencies, mainly because holding crypto eliminates any Sharpe ratios, no matter how many 10s are captured. But despite underlying model assumptions, strategies with high Sharpe ratios are good because they can exploit almost arbitrary risk thresholds. In traditional finance, a risk infrastructure of the type described in our previous article gives good market makers incredible lines of credit and access to leverage precisely because of the return characteristics of the strategies they execute.

Aptos ecological AMM trading platform LiquidSwap has launched the LayerZero cross-chain bridge UI and opened the test network: On October 14th, Aptos ecological AMM trading platform LiquidSwap has launched the interoperability protocol LayerZero cross-chain bridge UI interface, and the test network is now open , users will soon be able to cross-chain ETH to Aptos.

As previously reported, on October 11, the public chain project Aptos has integrated the cross-chain interoperability protocol LayerZero. The Aptos ecosystem will use LayerZero to unlock cross-chain opportunities. LayerZero will enable the Aptos team to provide a key foundation for the entire Move ecosystem and beyond. Facilities, applications and advanced tools. [2022/10/14 14:27:30]

A second key feature of these strategies is that they are scale-limited. Most of these companies, only manage their own money, because they don't actually gain anything from raising capital, but rather dilute their own ROI. Many people believe that a basic law of quantitative finance is that the Sharpe of a strategy is closely and negatively correlated with its size. It is worth noting that Numerai is a radical attempt to break this law, but so far, this law has been basically true in the capital market. In short, Citadel's Sharpe ratio is an opportunity for Uniswap.

Sudoswap launched a new sudo AMM joint curve "Concentrated XYK Curve", which allows setting a more concentrated price range: On September 16th, the NFT trading protocol sudoswap launched a new sudoAMM joint curve, the Concentrated XYK Curve (Concentrated XYK Curve), which can Let users replicate the full range of Uniswap V2LP, and set a more focused price range. LPs can configure "pooling" parameters, controlling pool depth and slippage. If the "centralized" parameter is set to 0, it is effectively the same as the full-scope Uniswap V2 pool, but for NFTs. LPs can also increase concentration to tighten price bands. [2022/9/16 6:59:58]

That said, the AMM tradeoff is a matter of policy scalability. AMMs will never compete with the dynamism and data dominance of market-making bots, but the benefits of scalability are undeniable. By locking liquidity providers into the exact same strategy, AMMs provide a more level playing field for liquidity providers. The biggest risk a market maker faces is so-called adverse selection, which is when the taker knows something you don't, whether it's a high bid for a coinbase or a protocol hacking attack, etc. A good AMM effectively dilutes the adverse selection of all LPs proportionally.

The AMM protocol Velodrome on Optimism is officially launched, and the airdrop application has started: On June 2, the AMM protocol Velodrome on Optimism launched by the Solidly ecological project veDAO and using Solidly as a template was officially launched, and the token airdrop application has been launched. [2022/6/2 3:57:24]

Uniswap does a good job of this, but LPs have several ways to avoid toxicity. Consider a sophisticated uniswap LP who is modeling price action and at the same time he is an arbitrageur. With the infrastructure in place, they can predict and spot flows that could cause substantial price changes (meaning prices won't "recover" anytime soon) and generate some losses for LPs. In this case, a sophisticated LP can simply remove its liquidity, execute an arb, and re-add it. Some researchers briefly discussed this possibility in the audit, but this does not mean that Uniswap is particularly weak in this dimension (in fact, quite the opposite). The point is, LPs need to be constrained because if some LPs have a material advantage over others, returns for most LPs will be inhibited, and the strategy is limited in its scalability.

The dynamic AMM-based index protocol PowerIndexv2 is officially launched ASSY index has been migrated to this version: DeFi governance aggregation protocol Power Pool officially launched the index protocol PowerIndexv2 based on the concept of dynamic AMM. The previously launched index product ASSY has been migrated to a dynamic AMM fund pool, including The assets are AAVE, SNX, SUSHI and YFI. Unlimited ETFs based on PowerIndex v2 to create automated trading portfolios and ETFs. Its advantages are: 1. Automatic trading strategy and rebalancing based on dynamic AMM; 2. Generate APY from consolidated assets, and the assets in the automatically traded portfolio/ETF can be used in Vault; 3. Meta-governance, if the assets in the pool Has governance functions; 4. Integrate with platforms such as 1inch and Balancer to attract more transaction volume and related fees; 4. No permission required, anyone can define a trading strategy in a basket, start a fund pool and provide attract liquidity. [2021/2/17 17:24:00]

Instead of providing options for makers so that AMMs become more and more like order books, it is better to rely on their own advantages and strictly constrain makers. One way to constrain the maker is to impose a time penalty on LP withdrawals as suggested by Tarun and Monet Supply, which would filter LPs on higher timeframes, and by setting the withdrawal fee to 100% for block 0 deposits, Eliminate LP sandwich attack. A consequence of restricting the maker is that this strategy is literally fungible. Consider a new kind of AMM that allows arbitrary curves to be mapped to each other, perhaps even a hybrid of discrete limit orders.

In this environment, all makers operate under different conditions, so their unique strategies cannot be tokenized, tokenization and maker constraints are almost duplicated, and an AMM with LP out of the game will not be able to provide secure tokenization Expressed. Early research on LP tokens is promising and could become a truly unique DeFi primitive by democratizing Goldman's structured product desk. But in practice, returns are a bit tricky to determine, and have come under fire for always comparing returns to a benchmark (impermanent loss) rather than cash. An ideal LP yield curve has a high amount/current ratio and a high Sharpe ratio. In this case, the AMM cannot limit AUM, it will have to accept new LPs in order to flatten yields to the point where inflows match outflows, resulting in hyperliquidity.

YFI/ETH LP returns denominated in ETH

This trade-off is often discussed deterministically, but in practice the outcome will be probabilistic and path-dependent. There is no doubt that Ethereum’s gas cost and latency semantics give Uniswap a healthy advantage over order book exchanges, which may be the bootstrap mechanism for making AMMs concrete. And newer, highly scalable order book solutions like Serum, with traditional market structures, can offload Uniswap’s trade flow. It goes without saying that AMMs offer universal access to (risk-bearing) transaction fee income, which is unique and only possible under strict policy constraints, after all constraints are liberated.

Tags：

DOGE