Clipper Exchange: A Case Study
Disclaimer: I have ~$400 deposited in Clipper's Core Pool (to test it out). Not enough to sway my opinions and certainly not enough to make me a stakeholder in any capacity.
Introduction
What is Clipper?
(Feel free to skip to 'How does Clipper work' since the following paragraphs were essentially taken from their website.)
Powered by Shipyard software, Clipper is a decentralized exchange (DEX) built for self-made traders, not hedge funds or whales. Clipper is designed to have the lowest per-transaction costs for smaller trades on the most popular crypto assets. This is accomplished through a novel architecture and a series of design tradeoffs that sacrifice price competitiveness on large trades for better prices for smaller trades.
Clipper is named after a type of mid-19th-century merchant sailing ship used for global trade. Clipper ships were designed for speed, partially by trading off cargo capacity.
How does Clipper work?
Unlike other DEX 'whales', Clipper is designed to provide the best fees, slippage, and gas to 'small-time' traders and orders, with their unique formula market maker (FMM). They claim their FMM is different from AMMs and PMMs so let's take a look at all three.
AMM: A Primer
Automated market makers, or AMMs, are decentralized exchanges that automate trades in various liquidity pools. One simple example is a single liquidity pool with two assets where the number of two assets should produce an invariable through adding (constant sum), multiplication (constant product), and/or various other operations.
Say we have an ETH/USDT constant product liquidity pool with 2 ETHs and 3,000 USDTs and trader A wishes to trade 1,000 USDTs for ETHs. The constant product should be 6,000 (2 x 3000). With A's 1,000 USDTs, this pool will have 4,000 USDTs and the number of ETHs should be reduced to 1.5 (6000/4000). Thus, trader A will get 0.5 ETH (2-1.5) for his 1,000 USDTs. All AMMs will charge a trading fee - by introducing a spread - and these fees will be split, most of the time pro rata, between all who provided liquidity for their pools.
You can see that AMMs are completely automated, as its name suggests, and completely path independent if we take trading fees out of the equation. It is not without its drawbacks, however, as the asset prices are prone to fluctuations from large deposits and yield will decrease over time as the pool grows larger.
AMMs are also prone to arbitrage flows when the market fluctuates. Many AMMs embrace this mechanism as part of the free market - it does tend to nudge asset prices towards their fair prices (market prices, that is). But this lies at the root of impermanent loss. There is no silver bullet to impermanent losses for constant product market makers and their variants so most AMMs try to mitigate this with trading fees which, most of the time, feels like a bandaid on a gaping wound.
PMM: A Primer (for me)
Contrasting to an AMM, a proactive market maker (PMM) takes a more hands-on approach to determine the asset's price by introducing a factor that is determined by market actions, keeping the liquidity more concentrated around the price it's more likely to be.
PMM was first introduced by DODO DEX. Their PMM algorithm is based on the trading patterns of order-book-based exchanges, that is, most CEXes by introducing a 'mark price' - pulled from oracles or set by market makers - and a 'slope' that can be adjusted to any value between 0 and 1. By tuning these two parametres, you can turn your PMM into an order-book-like market (when you make a big purchase, you would have to purchase a series of small orders from the cheapest available to the most expensive needed, others will also soon place orders at a lower price to fill the spread) or even an AMM.
Other PMMs include Amber, who added another layer of liquidity pool (backstop pool) as a buffer to mitigate fluctuation impacts and impermanent losses. It should be noted that although Amber claims in their whitepaper that they have 'eliminated' impermanent loss, your assets may still be exposed to market risks.
How does Clipper's FMM differentiate from these two?
Although Clipper claims that their FMM is 'Not an AMM, not a PMM', down to its core, their FMM still takes asset prices from oracles and adjusts their pools' behaviors with customizable parameters.
Clipper claims that impermanent loss can be minimized (and eliminated with their DRP, which we will get to later) with their integration of off-chain oracles (like DODO) to minimize arbitrage flows. They also try to make it less profitable for arbitrageurs with ginormous slippage - made possible by their small liquidity pools. This (at least to me) is the major difference between Clipper and other DEXes - it is specially designed for individual (small-time) traders with only a handful of assets or in their terms, trades below $10K.
They developed their formula as a more sophisticated version of invariant-based AMM formulas - by introducing a variant k. You can set k to 0 to have a constant price swapper, or 1 to have a CPMM, or anywhere between 0 and 1 to fine-tune its behaviour.
This approach sure has its perks but does come with certain risks since you (they) may set it to a value that seemed preferable or plausible at the moment but later proved somewhat inappropriate as it's only been out for less than 2 years (though it has gone through the bullest bull and one of the bearest bear). I do not know whether they could change the k value or have done so, but it can be seen as a lack of knowledge and confidence in their formula (at least to me).
Most of the time, Clipper would set their k value at a point (closer to 0, according to my limited knowledge of differential equations) where slippage is unevenly distributed - smaller for small trades but larger for large trades. This will, to some extent, make it less profitable for arbitrage actions and mitigate impermanent loss. They also charge less trading fees but manage to keep their fee seductive enough for small LPs by limiting the size of their liquidity pools.
Clipper also has its own ' Core Pools', which are similar to Amber's backstop pools - a buffer pool consisting of stablecoins, BTC/ETHs, and GLMR/MOVR/MATICs, depending on your network, while their liquidity pools in DeFi's common sense are aptly called 'Clipper Coves'. Think of the Core Pools as a slightly diversified tricrypto where you deposit any amount of any major coin in the portfolio and get exposure to all the assets but with a $10,000 limit (can be raised to $100,000).
Their Core Pools will automatically adjust the ratio of each asset based on market conditions - the 'Diversified Rebalancing Portfolio' (DRP). This rebalancing act is essentially buying low and selling high and a considerable portion of it can be done by traders swapping tokens. In theory, even without a single trader, the rebalancing mechanism should still be able to generate yield on its own. This is how Clipper could 'eliminate impermanent loss' - proactively adjusting the portfolio instead of waiting for the market to do it. A bit like PMM, yeah?
How does Clipper swap?
Let's do a simple comparison between Clipper and Uniswap, Sushiswap, and Curve. These numbers are taken approximately at the same time at 4 websites, both on Ethereum Mainnet and Arbitrum One. I did not execute these trades, however, as I didn't have that much ETH lying in my wallet for gas fees (think of this as a subtle call to potential sponsors) so I just typed down a bunch of numbers. Also, these numbers will fluctuate in the future so do your own comparison when you're looking for the best place to swap tokens.
| Clipper | Uniswap | Sushiswap | Curve | |
|---|---|---|---|---|
| 100USDT -> USDC (ETH) | 99.93 | 99.9908 | 99.4632 | 99.9957 |
| 100USDT -> ETH (ETH) | 0.063671 | 0.06366 | 0.063492 | 0.063660 |
| 100USDT -> USDC (ARB) | 99.94 | 100.001 | 100.00265 | 100.000084 |
| 100USDT -> ETH (ARB) | 0.063689 | 0.06368 | 0.06354 | 0.06367 |
Note that I did not take fees, slippage, and gas into account because I didn't execute these trades so take every number with a grain of salt. This also shows that Clipper does not always offer the best price but that may be compensated by their lower fees, slippage, and gas - in theory, if someone would conduct extensive tests across different platforms. I also didn't try to swap small-time tokens because there are just too many of them.
Core Pools & Clipper Coves
A humble farmer may examine all the DeFi protocols in the world with as much scrutiny as possible but in the end, what persuades these tokens out of his/her pocket is the ability to provide consistent APR (and safety, but I'm no safety expert so we'll talk about that later in other articles, hopefully).
As mentioned above, Clipper's Core Pools would rebalance their portfolio and generate returns in the process. We should have already covered most of what you need to know about it so let's dive into some analytics and see how they've performed and grown.
These data are extracted from DefiLlama and Clipper's website. Feel free to check them for yourselves for more information and future updates.
At the time of writing, Clipper has a TVL of $10.02m, only a fraction compared to established protocols like Uniswap or Sushiswap. Most of their TVLs are concentrated on the Ethereum Mainnet with WBTC and WETH taking up a little bit less than 1/3, respectively, followed by USDC, DAI, and USDT. These 5 tokens make for 98% of all assets, a robust combination that could in theory withstand most fluctuations.
Over the last week, Clipper claims a 4% per year profit yield and a 15.6% per year comparable APY - taking avoided impermanent loss as profit. This number has been marginally higher than Curve's tricrypto pool (and other stablecoin/major coin-based pools) but is on a bumpier ride recently. Considering its rebalancing actions, we should be able to see higher returns again when the market starts gaining traction again.
Clipper's Coves aren't tracked on DefiLlama possibly because of their entirely different mechanism. Although Clipper is designed for small-time traders and their coves accept any asset at any amount, their trading volumes may seem too small (even for me): $2,000+ on Polygon, $200+ on Optimism, single figure on Arbitrum and Moonbeam, and unavailable on the Mainnet. Even if they've rooted impermanent loss, I suspect, these trades are far from attractive for yield farmers.
Other things
Technically, Clipper doesn't issue their own token, but their ClipperLP represents your share in their Core Pool.
Clipper is governed by AdmiralDAO but it is not open to everyone and presumably is a vehicle for their founder, Shipyard software.
Clipper may show some potential since most would say we're still not much higher than rock bottom and as the market starts seeing some sunshine, so will Clipper's Core Pool.
This is not financial advice. Please always do your own research before investing in anything. I'm only doing this for fun because I can always find fragments of time scattered across the day so I decided to do something with it.
I only got a C in calculus so I may not be able to comprehend their whitepaper that accurately. Read it for yourself if you're curious about their FMM.
Although I have written Python a lot, I do not know much about smart contracts and haven't read their code. Clipper was not listed on Rugdoc but they have put their audit report out for everyone to see. Check the link at the end of this article if you wish to take a closer look.
Closing thoughts
This section has nothing (much) to do with Clipper so feel free to skip it.
Is impermanent loss a bad thing?
It is, theoretically, since it does prevent you from earning more. But it also prevents you from losing more. It's common sense in finance, both traditional and decentralised, that most of the time, less risk translates into less profit. By putting your assets into a pool (and locking them), you are essentially purchasing a fund consisting of several stocks. You would theoretically earn more if you held onto something through a major bull, but bull markets in crypto were never simple bulls but bulls and occasional bears. If you could withstand the emotional stress when seeing an abrupt bear among a seemingly long bull, could you overcome your fear and hold onto your assets as if they were locked? Probably, but there's no guarantee.
So next time when you're thinking about impermanent loss, think of it as insurance in Escape from Tarkov. It does cost you, but it also protects you from bleeding out after one single misjudgment. Even when the stock market is going sky high, there are still fix-term deposits with single-digit interest rates because they have always been solid hedges. Less risk does not only mean less profit but also less stress. Don't get greedy. No one takes it all.