M: Actually, I really want to talk about what you learned during the evolution from v2 to v3, and let us briefly talk about some key features of V2 and V3 and what changes have been made.
L： This is a very good question. What we learned from v2 is that liquidity is the first principle of financial products, which means we should focus on how to improve liquidity, including high capital efficiency, which leads to low slippage for traders, and lower cost to provide liquidity.
How did we discover this? V2 is a hybrid model with both order book and AMM. To achieve complete decentralisation, V2 is driven by AMM with the constant product pricing formula, that is, the formula x times y equals k. I guess many people are pretty familiar with it. Its problem is the low capital efficiency and high slippage. This actually doesn’t work for perpetual swaps because the leverage again amplified the low capital efficiency. So we meanwhile run the order book model. The idea here is that AMM can do the role of pricing and decentralisation and then the order book can provide real liquidity. Thus we found the problem — The cost of liquidity on the order book is too high.
For example, the liquidity around spread on centralised exchanges is provided by the market maker, which requires two parts of costs from the team. On the one hand, they need funds to place orders. This problem will be more serious in dex because in cefi, assets can be faked, but this is impossible for DEX, because all assets are on chain, funds cannot be faked. So this is the first part.
The second part is the market making risks. Since the exchanges rely on the market makers to provide liquidity so that the users, as takers, can enjoy better user experiences. The market makers have some risks by doing that.
Looking back, we found that V2 has two contradictions. The first one is that the liquidity cost for the order book is very high, however that for AMM is relatively low. Anyone can add liquidity for AMM, however only professional market makers can provide liquidity by borrowing team’s funds. The second contradiction is the cost of risks. AMM has much lower risk cost compared to the order book model. On AMM there might be impermanent losses. As a market maker, the risk is unavoidable, but just a matter of the size of it and whether the return can cover the risk cost.
The high cost of the liquidity for the order book leads to the fact that it’s hard to provide the liquidity rapidly.What is worse, for some long-tail assets, sometimes it’s even impossible to find the market makers. We all know that one of the key competencies of FTX is to be able to provide liquidity to long tail markets so that FTX can grow fast in its early stage. It is difficult for other exchanges to find market makers to provide liquidity for long-tail assets. To conclude, the issue of liquidity is the most important thing we learned from v2.
After we realize those issues, we kind of suspend the V2. Meanwhile, we also further validate the market’s needs for decentralized perpetuals. Since we raised the transaction fee to 1%, which is way much higher than centralized exchanges, we still had several traders back then.
Therefore, in V3, we are determined to solve the liquidity problem.
The first change is the AMM-only model. We redesigned the AMM pricing formula. For perpetual swaps, The price must be soft pegged to the index price and cannot fluctuate far from the index. The straightforward idea is to make full use of the existing index and use mathematical formulas near the spread to adjust the price. There are several ways to adjust the price. The first one is to let market makers always buy low and sell high. The second is to avoid market maker risks. When market makers hold net positions, we will have those positions sold with some discount, so as to encourage arbitrageurs to take those positions and reduce the market maker risks.
So that’s basically how we have a new AMM with much better capital efficiency. We have done some tests to prove that the capital efficiency is 1000 times higher than V2. We’ve referred to the market making strategies of CEX. Besdes, AMM allows anyone to provide liquidity instead of professional market makers. For the small-cap assets , they can create a perpetual market permissionlessly on our V3. By the way, we just saw the design of Uniswap V3. We have a similar approach here. The liquidity is concentrated in a certain range.
The second key feature of V3 is permissionless. Traditional way of adding market with permission requires lots of work from team. And usually team has a hard decision to chose what market to add. By making it permissionless, we the the power to the communtiy to do some intersting market. For example, if a project wants to create a market with their token as collateral, they are pretty easy to do it. No need to communicate with us. So being permissionless is the second innovation for our V3.