More on LPs
With LPs, you usually have to deposit both sides of the pool. For instance, if you're putting money in a USDC/USDT pool, you have to deposit both USDC and USDT in equal amounts. Some (and more and more) pools allow single-sided staking.
Examples:
Non-single side staking. I have $10K USDC. I want to LP USDC/USDT. I trade half for USDT and deposit $5k of each.
Single side staking. I have $10K USDC. I want to LP USDC/USDT. I deposit $10K USDC in the pool and it adjusts for me.
What To Look For
The key to LPs is balancing return versus divergence and market risk. You can lose money in two ways with an LP: either because the asset you're depositing in the farm goes down, or because the two assets you're depositing diverge wildly in price.* (*Impermanent loss isn't really a loss so much as an opportunity cost compared to holding the asset, but it's something to remain aware of.)
Stay away from native project token pools, especially native token/stablecoin pools. They have the highest farming rewards rate, but the token will usually peak early and then dwindle, often faster than the rewards make up for.
What you want are highly correlated pairs that either increase in value or stay flat. The clear choice is stablecoin-stablecoin pairs, but other pairs may also fit.
There is a worthwhile counterstrategy, though, that I call
Sharecropping
Sharecropping is farming with borrowed assets. Let's look at two Solana projects: Raydium and Mango.
Mango lets you deposit USDC or USDT and borrow ETH, SOL, BTC, and RAY, among others.
Right now, you can deposit USDC at 13.81% (down from the 18+% a few hours ago - like I said, it's volatile) and borrow SOL at 0.08%, BTC at 0.12%, and ETH at 1.25%. Basically free money.
Raydium has a bunch of farms, like BTC-mSOL, ETH-mSOL, ETH-RAY, etc.
You can get these farm rewards while staying market neutral by sharecropping. Deposit your USDC on Mango, borrow ETH and SOL (convert to mSOL through Saber or Mercurial), and deposit those in the Raydium farm.
If the crypto market takes a huge sh!t, you don't care - you borrowed X ETH, you pay back X ETH. If ETH is $4k or $2k it's all the same.
With any borrowing protocol, you do need to watch your collateral ratio so you don't get liquidated. If you borrow SOL and SOL moons, you may end up undercollateralized.
Depending on how sketchy the farm is, you've lowered your exposure of a rug pull too, because you can't borrow your full collateral. However, the non-stable farms usually compensate better, because they have to price in market risk. So sharecropping means you can get the same or better return on less capital risk.
Example. Project Y has a USDC/USDT farm returning 10%, and a FIN/BOARD farm returning 30%. Lending protocol Z is averaging 5% APY on USDC deposits. You have $10k USDC. You don't want to buy FIN and BOARD, because they're shitcoins and they'll dump soon.
If you put all $10k into Y, you get $1k annual returns.
If you put all $10k into Z, you get $500 annual returns.
If you put all $10k into Z and borrow $3k each of FIN/BOARD, you get $500 from Z + $1,800 from Y = $2,300 annual returns, AND, because Z is probably safer than Y, you lower your risk of getting jacked from $10k to $6k.
Real world example you can use: Mango and Raydium. USDC deposits have averaged 14.1446% over the past 30 days. SOL, BTC, and ETH borrow costs are basically nothing - 0.07% for SOL and 0.33% for ETH over last 30 days. Raydium's BTC-mSOL farm is returning 21.28% and ETH-mSOL is returning 27.63%. BTC, SOL, and ETH are probably correlated enough to minimize divergence costs on the strategy.
Let's look at sharecropping into ETH-mSOL.
$10k USDC into ETH/mSOL (60% LTV borrowed) would be:
$10k USDC Mango returns 1,414.46 annual
-(interest costs on $3k SOL borrowed) (2.13)
-(interest costs on $3k ETH borrowed) (10.14)
Returns from Mango: 1,402.18 annual
$6k into ETH-mSOL 1,657.80 annual
Total return 3,059.98 or 30.5% on the original $10k.