Gamma Swap: A New Type of Money Lego
TLDR; GammaSwap looks great! IL is a myth! Long Live IL!
One of the biggest “myths” that’s been foisted on retail traders is the idea of “impermenant loss.” This is a poorly understood concept that gets boiled down to traders complaining about receiving less quantity of tokens after providing them to a Constant Function Market Maker (CFMM) like Uniswap or Curve.
Without getting extremely technical, providing LP is analogous to selling straddles (uniswap v2) and strangles (uniswap v3). Upon depositing assets into the CFMM, the trader now becomes an options seller. They are “short gamma” (somewhat), in that they suffer “impermanent loss” for perpetually receiving income from trading fees. If prices for both assets supplied to the CFMM stay the same, the trader will get the exact amount deposited back, plus a little more from the fees. If the prices of both assets change a lot, well the trader is going to get back a different amount of each asset. For retail traders, the hang up goes something like “Hey I supplied 20 ETH and $25,0000 dollars a few months ago…. Now when the price has gone up, I’m getting back 15 ETH and $35,000. If I had just held the ETH and dollars, I’d have more money right now.”
The problem here though is that “option traders” don’t think in terms of unit cost. They only think in cash balance terms. In the above example, the trader started with a cash balance of $50,000 (25 ETH + $25,000). When they withdrew their cash balance was $70,000. Had they held they would have had substantially more, but ultimately the trade they took was a winning one. They profited at the end of the day and their cash balance ended up higher than what they started with. If prices moved downwards, then the trader would end up with MORE ETH and less dollars. Thus, the goal for retail LPs at the end of the day should be simple; end with more money than you started with.
There’s a widespread mental problem that affects many traders. They think up and down, but eschew time completely. Every trade turns into “THIS IS GOING TO DO 1000% TOMORROW GET READY [ROCKET SHIP EMOJI x10].” And sure, that might be the case in 1 in every million trades, but in reality price goes up or down slightly, then none, then a lot, then crashes, recovers, etc. Straight up and down is rare.
For the sake of simplicity, there are roughly 5 different outcomes that can happen with price in a given period:
Every trade is a choice of what expected outcomes are going to be. If you think 1 or 5, you want to be max bid or short and happily not collect any fees. If 2, then LP or holding might be a good option, the call to make will be “will my cash balance be higher from LP fees over X period of time?” if not then hold. 3 is always best for LP since you earn fees over the time period. 4 is only good for LPs when the fees outweigh the price decline. This rudimentary model for how to allocate assets assumes the goal at the end of the period is to maximize cash balances.
Most retail traders won’t go further than this model. The next level is to introduce hedging on one or both assets. At this point IL really goes out the window as the strategies assume cash balance to begin with. Don’t fear the impermanent loss, fear capital losses.
Ok, now that we’ve covered the IL myth for retail traders let’s get into Gamma Swap. Since LPs are options sellers, they are “short gamma,” which means they “lose” money the further prices get away from their starting point. The chart below created by the Gamma Swap team shows how as prices move in either direction, LPs gain more of one asset and lose the other.
Gamma Swap creates a market for traders to bet on either LP + fees or holding assets outright.
Traders deposit their CFMM LP tokens into the Gamma Swap contract. LPs are always short gamma. They will continue to earn fees while their LP is deposited to Gamma Swap.
Once deposited, borrowers can come and go “long gamma” by posting collateral and borrowing the reserve assets used in the CFMM LP. Borrowers are betting on either option 1 or 5 above as the outcome. If volatility is high and prices move a lot in one direction, borrowers profit as the “impermanent loss” becomes an “impermanent gain” for them.
Let’s take a more detailed look.
Assume ETH is $1000
If ETH moves to $2000
Difference between LP and GS Borrower = $150,000 - 141,421.36 - Fees?? = 8,578.64
So in this case the GS Borrower would potentially profit $8.5k USD minus whatever fees are collected by the LP over the period. If the fees are less than the impermanent gain difference, then they profit. This example can be reversed for prices going down. It works either way.
Gamma Swap keeps track of the assets that are borrowed and the yield accrued to the LPs during that time. Interest accrues and must be paid back when the position is closed. If the interest exceeds the borrowers collateral, its liquidated and used to repurchase assets for the LPs.
Liquidations happen without any oracles. One of the common attacks on lending protocols is flash loans. In a single block the price of an asset is moved up or down a significant amount, allowing the attacker to manipulate lending protocols with the goal of extracting unearned value and leaving bad debt. Because of the risk curve on Gamma Swap, it makes no difference whether the price goes up or down significantly, as the only outcome that matters is the number of LP tokens delivered back to the protocol. The Gamma Swap team thinks their model can be applied to all token pairs safely and securely, creating new markets for long gamma traders.
What’s so cool about Gamma Swap is that they enable long gamma positions without LPs having to do anything except deposit into their smart contract. LPs always explicitly will be short gamma, Gamma Swap is simply enabling a new type of trade to be derived from the underlying asset. It’s a very smart implementation and will unlock new ways for sophisticated traders to profit in different market conditions.
Ultimately, retail LPs can keep their same simple strategies and deposit their funds into GS. Pro traders can then use their reserve assets to trade volatility and the LPs earn extra interest on top of the fees they already collect. It’s a very simple, but effective asset usage strategy. We’ll be watching intently to see how it performs and if there is actual demand for long gamma once it launches.
How could Frax fit in?
What’s super cool about Gamma Swaps model is that if Frax/USDC LP is added as collateral, the long gamma position taken by the trader now becomes peg insurance. With stable/stable LPs the volatility should very, very, low all the time. If the peg were to break and price moved significantly away from $1, it would represent a huge move in gamma. Thus a Frax LP who wants peg insurance can go long the Frax/USDC LP in Gamma Swap for the time they think the peg is under stress.
Leverage through Fraxlend - To qualify this, I don’t know how the Frax team would integrate the GS token into an oracle or TWAP function to get a good price feed. I don’t know if the GS tokens are composable. But even so, leverage on long gamma positions will be really desirable. For the bluechip crypto pairs traders will probably earn 2-5% on their trades. With leverage using Frax as the loan instrument, those returns could be juiced 10-20x higher.
In theory locked LP could be added to GS and pose no risk to the protocol so long as the borrower collateral is golden. LPs would earn increased fees and the protocol isn’t at risk (Other than smart contract risk). While this might introduce too much risk for the core team to implement, maybe the GS team, or another team like Pitch would integrate locked LP and GS somehow. Again, all theory.