Uniswap v3: Towards capital efficiency or magnifying LP losses?

Thanks to the original author Momir Amidzic (IOSG) and co-author Danning Sui (0x Labs) for supporting the article!

Uniswap v1 and v2 implement a simple unified XYK pricing curve. When there is a large transaction, liquidity can be guaranteed at any price range with low asset utilization and exponential price growth. Therefore, the only way to reduce price slippage is to increase the K value, which is achieved by attracting more funds into the capital pool, that is, increasing the total lock-up (TVL).

V3 is much less dependent on TVL as it allows liquidity providers to open multiple positions within a specified price range. Thus, the concept of "unlimited" liquidity is replaced by the concept of concentrated liquidity around the market price.

Source: Uniswap v3 Whitepaper

The method of pooling liquidity is not new in DeFi. Curve and DODO have been practiced for some time from different angles. However, Curve only focuses on stablecoins, while DODO uses price oracles to pool liquidity around market prices.

Both Curve and DODO were designed around the provision of passive liquidity, V3 nevertheless achieves capital efficiency by relying on the concept of rational liquidity providers who "can pool their liquidity around current prices narrow areas and keep its liquidity active-zce by adding or removing tokens as prices change, thereby reducing its cost of capital."

UNI breaks through the $37 mark with an intraday increase of 8.36%: According to data from Huobi Global, UNI rose in the short term and broke through the $37 mark. It is now reported at $37.0055, with an intraday increase of 8.36%. The market fluctuates greatly, so please do a good job in risk control. [2021/4/15 20:21:46]

Conceptually, this makes sense; but in practice?

To see how v3 performs, it is natural to first check whether it actually reflects capital efficiency improvements. One way to measure increased efficiency is to look at the velocity of capital turnover in v3 relative to v2. Therefore, we observe the daily TVL turnover rate and compare it with v2 and Sushiswap.

As shown in the figure above, the TVL turnover rate of v3 is much faster than that of v2 or Sushiswap. For example, the market crash on May 19 during the peak period, the TVL of 1 USD provided to v3 can be converted into a daily trading volume of more than 1.7 USD. V3 is a big improvement over V2. Under the same circumstances, V2 only generates about $0.2 of transaction volume per day.

Capital efficiency is indeed a hallmark of the v3!

Another aspect is whether v3 offers a better price than v2. Usually when using v3, we will get the following information:

UNI breaks through the $31 mark with an intraday increase of 4.41%: Huobi Global data shows that UNI has risen in the short term, breaking through the $31 mark, and is now reported at $31.0001, with an intraday increase of 4.41%. The market fluctuates greatly, so please do a good job in risk control. [2021/4/3 19:42:39]

One way to compare is to examine the volume allocated by DEX aggregators for v2 vs v3. For this, we checked two of the largest DEX aggregators, Matcha and 1inch. Both aggregators offer the best prices to end users, so they send most of their volume to the most competitive venues.

In general, as shown in the graph below, we observed a trend where aggregators allocated most of their volume to the latest Uniswap release, implying better pricing.

Source: https://duneanalytics.com/queries/49999/98616

UNI breaks through the $20 mark with an intraday increase of 8.75%: Huobi Global data shows that UNI rose in the short term and broke through the $20 mark. It is now reported at $20.0002, with an intraday increase of 8.75%. The market fluctuates greatly, so please do a good job in risk control. [2021/2/2 18:39:13]

Source: https://duneanalytics.com/queries/50020/98653

There are two ways for an aggregator to bring in on-chain liquidity - via a bridge contract, or via "VIP" routing directly into a Uniswap pool. The latter is optimized to be more gas-efficient than Uniswap's own routing. Therefore, taking the gas fee into account will adjust to a better price. Currently, neither Matcha nor 1inch has enabled VIP routing. This shows that the liquidity of Uniswap V3 can be more competitive than the current data shows. the

After discussing the advantages of the recent upgrade, we also need to examine the potential disadvantages of the new design.

Here is a reminder to readers that one of the problems with v2 is the impermanent loss faced by liquidity providers. Assuming that price discovery occurs primarily on centralized exchanges, any price discrepancy presents an arbitrage opportunity: buy undervalued tokens from the pool or sell overvalued tokens to the pool.

Uniswap’s current lock-up volume has reached US$1.1 billion, an increase of 115% in 24 hours: According to OKLink’s data, the Uniswap lock-up volume has rebounded significantly in the past 24 hours. The current lock-up volume is US$1.1 billion, an increase of 115% from the previous day. Sushiswap lock-up volume is about 1.09 billion US dollars. Affected by the previous Sushi migration, the amount locked in Uniswaps fell to $510 million yesterday. [2020/9/11]

V3 does not solve the problem of impermanent loss, but the size of the loss can be determined by the behavior of LP. That is to say, in V2, LP is relatively static with respect to arbitrage, while in V3, both LP and arbitrageurs have pricing power. This creates an interesting dynamic between the two, with roughly two possible scenarios.

In the first case, LPs can limit arbitrageurs. This requires mature LPs to constantly adjust their price ranges and correctly map market price changes before arbitrageurs take action, thereby protecting LPs from arbitrage in a highly volatile market.

The second situation is aimed at less mature LPs, which is beneficial to arbitrageurs. That said, a narrow price range means greater depth of liquidity and a higher risk of loss in a volatile market environment.

Suppose we provide the liquidity of ETH, as shown in the figure below. Our capital does not become active-zce until the ETH price exceeds $2817.5. Assuming that the price of ETH eventually rises above $3138.8, LP's position will consist entirely of DAI, and LP's exposure to further rises in ETH is zero. After that, DAI's liquidity will be inactive until the ETH price falls back into range. Assuming ETH falls back into the range at some point and continues to fall below $2817.5, the LP position will consist entirely of ETH.

Voice | UNI founder Sunny: UNI built the UNI main chain and Phoenix cross-chain system to solve cross-chain asset transaction problems: On December 18, 499Block Korea Award Night was held at Merry Garden in Gangnam District, Seoul, South Korea. UNI founder Sunny was releasing It was stated at the meeting that UNI built the UNI main chain and the Phoenix cross-chain system to provide more personalization for parachains, realize "cross-chain flash swaps" such as cross-chain atomic swaps, and solve the problem of cross-chain asset transactions. Secondly, the Uni ecosystem improves the cross-chain interface service system Phoenix has a variety of functional modules, which facilitate third parties to achieve cross-chain consensus through SDK and API, call the main chain service, and reduce the development difficulty of multi-chain DAPP. For the first time, UNI introduced SGX hardware homomorphic encryption technology into cross-chain, providing a smooth experience of cross-chain switching scenarios, solving the pain point of inconvenient multi-chain management and improving asset cross-chain liquidity. [2019/12/20]

Therefore, in the case of an ETH bull market, LPs lose exposure to rising prices, and in the case of a bear market, LPs get 100% exposure to downside risks. Assuming a price lag between centralized exchanges and Uniswap, arbitrageurs will crush LPs.

Although it is too early to judge, based on experience, we can see how much trading volume in V3 and V2 came from the top arbitrage bots, and we assume that more bot activity means greater LP losses.

As you can see in the table below, the largest bot accounted for 15.5% of the total Uniswap v3 volume since inception! This address alone has generated $3 billion in volume since v3 launched. In addition, the top 5 arbitrage bots accounted for about 22% of the total trading volume, much higher than the level of V2, and the top 5 arbitrage bots accounted for about 11.2% of the total trading volume during the same period.

Source: https://duneanalytics.com/queries/51415/101669

Source: https://duneanalytics.com/queries/51461/101708

This also shows that LP suffered a significant loss in the early days of v3. But why such a big difference relative to v2?

While the flattened XYK pricing curve implemented by v2 is not the most capital efficient solution, it still provides some level of protection for LPs as slippage increases exponentially. Therefore, even if v2 tokens are mispriced, there is no possibility of an outflow of underpriced tokens from the pool. On the other hand, unless v3 LPs aggressively adjust their price ranges, they risk losing exposure to low-priced assets entirely, or having positions made up entirely of high-priced assets, becoming the purse of arbitrageurs.

Trading volume ratio of the top 5 robots in terms of activity; Source: https://duneanalytics.com/queries/51508/101777

We expect arbitrage participation in V3 to gradually decrease over time for the following reasons:

LP can better grasp risks after learning from the past

Further specialization of liquidity provision, with more LPs utilizing dedicated vaults for active-zce management (https://alpha.charm.fi/).

Layer 2 deployment and a low gas fee environment support a more active-zce LP strategy

Potential MEV-proof strategy can benefit traders, reduce slippage, and reduce gas loss

Uniswap V3 is off to a good start. Although there are areas for continuous improvement, we are still optimistic about the future development of V3. Eventually we will see more professional market makers become liquidity providers, crowding out arbitrageurs with better pricing. And passive liquidity will be a thing of the past, offered only through specialized vaults. Overcoming the current drawbacks requires an ecosystem layer of liquidity yield management, including dApps like Alchemist, Charm, and Visor, which will support DeFi products to develop to a higher level.

Incubating projects for ambitious teams around the world and looking for the most potential staking projects.

Building a new world，the bridge of the time.

Tags：

USDT