I’ve referenced impermanent loss throughout earlier articles in this series as the risk that catches more beginners off guard than almost anything else in DeFi, and I want to finally give it the full, careful treatment it deserves — including the actual math, because understanding the shape of the problem matters more than just being told “it’s risky.”
This article directly extends what is Uniswap and liquidity pools explained — both are worth reading first if you haven’t, since impermanent loss is a direct mathematical consequence of how automated market makers function, not a separate, bolted-on risk.
What Impermanent Loss Actually Is
Impermanent loss is the difference in value between holding two assets in a liquidity pool versus simply holding those same two assets separately in your wallet, without ever depositing them anywhere. It occurs whenever the price ratio between the two pooled assets changes after you deposit — which, given how often crypto prices move, is essentially always to some degree.
The “impermanent” part of the name refers to a specific, genuinely important nuance: the loss exists on paper as long as you remain in the pool and prices have diverged, but it only becomes a realized, permanent loss the moment you actually withdraw. If prices happen to return to the exact ratio they were at when you deposited, the impermanent loss fully disappears — hence the name. In practice, prices rarely return to the exact same ratio, which is exactly why this “impermanent” loss so often becomes very permanent in practice.
Why It Happens: The Mechanics
As covered in liquidity pools explained, most AMM pools maintain a constant product between their two reserves (the formula x × y = k). When the market price of one asset changes, arbitrage traders buy from or sell into the pool until its internal price matches the broader market — and every one of those arbitrage trades shifts the pool’s reserve ratio away from what you originally deposited.
Here’s the part that’s genuinely counterintuitive: this rebalancing always works against you, regardless of which direction the price moves. If the volatile asset in your pool rises in price, the pool sells some of it on your behalf as arbitrageurs buy it from the pool — meaning you end up holding less of the asset that just became more valuable. If it falls in price, the pool buys more of it on your behalf — meaning you end up holding more of the asset that just became less valuable. Either direction leaves you worse off than simply holding the original amounts unchanged, and the loss grows larger the further the price moves from your original deposit ratio, in either direction.
The Actual Math, Worked Through Properly
Let’s go through this with real numbers, because the shape of the relationship matters more than any verbal description can convey.
Say you deposit into an ETH/USDC pool when ETH is priced at $2,000, contributing 1 ETH ($2,000) and $2,000 in USDC — a $4,000 total deposit, with the pool’s constant product (k) therefore equal to 1 × 2,000 = 2,000.
If ETH rises 2x, to $4,000: Using the constant-product formula, your new position would hold approximately 0.707 ETH and $2,828 in USDC (since the pool rebalances to maintain the constant k while reflecting the new price ratio). Your position’s total value: (0.707 × $4,000) + $2,828 = $2,828 + $2,828 = $5,656.
Compare that to simply holding the original 1 ETH and $2,000 USDC without depositing anywhere: (1 × $4,000) + $2,000 = $6,000.
The difference — $6,000 minus $5,656 — is your impermanent loss: $344, or roughly 5.7% of what you would have had by simply holding. This is true even though your pooled position still grew substantially in absolute terms, from $4,000 to $5,656 — a real, positive return, just a smaller one than simply holding would have produced.
If ETH instead falls 50%, to $1,000: Your position would rebalance to approximately 1.414 ETH and $1,414 USDC, for a total value of $2,828. Simply holding the original amounts would leave you with (1 × $1,000) + $2,000 = $3,000. The impermanent loss here is $172, or roughly 5.7% — notice this is the exact same percentage as the 2x upward move, which is a genuinely important and non-obvious property of this math: impermanent loss is symmetric around the original price, and it depends only on the magnitude of the price ratio change, not its direction.
A Quick Reference for the Shape of the Loss
This relationship follows a predictable pattern as the price ratio diverges further from where you deposited:
| Price Change | Approximate Impermanent Loss |
|---|---|
| 1.25x (25% move) | ~0.6% |
| 1.5x (50% move) | ~2.0% |
| 2x (100% move) | ~5.7% |
| 3x (200% move) | ~13.4% |
| 5x (400% move) | ~25.5% |
The key takeaway from this table: impermanent loss grows non-linearly as price divergence increases. A modest price move costs you very little; a dramatic one costs you substantially more, in a curve that steepens rather than growing in a simple straight line.
Why Fee Income Doesn’t Automatically Offset This
A common, reasonable-sounding assumption is that trading fees earned along the way simply cancel out impermanent loss. Sometimes they do; often they don’t, and the difference comes down to a race between two separate variables: how much the price ratio has diverged versus how much fee income the pool has actually generated during the time you were in it.
A pool with very high trading volume relative to its total liquidity can generate enough fee income to comfortably outweigh a moderate amount of impermanent loss — this is genuinely the normal, profitable case for many established, high-volume pools, and it’s why liquidity provision remains a viable strategy despite this risk. A pool with low trading volume relative to a large price move, on the other hand, can leave liquidity providers with a net loss even after accounting for whatever fees did accumulate. There’s no universal answer here — it depends entirely on the specific pool’s volume-to-liquidity ratio and the magnitude of price movement during your specific holding period, which is exactly why a displayed APY in isolation, without context on the underlying asset pair’s volatility, doesn’t tell you enough to make a genuinely informed decision.
Which Pools Are Actually Most at Risk
Highest impermanent loss risk: Pools pairing two assets with no fundamental reason to move together — an established Layer-1 token paired with an unrelated smaller-cap altcoin, for instance, where the two prices can easily diverge substantially and independently of each other.
Moderate risk: Pairs like ETH/USDC or BTC/USDC — one volatile asset against a stablecoin. The volatile asset can move significantly, and since the stablecoin side doesn’t move at all, the full magnitude of any price change in the volatile asset translates directly into impermanent loss exposure.
Lowest risk: Stablecoin-to-stablecoin pairs (USDC/USDT) or pairs of closely correlated assets that tend to move together, like ETH and a liquid staking token such as stETH. Since both assets in the pool track a similar underlying value, the price ratio between them rarely diverges much, keeping impermanent loss minimal under normal conditions — though not zero, particularly if one side experiences a depeg event.
Ways to Reduce, Though Not Eliminate, This Risk
Choose correlated or stable asset pairs deliberately. This is the single most effective lever — if impermanent loss concerns you, the asset pairing you choose matters more than any other single decision.
Favor higher-volume pools relative to their total liquidity. More trading activity relative to pool size means more fee income generated to offset whatever divergence does occur.
Use concentrated liquidity thoughtfully, if you’re willing to manage it actively. As covered in liquidity pools explained, concentrating liquidity within a realistic, well-considered price range can meaningfully increase your fee income per dollar deposited — though it requires genuine ongoing attention rather than a deposit-and-forget approach, and choosing too narrow a range increases the chance your position moves entirely outside it and stops earning anything.
Some newer protocols offer explicit impermanent loss protection mechanisms. These exist, but they typically involve their own tradeoffs — often a longer required holding period to qualify for full protection, or funding the protection mechanism through the protocol’s own token, reintroducing exactly the kind of token-price-dependent risk covered in what is yield farming.
Final Thoughts
Impermanent loss isn’t a bug, a rare edge case, or a sign that something has gone wrong with a specific pool — it’s a direct, predictable mathematical consequence of how automated market makers maintain pricing, and it applies to every constant-product pool to some degree, scaling with how much the underlying asset pair’s price ratio moves apart. Understanding the actual shape of this relationship — symmetric, non-linear, and dependent entirely on price divergence magnitude rather than direction — lets you evaluate any specific liquidity pool honestly: which asset pairing you’re choosing, how much trading volume is realistically likely to offset the divergence you might experience, and whether the displayed APY genuinely compensates for the risk you’re actually taking on.
This article is for educational and informational purposes only and does not constitute financial advice. Liquidity provision carries risk of loss, including impermanent loss, and past pool performance does not guarantee future results. Always do your own research before providing liquidity to any pool.