by Akshay Ramasubramanian
This analysis examines liquidity providers (LPs) on SushiSwap and what they do with the farmed SUSHI tokens.
The top destination for farmed SUSHI (29.7%) is SushiBar. SushiSwap's own SUSHI-WETH pool, as well as Cream and Uniswap, are the other key destinations for farmed SUSHI.
The Nansen team also noticed some interesting exploits on the SushiBar contract, which we'll dive into at the end of this article.
SushiSwap is an automated market making (AMM) decentralized exchange (DEX) currently on the Ethereum blockchain.
In this article we will dig into SushiSwap, and focus on three things.
- Liquidity Provider Behavior: Identifying the different LPs, and look at what they do with their SUSHI tokens.
- LP behavior on different protocols, specifically Uniswap, Balancer, and Cream.
- Deep-dive on something interesting we noticed going on with SushiBar from the data gathered from the previous steps.
Liquidity Provider Behavior
Total Number of SushiSwap LP's
At the time of writing this report there are about 10,176 unique addresses providing liquidity to SushiSwap. These are the unique Ethereum addresses that interact with the Sushi Liquidity Pools and lock two tokens in order to receive Sushi Liquidity Pool Tokens (SLP) that symbolize the liquidity provided to the protocol. Note these addresses may not necessarily be staking their SLP tokens.
Number of Pools per LP
If we break down the previous data by looking at the number of SushiSwap pools each of the LP’s provide liquidity to, we can plot a simple chart that summarizes this data.
A majority of LPs (66.2%) provide liquidity to only 1 pool. Only a small percentage (4.2%) provide liquidity to 5 pools or more as seen in Chart-1.
Nansen labels can help us identify who the notable addresses that provide liquidity are. These labels also include ENS domains that the users have registered on their addresses. Digging into the labels of each of the LPs in Chart-2 below we notice most labels have
Dex Trader. All this means is that these addresses trade on decentralized exchanges (Dex).
We also notice that SBF (17 pools) and boringcrypto.eth (15 pools) are a few labels that stand out and are known by the community. Let’s remember this as we continue our analysis.
At this point it's fair to assume that these addresses will show up again when we look at the rewards per LP or total liquidity locked per LP since they seem to be interacting with multiple pools.
How many LPs account for X% of total liquidity?
In this section we will look at the gross liquidity provided by the LP. We will focus on all the transfers (ETH and tokens) from LP addresses to the SushiSwap pools to calculate the gross liquidity for each LP. Note if a LP provides liquidity of 10k USD to SUSHI-WETH pool and withdraws a few days later and then repeats the same process, this will be counted as gross 20k liquidity.
From Chart-3 we can see the labels that stand out are Alameda and also Three Arrows Capital both providing over 35 million USD in liquidity over all time. Note that this is not the same SBF MultiSigner address from Alameda that we noted down in the previous section. This probably means Alameda seems to use multiple wallets to interact with SushiSwap.
Most of these addresses share the
Heavy Dex Trader label. These are just addresses who have done over 500+ DEX trades.
To get a better understanding of the distribution we can plot a histogram. We split the total liquidity into groups to make things easier to picture. Chart-4 shows the total liquidity range in USD and the percentage of LP’s that fall into this category.
From this it's clear that the majority (70.9%) of LP’s supply between 100-100,000 USD in liquidity over all time. There is quite a large percentage (10.2%) that supply between 100,000- 1,000,000 USD in liquidity. Finally there is only about (2.6%) that have supplied over 1 million USD in liquidity.
To summarize we can say that the majority of SushiSwap LP’s seem to be locking up between 100-100,000 USD in tokens to provide liquidity in the SushiSwap pools since the migration.
How many LPs provide liquidity elsewhere?
As it can be seen a large percentage (40%+) of SushiSwap LPs’ provide liquidity on Uniswap too. Surprisingly a big chunk (24.2%) also provides liquidity on Balancer. A similarly sized chunk (22.5%) only provides liquidity to only SushiSwap.
This is interesting as one might have expected that most of the SushiSwap LPs have provided liquidity in other venues and got interested in moving due to the higher APY, but based on the data it seems like there is a decent chunk of users (22.5%) who only provide liquidity to only SushiSwap.
Note that since in this section we only focus on Cream, Balancer, Uniswap and SushiSwap, these users may be interacting with some other protocols that we have not been looking out for.
What's the distribution of SUSHI claims per address?
About 11.9% of the farmers have farmed 1,000-10,000 Sushi. Only a handful (2.9%) have farmed more than 10,000 SUSHI. Majority (68.5%) have farmed between 10-1,000 SUSHI.
Let’s take advantage of the Nansen labels to dig into these Sushi farmer addresses to figure out if we can spot anything interesting.
In the table above we can see Alameda is one of the top farmers. Even though SBF MultiSigner wallet and boringcrypto.eth seemed to be providing liquidity to the most pools, the same addresses don't show up in the top 35 list above. (SBF is the CEO of Alameda and may just be using multiple different wallets since Nansen labels seem to have picked up multiple labeled Alameda addresses)
Something interesting to note is that there are a few addresses that seem to be making a lot of rewards from SushiBar farming. You might expect the all addresses to primarily be using MasterChef to farm SUSHI tokens, however from the table above it can be seen that this is not true. Several farmers are making significantly more SUSHI from SushiBar compared to MasterChef.
MasterChef involves locking SLP tokens (liquidity tokens) to mine Sushi, while SushiBar involves locking SUSHI. For a more detailed explanation click here.
Note that if a user hasn't claimed SUSHI from SushiBar it will not show up in the analysis.
We will dig into this in more detail in the last section to discover what exactly is going on here.
Where does SUSHI from Liquidity Provider Farming go?
In order to determine this we can take advantage of Nansen’s labels once again. For a big picture we can try to group all our entity labels together to try to understand where most of the SUSHI farmed goes.
Note that these amounts may be higher than the SushiSupply since it's the gross aggregated SUSHI received for each of the entities below. Also grouping the labels doesn’t tell us the exact action the user is taking. Since they may be either locking their Sushi on SushiSwap or swapping it. A more detailed breakdown will be provided below so we can differentiate between these actions.
The top known destinations for farmed SUSHI are SushiSwap (23.5%), Cream (18%), and Uniswap (16.4%).
If we look at centralized exchanges, Binance is the first one in the list, but it's only a 3.6% chunk, compared to the DeFi protocols (SushiSwap, Uniswap, and Cream) which have magnitudes of higher SUSHI received.
We could look into the registered ENS domains for the ENS-labeled Entities to see if anything interesting shows up. This is displayed in the table below.
Some known names from the Sushi community such as zippoxer.eth can be seen here.
In order to find out more about the transfers into SushiSwap specifically the actions users take, let's dig into more detail and look at the top addresses and labels instead of grouping them by entity.
It can be seen from the pie chart below that the majority (29.7%) of the Sushi goes to SushiBar. The second biggest receiver is the SUSHI-WETH SushiSwap pool (22.8%). After digging into more detail for this section, it seems like all these transactions to the SUSHI-WETH pool are actually mint events which means they are actually LP’s locking SUSHI + WETH in the liquidity pool.
The third largest is Cream with 18.1%. We will dig into this in more detail in the next section.
We can summarize the information above and come to the conclusion that a lot of the SUSHI liquidity followed this path:
- Lock SUSHI in SushiBar
- Add liquidity to the SUSHI-WETH pool
- Send LP tokens to Cream
The SushiBar Mystery
As noted earlier, there are a few addresses that seem to be farming most of their SUSHI from SushiBar and not MasterChef.
The Nansen team started digging into these addresses and noticed a bunch of really profitable transactions. Let’s have a look at one of the top SushiBar farmer addresses in more detail.
This externally owned account (EOA, i.e. non-contract) seems to be entering and exiting SushiBar and making 10+ ETH of profits in less than 12 minutes of entering the SushiBar. A simple breakdown of what they're doing is summarized below.
- Borrow Sushi from Cream
- Stake Sushi -> xSushi (enter)
- 12 mins later
- xSushi -> Sushi (leave)
- Repay Cream Loan
- Convert profited Sushi to WETH
Let’s dig into this in more detail with token amounts.
Nansen users can inspect this address in our Wallet Profiler For Token dashboard.
Alternatively, you can follow these transactions via Etherscan yourself by clicking on the timestamps below.
- Oct-06-2020 11:30:51 PM +UTC: Interacts with CreamFinance contract to call function 'borrow', gets 5,700,00 SUSHI ()
- Oct-06-2020 11:31:32 PM +UTC: Stakes 5,700,000 SUSHI for 5,479,364.01 XSUSHI by calling enter function in SushiBar Contract (41 seconds later)
- Oct-06-2020 11:38:41 PM +UTC: Leaves SushiBar by calling the leave function in SushiBar. Converts 5,479,364.01 XSUSHI’s to 5,705,876.79 SUSHI (7 minutes 7 seconds later) (5876.79 SUSHI profit)
- Oct-06-2020 11:39:37 PM +UTC: Sushi is sent back to CreamFinance 5,700,097.25 SUSHI (an additional 97.25 SUSHI) ( 56 seconds later) function repayBorrow (Net profit of Sushi still at 5779.54 SUSHI)
- Oct-06-2020 11:42:49 PM +UTC: Convert all profits to WETH using SushiSwap (Swap 5779.54 SUSHI for 10.815 WETH)
The wallet then repeats a similar action on the next day. The first transaction is linked here. If we look carefully, it's at a slightly different time Oct-07-2020 10:46:51 PM +UTC to Oct-07-2020 10:57:17 PM +UTC .
Initially we thought perhaps there was a spike in volume during those times, but digging into the volumes between those times in more detail reveals that there was only about 553 ETH equivalent traded in the time interval this trader borrowed Sushi, staked it and converted back to Sushi.
This would account to about 5553.53 * 30bps * 5bps = 0.000083295 ETH in fees in those 7 minutes. This is based on the rewards highlighted in Sushi docs.
This means that this EOA actually seems to be taking advantage of SushiBar and reaping huge rewards in return. If we look at the Sushi Bar contract source code it looks like anyone can call the convert function and distribute the rewards.
Let’s try to get a bigger picture of the sushi bar to see the realized gains of users.
From the above charts it can be seen that jokap.eth seems to have the most rewards from the SushiBar. This is the same address taking advantage of the Cream loan to get the most rewards from SushiBar. Let's split this into groups to get a better understanding of the distribution.
Only a small percentage of addresses (0.18%) makes more than 10,000 SUSHI from SushiBar, majority make 0-100 SUSHI (94.8%).
After looking at the above exploit (if we can call it that), it seems like it might be worth rewriting the SushiBar logic to take into account the lock duration of SUSHI tokens. It would make more sense to reward users who have locked their SUSHI for a long time frame (more blocks) at a much higher rate vs users who lock their SUSHI for fewer blocks. This would make it harder for users to take advantage of the system as seen in the above example.
This analysis was funded by SushiSwap with the purpose of better understanding liquidity providers. The Nansen team had full editorial freedom in publishing it.
Contact the Nansen Research Desk if you'd like us to perform an analysis.