Uniswap recently launched the third version (V3) of its acclaimed Decentralized Exchange (DEX). The external response has been mixed, and the debate is mainly about how V3 changes the dynamic uncertainty between active and passive liquidity providers (LP). People’s response has been mixed. Considering the forecast article ” Uniswap’s success opportunity lies in replicating the Sharpe rate of Citadel, the king of hedge funds ” published by Parsec Research last fall , we need to conduct a follow-up analysis.
The design of V3 proposes a feasible attempt to invalidate the trade-off space described in the article, reconciling active and passive LPs.
On the one hand, V3 is a superset of V2. If all LPs are to provide liquidity from [-∞,∞], then your pricing curve is the same as V2. On the other hand, LP can now provide liquidity in a separate “bucket” in a manner similar to a limit order. This upgrade is clearly aimed at improving the taker’s experience. The question naturally becomes: if we are already sacrificing lazy liquidity, why not just build an order book?
The answer is: on the one hand, because of the constraints of the computing environment, on the other hand, it is necessary to challenge the view that inert liquidity must be sacrificed.
Order book v. Uniswap V3
In a standard limit order book (LOB), the trading venue provides a quotation amplitude, which is usually $0.01 in stock trading, and cryptocurrency trading varies from trading venue to trading venue. According to the magnitude of the amplitude, the maker can place orders one by one on any quotation, but all orders must be managed independently.
In Uniswap V3, a set of similar price amplitude services has also been determined, but the order holder can specify a range of price changes in V3, and use spread orders to fully cover this range. In this way, passive strategies and active strategies can be mixed, but what is important is that for some semi-passive strategies, the number of operations can be significantly reduced.
Assuming that LP believes that the price range of ETH/USDC is between 1900-2100 US dollars, the V3 spread position can complete all operations at one time without adjusting bid and inquiry orders one by one and managing all orders, thereby significantly reducing transaction costs and Maintenance costs.
Range Strategy on V3 vs. Order Book
Since V3 covers the entire range of the order based on the spread, when the order crosses a price, it will be proportionally matched with the LP capital. This is completely different from LOB, which is matched with different limit orders according to the first-in first-out (FIFO) method. This type of detail sounds trivial, but I assure you, absolutely not. In every liquid exchange environment, very small execution details can have a big impact on the optimal strategy of algorithmic traders. This is particularly prominent in the context of public chains. In this case, the best strategy may have negative externalities to the entire chain. Since the execution of orders and fees accumulate in proportion to the order holders, it seems that the competition for the execution of orders within a single price will be reduced.
When I thought about the article on the pros and cons of AMM linked earlier, there was a loophole in theory, namely the effect of aggregators. The accumulation of liquidity has a parasitic effect on inert LPs, especially the PMM (Private Market Maker) model adopted by most leading aggregators in highly liquid exchanges. This is parasitic, because active market makers selectively join transactions to fill the flow of non-toxic information without having to face huge and constantly changing market orders as their counterparties. Over time, this will erode the return of the LP, because the active LP can jump in the aggregator order, and the marginal gas fee for the final trader increases.
This made me understand the real importance of V3 in my personal view-a reliable attempt to become the king of decentralized exchanges.
Uniswap tries to match the two parties by opening up options for active market makers and inert LPs, and forces them to participate in transactions in accordance with the same set of rules. This is the beauty of looking back at the proportional mix of orders. By mixing inert/active orders equally, the problems of parasitic aggregator market makers can be eliminated. This design captures passive and active LPs into the same order queue. The result is that active market makers assume greater price/inventory risk, but they can in exchange for a significant reduction in transaction costs. Passive market making will still charge fees and benefit from the increase in taker traffic caused by the increase in active traffic.
But this only works when Uniswap routes directly to a large number of transactions.
In the long run, direct routing is difficult to grasp, but it can be achieved in two ways:
- The latency/infrastructure cost of fighting market makers one by one exceeds the price advantage of routing outside of Uniswap. This happens only when Uniswap is extremely liquid and the congestion time is less than 15 seconds (for example, rollup). In addition, most aggregator routing schemes involve a marginal increase in user experience and gas costs (although this is mainly a cost related to approvals, as accounts are abstracted, the approval flow will be deprecated and this cost will gradually decrease).
- The use of UNI tokens encourages direct routing. Uniswap governance sits on UNI tokens worth about 5 billion U.S. dollars, which can be used to incentivize direct transaction flow, reward taking orders and routing interactive interfaces. It is not sustainable when t→∞, but it is worth it if the first can be made a reality.
There are still many problems with this design mechanism. Preemptive deals/sandwich deals will still be a persistent threat? For the maker and taker, the order book exchange is actually an environment with great advantages? Nevertheless, the design is ambitious and changes the DEX dimension again.