LP Math: When Yield Beats Impermanent Loss
The previous two lessons gave you the two halves of the liquidity-provision equation: fee income (potential reward) and impermanent loss (potential cost). This lesson puts them together into a single decision framework, and introduces the single most important variable for predicting which side wins: correlation between the two assets in the pool.
When the two tokens in a pool are highly correlated — most obviously, two stablecoins that are both designed to stay near $1, like USDC and USDT — their relative price barely moves, which means the price ratio driving the impermanent loss formula stays close to 1, and IL stays close to zero almost regardless of what happens to the broader crypto market. In this case, nearly all of your fee income is close to "free" yield, since there's very little rebalancing cost eating into it. When the two tokens are uncorrelated or only loosely correlated — for example, ETH paired with a small, independently-priced token — their relative price can swing dramatically, and the impermanent loss formula from the previous lesson shows that even moderate relative price moves (2x, 3x, 4x) produce IL percentages that can easily exceed double-digit trading-fee APRs, especially for pools that haven't been running long enough to accumulate meaningful fee history.
The practical rule that follows: stablecoin/stablecoin pairs are generally the lowest-risk way to earn liquidity-provision yield, because IL risk is structurally minimized by design; volatile/volatile pairs can offer eye-catching advertised yields, but require you to seriously model realistic price-move scenarios (as you did in the previous lesson) before trusting that the advertised number reflects what you'll actually keep.
Worked Example
Consider a USDC/USDT pool advertising a modest 4% APR in trading fees. Because both tokens are stablecoins, even a relatively large de-pegging event (one of them briefly trading at $0.98 instead of $1.00, for instance) produces only a small impermanent loss, since the price ratio barely moves from 1. In almost every realistic scenario, LPs in this pool keep close to the full 4% APR, because there's essentially no IL to subtract.
Now compare this to an ETH paired with a newly-launched governance token, advertising a flashy 150% APR — likely boosted heavily by temporary incentive rewards, a topic the next lesson explores. If that new token's price swings 3x relative to ETH over the following month — entirely plausible for a low-liquidity, newly-launched asset — the impermanent loss formula from the previous lesson shows an IL in the range of 13-14% for that single move alone. Even a very generous 150% annualized fee rate, prorated to one month (roughly 12.5%), would be entirely wiped out by a single 3x price swing, leaving the LP with a net loss despite the "high yield" headline number that attracted them in the first place.
There's a useful middle category worth naming explicitly: correlated-but-not-identical pairs, such as ETH paired with a liquid-staking derivative of ETH, or two different wrapped versions of the same underlying asset. These pairs sit between the two extremes — their relative price can drift modestly (for example, if a liquid-staking token temporarily trades at a small discount or premium to the ETH it represents), producing some IL, but nowhere near the magnitude seen in genuinely uncorrelated pairs. These "correlated pair" pools have become a popular middle ground for LPs seeking meaningfully higher yield than pure stablecoin pairs without taking on the full IL risk of unrelated volatile assets, and they're worth knowing about as a specific category when you evaluate real pools.
Common Pitfalls
- Treating "stablecoin pair" as automatically zero-risk. A stablecoin can de-peg significantly during a crisis (several historical examples exist), and a severe de-peg produces real IL even in a "stable" pool — correlation reduces risk, it doesn't eliminate it.
- Comparing advertised APR across pools without adjusting for each pool's realistic IL profile. A 20% APR stablecoin pool and a 20% APR volatile pool carry very different realistic net returns.
- Ignoring how long an incentive-boosted APR (from the next lesson) is scheduled to last relative to your own time horizon for holding the position.
Exercise
Take two candidate pools you've researched or can reasonably estimate: one stablecoin/stablecoin pair and one volatile/volatile pair. For each, write down the currently advertised fee APR (or your own estimate from the "How Liquidity Pools Work" lesson), and a rough worst-case impermanent loss estimate for a plausible price move over the next month — using the IL formula or calculator from the previous lesson, and a price-move assumption you consider realistic for that specific pair given its historical volatility.
Compare the two numbers for each pool and decide, in each case, whether the fee yield clearly, marginally, or does not at all clear your personal bar for "worth the IL risk." Write one sentence of reasoning for each pool. This comparison — fee yield versus realistic worst-case IL — is a template you should reuse for any pool you ever consider in real life, not just in this exercise.
Recap
- Correlation between pool assets is the single biggest driver of impermanent loss risk: stablecoin pairs carry minimal IL; volatile pairs can carry severe IL.
- Flashy advertised APRs on volatile pairs can be entirely wiped out by a single realistic price swing, as the worked example's 3x move shows.
- Always compare realistic worst-case IL against fee yield before providing liquidity — this comparison template applies to any pool you'll ever evaluate.