Nate Hindman: “The early design of Bancor v2 used an external price oracle to manage the dynamic weighting system, which was designed to protect LPs from impermanence losses. However, we soon discovered that this method is vulnerable to preemptive transactions by the oracle-practical It will increase the impermanence loss of LP.”
Interview & Record: Wang Lujie (Winkrypto)
On February 4, 2021, Winkrypto invited Bancor’s Chief Growth Officer Nate Hindman to the WinDeFi community to hold an AMA called “AMM 2.0 DEX Hundred Regiments Battle, How Does Bancor Break Through?” A detailed and in-depth discussion on Bancor’s changes to the AMM model.
The following is an arrangement of AMA text.
Question 1: Why was Bancor designed? What kind of problem is Bancor trying to solve?
Nate Hindman: Now, Bancor v2.1 provides two breakthrough features for token holders who want to earn token revenue by providing liquidity to automatic market makers (AMM). These features have the functions of unilateral exposure and impermanent loss protection, and users can access them on bancor.network.
First, we discuss the key issues of the first-generation AMM, such as Uniswap, Sushiswap, and Balancer, all of which copied the first version of Bancor’s model launched in 2017.
Take a LINK holder as an example, if you want to provide LINK to Uniswap’s LINK/ETH pool, you need to provide both LINK and ETH tokens to the pool. If the price of LINK relative to ETH rises (which will always happen), then there will be an arbitrage opportunity to rebalance the pool of funds.
This will cause impermanent losses to the liquidity provider, because the LINK / ETH pool is selling LINK with rising prices, and at the same time buying ETH with falling prices relative to LINK. This is the reason why many token holders do not participate in the AMM pool: Why should token holders sell their favorite tokens when the price rises? In addition, why should token holders turn the profits of paying fees into impermanent losses?
Bancor v2.1 allows users to use tokens for a long time and maintain 100% exposure while providing liquidity. Provide liquidity with a single token, earn trading fee profits, and there is no risk of impermanent loss.
E.g:
- Provide 1 LINK = 25 USD
- LINK price has doubled (now it is $50)
- Users can withdraw 1 LINK value ($50) + transaction fee profit and liquidity mining rewards
Question 2: So far, has Bancor been developing along the path of the original design?
Nate Hindman: The changes in this area of AMM are amazing. Considering the speed of change, we tend to plan timetables and roadmaps on a quarterly basis. After all, we should be in the very early stages of the AMM field.
The key to Bancor’s success is not marketing hype or various meme marketing named after food, but based on the fact that we created AMM on Ethereum in 2017. Therefore, we know them best, and our designs are usually much ahead of our competitors.
For example, after Bancor started discussing these concepts in early 2020, everyone was talking about unilateral exposure and impermanent loss protection. Since then, Bancor is the only AMM that can actually provide solutions with these features.
Question 3: Is there any important update to Bancor recently?
Nate Hindman: Yes, Bancor v2.1 has been active for about 3 months. During that time, the total value locked (TVL) increased by more than 2200% (now there is approximately $350 million in liquidity and is growing rapidly). The liquidity providers on the platform have obtained high APY, locked their favorite tokens on the protocol, and there is no risk of impermanent loss.
APY not only comes with transaction fees, but also comes from BNT liquidity mining rewards-it was only launched on the front end of bancor.network two weeks ago. Some users only see from the mining rewards that their holdings have 6-digit returns.
At present, the Lp in most mining projects simply sells the mining assets it owns. Unlike them, we found that about 90% of the Lp will re-enter the rewards they get to the platform to get rewards again. Lp can be increased Income and get more transaction fee income and BNT rewards, these income can be regained to pledge. In most AMM reward systems, regaining mining rewards requires selling part of your mining rewards, or matching your rewards with assets such as ETH to increase revenue. However, because Bancor is the only agreement with one-sided liquidity, LPs can increase their returns without selling BNT or using other tokens.
Other updates include our upcoming cross-chain bridge from Ethereum to Polkadot, our implementation of Ethereum Layer-2 on the Arbitrum testnet, and more!
Question 4: Why does Bancor design a unilateral liquidity pool?
Nate Hindman: For most liquidity providers, it is troublesome to provide two tokens to the liquidity pool. They want to hold certain tokens for a long time, do not want to sell some of them to obtain another token, and do not want other tokens to face the risk of loss.
Some AMMs try to solve this problem through “zaps”-split your single token into two halves-you provide 50% tokens and 50% ETH. But this means that 50% of your ETH will remain exposed.
Bancor v2.1 allows you to maintain 100% exposure. This eliminates friction, and we see token holders who have never considered the AMM pool before start joining our pool and love the experience. Just sit down-set it up, then forget it-and earn fees and mining rewards.
Question 5: How does Bancor protect impermanence losses?
Nate Hindman: When users deposit unilateral liquidity into the fund pool, the process of supporting deposits is as follows. (We will take LINK as an example again).
- In order to match the user deposits of unilateral LINK, new BNT is jointly invested in the TKN pool by agreement (for example, depositing $100,000 of LINK into the LINK pool will trigger the release of $100,000 of BNT).
- The BNT invested by the agreement is injected into the pool instead of the external market.
- BNT remains in the pool and earns fees until BNT and its accrued fees are finally destroyed.
- When the LINK provider withdraws its deposit, the BNT invested by the agreement and its related expenses may be burned; or when the BNT holder puts his BNT in the pool to take over the BNT of the agreement.
Therefore, the fees charged by BNT invested by the agreement can make up for any impermanent losses. When the cost earned from the BNT invested in the agreement is greater than the IL cost of the equity, the agreement can cover the IL cost without generating any new BNT. The excess burning costs put deflationary pressure on the token.
If there are not enough tokens in the pool to fully compensate for the impermanent loss in LP LINK tokens, part of the compensation can be paid with equivalent BNT.
E.g:
- Provide 1 LINK = 25 USD
- LINK price has doubled (now it is $50)
- You can withdraw 1 LINK value (45 USD LINK and 5 USD BNT) + transaction fee income and liquidity mining rewards
It is worth noting that the BNT issued by the agreement to support unilateral deposits will not enter the open market as inflation. Instead, the BNT is sent directly to the storage pool of the agreement, usually within the earning fee of the agreement, until it is burned.
Further analysis of these mechanisms and their impact on BNT supply can be found in Bancor’s Protocol Health Report. The results of the report show that the agreement covers IL fees and generates approximately 7 times the fees, creating high profits for BNT holders and liquidity providers. There is an explanation on Twitter.
If someone tells you that the loss of impermanence is not a big deal, because the cost of trading can be made up, please show them this picture.
In the $YFI/ETH pool on Uniswap, the income of LP.
After including the swap cost, LP will have a loss of -40%, and without IL, LP will have a gain of +30%.
Question 6: We all know that impermanence loss is very important to LP. But why is this not a problem for Uniswap?
Nate Hindman: Good question. Uniswap’s APY data is extremely misleading to liquidity providers because they completely ignore the risk of impermanence.
For this reason, I recently made a graph:
Here is another graph:
I think Uniswap, Sushiswap and other companies will soon have to bear the cost of Lp’s impermanence. I think they have good reasons to hide it now.
So, if LPs often lose money, why does Uniswap still have billions of dollars in liquidity? In some cases, this may be profitable for Lp. Sometimes the gain from transaction fees is higher than the impermanent loss, so Lp gains. But many times this is not the case. Many times, liquidity providers wake up one day and see that the price of one of the tokens has risen, but they have lost a lot of money.
More and more Lp are aware of this high risk. More and more Lp recognize Bancor and transfer its liquidity to our fund pool. People just want to put their tokens in a pool, knowing that they will always maintain their value and make money. Bancor supports this approach.
In fact, the fact that sometimes LPs in Uniswap can be profitable is precisely why the Bancor protocol can make money by providing such IL protection services to LPs. Some pools are generating profits for the network (because fees are greater than IL), and some pools are generating losses (when fees are lower than IL). By diversifying the risk of IL among a large number of whitelisted asset pools (currently about 60), the protocol can effectively manage this risk and make the LP experience stress-free.
You can check the impermanence loss in the AMM pool on Uniswap, Sushiswap and Bancor here
Question 7: Can you further explain Bancor’s liquidity mining? Is there anything that needs special attention?
Nate Hindman: Okay. First of all, if you provide unilateral BNT liquidity to a reward pool on bancor.network, then the APY obtained from liquidity mining rewards is expected to be between 100-200%.
If you provide the ERC20 end (ie, unilateral LINK or WBTC), you will still get BNT liquidity mining rewards, and APY is close to 20-50% (depending on the pool).
The incentive to mortgage BNT will be higher, because providing BNT is more helpful to the network, and it reduces the pressure of agreement to jointly invest in BNT to support unilateral deposits.
Regarding our reward plan, the most unique thing is that you can easily get benefits from the double pledge rewards. As mentioned earlier, because Bancor’s pool allows unilateral liquidity, you can re-stake the BNT you get from mining without providing additional tokens or selling BNT. This can earn you more transaction fee rewards and mining rewards, which can still be pledged again.
In most reward programs, the tokens obtained from mining are often sold by LP as soon as possible. In contrast to them, Bancor’s ability to obtain extremely high compound returns by providing liquidity is unique to Bancor.
Question 8: Some people in the community believe that it is meaningless to improve AMM with oracles, because there is a time difference between block time (10 seconds) and exchange transactions (milliseconds). There will definitely be potential arbitrage opportunities. How is Bancor? Regarding this view?
Nate Hindman: Bancor v2.1 does not use oracles. The early design of Bancor v2 used an external price oracle to manage the dynamic weighting system, which was designed to protect LP from impermanent losses. However, we quickly discovered that this method is susceptible to preemptive transactions by oracles—in fact, it will increase the impermanence loss of LP.
After the release of our AMM, other AMMs such as DODO and Cofix have also tried similar oracle-based designs and are now dealing with the same problem. Obviously, this method is flawed.
This is why Bancor v2.1 does not use an oracle to change the weight. The weight is fixed at 50/50, and the impermanent loss insurance in Bancor v2.1 does not require the use of oracles. We think this is a very good design, and while actually protecting users from IL, the protocol benefits obtained prove its effectiveness.
Question 9: If we trade tokens on Bancor and Uniswap at the same time, will there be subtle differences? How does Bancor attract Uniswap investors again?
Nate Hindman: Because the LINK pool has such high liquidity, it can provide cheaper prices for LINK transactions, and we can see from this graph that Bancor’s share of LINK transaction volume is increasing:
We hope this will happen to most tokens. First, LP realized that it was profitable. Then LP increased liquidity and created better prices, and then traders came to Bancor to seek better prices.
Bancor.network was originally built mainly for liquidity providers. Recently, we have invested a lot of effort to improve the trading experience on Bancor.network. We are now building a world-class trading experience to ensure that we get traffic not only from third-party aggregators, but also from traders who directly access the website, because the homepage transaction volume is not just LP transactions, but also token transactions the amount.
Question 10: I have tried to buy BNT on Bancor, but the gas fee may exceed 100 USDT to stop me. Will Bancor try to solve this common problem?
Nate Hindman: Now for every DeFi application, Ethereum fees are a challenge. We will definitely invest a lot of time in optimizing smart contracts to reduce gas costs, and we will also expand Ethereum’s Layer 2 solution.
We have announced Bancor’s real-time application on the Arbitrum testnet. I look forward to the launch of the Arbitrum mainnet, and Bancor will be one of the first applications to be launched!
Question 11: Can you share Bancor’s plans with us in the near future, especially in China?
Nate Hindman: Okay, Bancor’s priorities are:
- In order for the liquidity of the agreement to grow faster, the joint investment ceiling will be increased;
- Release Bancor Vortex, an innovative method to stimulate demand for BNT;
- Continue the liquidity mining plan, and conduct joint liquidity mining and remortgage with third-party projects;
- Expand to L2;
- Expand cross-chain;
- In order to increase the large number of transactions, the top secret pool design will be launched at the end of the first quarter of 2021 and the beginning of the second quarter
- Improve user traffic and token issuance in Bancor through a progressive impermanent loss insurance model
- Release an updated version of bancor.network
We are recruiting people who can translate our materials and applications, more participants in domestic events and hackathons,
For partners who are carrying out more marketing activities in China, please contact us if you have suitable ones!
Question 12: The last question: Do you think BNT is underestimated?
Nate Hindman: It’s up to you to decide! ! As the inventor of Ethereum AMM, we believe that we have the best AMM technology in the world. And we believe that we have the best token economy and value capture function in the AMM field.
After deducting unpaid losses, BNT holders account for approximately 25% of the agreed income. In contrast, UNI (0% of agreement revenue, SUSHI 0.05% of revenue).
If UNI and SUSHI increase the value capture of their token holders, I think LP will quickly leave their platform, because above IL, they earn less. Therefore, I think these projects, which are our main competitors, are in trouble.
Other reference links:
https://trunks-jtcstudio.medium.com/%E5%B8%B8%E8%A7%81%E9%97%AE%E9%A2%98-87162441fe19
https://t.me/bancor_cn
