DeFi Yield Farming Guide Part II: Outpacing Impermanent Loss

This article first appeared on (reposting with permission). 

Billionaire Mark Cuban has one piece of wisdom for the new crypto-rich: “Learn to hedge.” This article discusses an incredible tool in the hedging toolkit, the Liquidity Pool.

Note: This is not financial advice. It’s a math discussion and some strategic analysis around the Liquidity Pools returns found specifically in Uniswap, Sushiswap, and Pancakeswap.

Automated Market Makers

Automated Market Makers are a powerful new technology in cryptocurrency. They transformed trading on the blockchain since their debut in 2018. Uniswap, on the Ethereum blockchain, is almost always the largest user of the chain’s capacity. Uniswap changed trading from a competitive winner-take-all game into a cooperative endeavor between thousands of anonymous participants. Old decentralized exchanges became obsolete. The core concept of the automated market maker is the Liquidity Pool (LP), in which Liquidity Providers add tokens. These are cooperative pools- when traders trade with the pool, fees are split equally among participants in proportion to their contributed capital.

Liquidity Pools offer Liquidity Providers revenue in the form of trading fees and rewards. As soon as they were invented, people began discussing ways to improve them. A very large focus was placed on Impermanent Loss (IL). Impermanent loss, also known as divergence loss, is the unavoidable loss of value of your LP investment from any change in asset prices. Without revenue, LPs are terrible investments which lose versus stablecoins on the downside and lose against holding the volatile asset on the upside. In fact, when Vitalik Buterin proposed the AMM idea, he viewed it as an act of public service!

The discussion around Liquidity Providers (LPs) in Automated Market Makers (AMMs) needs to move forward from endless arguments about impermanent loss. This analysis shows how revenue changes the game significantly. With enough revenue, LPs become a hedge asset, giving you a way to out-earn changes in asset values.

As the bull market advances, drops become more and more likely. Because of this, certain LPs will become more attractive as returns from pure token appreciation diminish. After using Uniswap, Bancor, and Pancakeswap through highly volatile market conditions in the past few months I’ve had some thoughts as to how to reframe the understanding of LPs. When people complain about impermanent loss, they’re really complaining that their LP does not make enough revenue.


Basic LP strategy is simple: only invest in the biggest AMM on your blockchain (by Total Value Locked, TVL), and only invest in the long-term best earning LPs on that exchange. The risk of people running off with your money is far too great otherwise. Embrace large pools- the large pools will attract more traders, increasing fees faster than the dilution of having more people invested.

Only invest in the biggest AMM

The only viable protection on the blockchain is size and network effect. Do NOT invest in tiny AMM’s. They will never generate the revenue to cover your investment. Unknown coins on a reputable exchange are also vulnerable to exit scams known as “rug pulls” where all the valuable coins are drained out of the LP. It's not smart to buy totally unknown coins in pursuit of high returns either.

Now for advanced strategy. This article discusses the classic 50/50 LP, which are made of two assets in equal dollar value amounts. For example, $1000 worth of BTC and $1000 worth of USDC.

LPs were born in an environment of relatively lower volatility and depressed asset prices. The early conversation focused significantly on Impermanent Loss (IL). Impermanent loss is partly the idea that you underperform versus going all-in on a rapidly appreciating (or rapidly falling) asset.

Impermanent Losses

Impermanent loss: lose money from any price drifts away from your entry

I believe this focus needs to be significantly updated given new market conditions and new LP options. Right now, at a high point in the bull market, a lot more focus needs to be paid to optimal strategy selection and avoiding losses outright.

With that in mind, let’s introduce the Revenue Factor “r” to represent (LP revenue/initial investment + 100%)

Define revenue factor r

Revenue factor r

This revenue can be derived from any source- trading fees, subsidies, reward tokens. This revenue must be reinvested into the pool for compounding and not removed as payout.

Let's define p1 is the price of the underlying asset when you buy into the LP. Define p2 to be a price sometime in the future.

Revenue is the key factor for an LP: with revenue, can you out-earn volatility? A LP can beat impermanent loss up to:

LP wins up to r squared

A LP can even tolerate significant drops in asset price, down to:

LP wins down to 1 over r squared

That leads us to the following chart:

Optimal Strategy Graph

Revenue is a very important factor, and the market with the most persistent trading fee stream is Uniswap. Uniswap *real* returns are somewhere between 10%-30% APY for the biggest pools like ETH/USDC. That means Uniswap pools with a one year investment horizon can tolerate between a 69% price rise or -41% price drop and still be the best investment relative to all-in token or all-in stablecoins. In crypto, that isn’t much buffer. Doubling is common in token prices, as are 90% crashes. However, the winner-take-all nature of AMM’s means Uniswap has a decisive edge on the Ethereum chain. The biggest liquidity directly leads to the biggest trading volumes and revenue. The lower returns make it a somewhat underperforming investment, unless you are holding for multiple years.

There are other important sources of income like token reward streams. On Binance Smart Chain, PancakeSwap reward pools are much stronger when coupled with an autocompounding contract like A normal rewards-based PancakeSwap pool yielding 50% APY can tolerate a 125% price rise or -56% price drop and come out ahead of other strategies. That is much better hedging! Pancakeswap’s gamified, high-reward environment offers REAL protection against impermanent loss. The low fee environment on Binance Smart Chain means autocompounding is significantly more powerful and uses far less gas than on Ethereum. Ultimately, that means better returns.

A normal rewards-based PancakeSwap pool yielding 50% APY can tolerate a 125% price rise or -56% price drop and still come out ahead of other strategies.

Optimal Strategy Graph

Other AMMs and yield farms are experimenting with subsidizing reward tokens from additional revenue streams to promote the underlying reward token or providing insurance on various aspects of pooling. If a stable solution is found it could significantly increase the innovative AMM’s TVL and give it stickiness in the market.

LP Revenue

Beats All-In for price gains up to:

Beats Stablecoins for price drops down to:








































The second point on IL I’d like to reframe is what assets are used to initiate the LP. If you initiate the LP using the volatile asset, you are considering returns versus pure holding. When you consider LP performance versus cash or stablecoin, impermanent loss doesn’t really exist. It’s either real gain or real loss. IL is only a problem if you start your position from a full or half stack of volatile assets. A simple solution is to *always* start your LP positions from cash, and use an intentional strategy selection based on the availability of revenue in LPs. That is, don’t start by selling half of your stack. Start from cash or stablecoin.

For advanced strategy, we have clear indications when not to use LP’s. First, if you expect a clear 100% or more gain from the underlying asset just buy the asset and don’t try to earn more via LP. That means for unknown coins, you’re almost always better off just going YOLO into a full stack. For big coins like BTC and ETH, only enter the LP when you’re worried about future drops below the current level.

If you expect a loss of over 60% from the underlying asset it will be incredibly hard to out-earn the loss. In the case of high downside, stablecoin-stablecoin LP pools are always available and would be the best available alternative strategy. Stablepools are going to be an important investment asset going forward - they serve an important role enabling smooth commerce in a multi-chain future. They’re a good idea if you expect they will maintain peg. Just hunt for the best return!

Finally, if you know your expected return, you can clearly calculate your LP’s volatility tolerance: 1/r^2 on the downside and r^2 on the upside. From this we know volatile LP’s are amazing if it earns a high APY, you’re in it for a long time, and you expect we’re somewhere in the midpoint of the market and future values. Since LPs are capital-gains taxed it can be incredibly attractive to park it in the market and just let it go til long term gains. The longer you run, the more insulated it is from any market events. For that you need to choose a blue-chip AMM that won’t ever die - generally speaking the #1 on its blockchain.


  • Investors are over-invested in tiny altcoin LPs. The returns don't make sense and it would just be better to be all-in on the coins.
  • Investors are under-invested in stablepools
  • Investors should start their positions from cash or stablecoin and not sell their assets to fund a LP
  • Investors are under-invested in high return big-cap AMMs on the newer blockchains
  • LPs provide significant and well-defined protection on both upside and downside when they are really earning revenue for a long period of time

    If you like this article please follow the author on twitter @bscgateway to get notices on future articles!


    • There are no comments yet. Be the first one to post a comment on this article!

    Leave a comment

    Please note, comments must be approved before they are published