DeFi Research

13 Jul 2026

Impermanent Loss Explained: A Guide for Uniswap LPs

Impermanent loss explained in simple terms. Learn what causes it, how to calculate it, and see strategies to mitigate it, including new tools for Uniswap v4.

You're probably in a familiar spot. You deposited into a Uniswap pool, watched fees come in, and still ended up staring at a portfolio value that looked worse than just holding the tokens in your wallet.

That disconnect often prompts searches for impermanent loss explained. And it's why so many LPs get blindsided. The fee counter goes up, but your relative position can still go down.

The hard lesson is that LPing isn't just about earning swap fees. It's about taking on a specific kind of inventory risk. In full-range pools that risk is already real. In concentrated liquidity, it becomes an ongoing operational problem. You're no longer just a passive capital provider. You're managing exposure, timing, and execution quality.

Table of Contents

Why Your LP Position Value Can Drop Even When Earning Fees

The first mistake most new LPs make is thinking fees and portfolio performance are the same thing. They aren't.

You can earn fees every day and still underperform a simple wallet holding of the same two assets. That's because the pool keeps changing what you own. When one token starts running, the AMM sells some of your winner and gives you more of the lagging asset. You're collecting fees, but you're also giving up upside relative to HODLing.

Why the name trips people up

The phrase impermanent loss sounds gentler than the experience feels. It suggests a paper fluctuation that may disappear on its own.

That's only true if prices return to the original ratio before you withdraw. A guide from Speedrun Ethereum makes the key point plainly: the term is misleading because the loss only reverses if prices return to the starting ratio, and in volatile markets it often becomes a realized permanent loss when you exit at a new ratio. It also notes that many LPs don't realize they're effectively betting on mean reversion, not just collecting fees through time (Speedrun Ethereum on why impermanent loss often becomes permanent).

Hard-won lesson: If you measure LPing only by fees earned, you'll miss the actual question. Did the position beat doing nothing?

That's why experienced LPs compare every position to a HODL baseline. If the same starting assets would be worth more outside the pool, the fees didn't save the trade. They just softened the underperformance.

The real trade-off

Impermanent loss isn't a bug in the interface. It's part of the deal you make with an AMM.

You hand the pool two assets. In exchange, the pool lets traders rebalance against you. You get fee income for taking that other side. Sometimes that bargain works. Sometimes it doesn't. But once you see IL as a structural trade-off instead of a weird edge case, your whole approach to LPing changes.

What Is Impermanent Loss and How Is It Calculated

The cleanest way to understand IL is to stop thinking about the pool as a passive vault and start thinking about it as a self-balancing scale.

You place two assets on the scale. The AMM keeps them in a fixed relationship. When outside prices move, traders push the pool back toward the market price. That process changes how much of each asset you own inside the pool.

Why the pool keeps changing your asset mix

In a standard constant-product AMM, the pool follows x * y = k. You don't need to live in the formula to understand the effect. If one asset gets more valuable in the external market, arbitrage traders buy that asset from the pool until the pool price catches up.

That means the pool ends up holding less of the asset that went up and more of the asset that didn't. As an LP, your share follows that same shift. So even if your total position value rises in absolute terms, you can still end up with less than you would have had by holding the original tokens.

A simple way to read the formula

The mathematical definition is:

IL = 2 * sqrt(price_ratio) / (1 + price_ratio) - 1

Ledger's glossary explains two parts of this well. First, IL is a function of price divergence, not direction. Second, a 2x price divergence results in a 5.7% loss relative to holding, and that loss only becomes realized if you withdraw while the ratio is still different (Ledger's explanation of the IL formula and 2x example).

Here's the important intuition:

  • If price barely moves, IL is small.
  • If price diverges more, IL grows.
  • If price later returns to the starting ratio before exit, the gap can disappear.
  • If you exit while the ratio is still changed, that underperformance gets locked in.

The pool doesn't care whether the move was up or down. It only cares that the ratio changed.

A quick comparison table

A simple way to think about the formula is to compare the path of a hold portfolio with the path of an LP position.

Situation HODL outcome LP outcome
Price ratio stays near entry You still own your original mix The pool mix stays close, so IL stays limited
One asset rises sharply You fully keep that upside in the wallet The pool sells part of the winner as traders rebalance
One asset falls sharply You absorb the loss directly The pool buys more of the falling asset during rebalancing
You withdraw at a changed ratio Your wallet value is whatever the market says Your LP value is lower relative to HODL because of the changed asset mix

Milk Road gives two useful reference points for constant-product pools: a 50% price increase corresponds to about 5.6% IL, and a 100% increase corresponds to 14.8% loss relative to holding in that framing (Milk Road's constant-product IL examples). You don't need to memorize those values. You just need the shape of the curve in your head. Bigger divergence means bigger drag.

That's the heart of impermanent loss explained in plain English: the AMM keeps rebalancing your inventory, and that inventory path often lags a simple hold strategy when prices move apart.

The Race Between Trading Fees and Price Divergence

Once you understand IL, the next question is practical. If LPing creates this structural drag, why do people do it at all?

Because there are two forces in the position. Price divergence pulls you backward. Trading fees push you forward. Your result depends on which one wins over the life of the trade.

Fees are your compensation for taking inventory risk

Every swap that touches your liquidity pays you for making a market. That fee stream is the reason LPing can still make sense even though IL is built into the mechanism.

A January 2024 simulation study on arXiv reached a result that many LPs find surprising: providers can remain profitable relative to a simple hold strategy as long as prices do not drop by more than 75% or increase by more than 300% within a year. Within that boundary, accumulated fees often offset impermanent loss, which suggests the risk is less severe than many people assume during more standard volatility regimes (arXiv simulation on fee accumulation and IL boundaries).

Where the fee cushion tends to hold

That doesn't mean fees always rescue the trade. It means IL is conditional, not automatic.

A useful way to understand this:

  • Calm or two-sided markets: fees have more room to offset divergence.
  • Strong one-way trends: IL can outrun the fee stream.
  • Violent moves that persist: the position starts looking less like market making and more like forced inventory rotation at the wrong time.

Practical rule: Don't ask whether a pool “has impermanent loss.” Every volatile pool does. Ask whether the expected fee stream is strong enough to outrun the divergence risk you're taking.

That framing changes behavior. Instead of chasing headline APR or fee activity, you start judging pools by regime, pair quality, and how likely the move is to become one-directional.

How Concentrated Liquidity Amplifies Both Fees and Risk

Full-range LPing already asks you to accept inventory drift. Concentrated liquidity raises the stakes by making your capital more efficient inside a chosen band.

That's the attraction. More of your liquidity sits where trading happens.

Capital efficiency cuts both ways

The same design that boosts fee capture also makes the position more sensitive to movement. Binance Academy notes that concentrated liquidity changes the IL risk profile because providers choose a price range rather than offering liquidity across the full curve. Its explanation highlights that a 20% price move within a narrow 10% range can generate IL exceeding 15%, while the same move in a full-range V2 pool would be around 5.6% IL. The same source also notes that active pairs such as ETH/USDC can see 0.1% to 0.3% daily fee income, which is why narrow ranges can look so tempting in the first place (Binance Academy on concentrated liquidity, fee potential, and amplified IL).

The setup is seductive. Narrow the range, concentrate the capital, collect more fees while price stays where you expect. But that “while” is doing a lot of work.

Out of range is where the pain changes shape

In Uniswap v3 and v4 style concentrated liquidity, being wrong isn't just less efficient. It changes the nature of the position.

If price moves outside your selected band, the position becomes 100% of one asset and earns zero fees. A Substack analysis focused on Uniswap v4 explains that this amplifies IL compared with full-range LPing because you stop participating in fees while also missing subsequent appreciation or depreciation in the way a holder would (Gogol on concentrated liquidity positions going fully into one asset out of range).

That's why concentrated liquidity isn't really passive, no matter how platforms market it. It's an active exposure management problem.

If you're trying to reduce blowups, broad process discipline matters as much as entry logic. A strong checklist for risk management best practices for LP positions helps because most mistakes don't come from not knowing the theory. They come from reacting late, re-entering badly, or widening a range after the damage is already done.

A quick walkthrough makes the mechanics easier to visualize:

Manual vs Automated Strategies for Managing IL

Once concentrated liquidity enters the picture, most LPs split into two camps. One group manages positions by hand. The other group reaches for automation.

Both can work. Both can also fail in predictable ways.

What manual LP management gets right

Manual management forces you to think about context. You choose the pair, the range width, and whether the market even deserves your capital right now. That discretion matters.

Most experienced LPs use some combination of these approaches:

  • Wider ranges: You sacrifice some capital efficiency for fewer emergency interventions.
  • More correlated pairs: You reduce the chance that one asset runs away from the other.
  • Selective deployment: You stay out during unstable conditions instead of assuming every market deserves an LP position.

The problem is operational. Manual LPing asks you to monitor price, volatility, fee conditions, gas cost, and your HODL comparison at the same time. Few people do that consistently. Fewer do it well when the market is moving fast.

Good manual LPing isn't just clicking rebalance. It's deciding when not to be in the pool at all.

Where basic bots still fail

Automation helps with execution fatigue, but the first generation of LP bots often copied the wrong human behavior. They turned “price left range, recentre now” into code.

That sounds sensible until the market starts chopping. Then the bot can repeatedly move the range into noise, buy back exposure after an up move, and sell it away after a down move. You automate effort, but also automate whipsaw.

A better way to compare the two styles looks like this:

Approach Strength Typical failure mode
Manual management Human judgment about context Slow response, inconsistency, emotional decisions
Simple rebalance bot Fast and tireless execution Over-trading and repeated bad re-entries
More selective automation Rules-based discipline Still depends on strategy quality and risk design

That last row is the one that matters. Automation by itself isn't the edge. Decision quality is the edge. If the logic only reacts to price crossing a line, it's still a weak strategy. It just fails more efficiently.

Beyond Rebalancing with Intelligent Liquidity Automation

The next step past a basic rebalance bot is to treat LP management as a stateful decision problem.

That means the system isn't only asking whether price touched a boundary. It's asking whether conditions support LP exposure at all, whether the current move looks like noise or a regime shift, and whether capital should remain in LP, rotate toward the volatile asset, or move toward the stable asset.

Good automation is selective

Here, an intelligent liquidity management system differs from a simple recentering bot.

A stronger setup uses volatility-aware, rules-driven, and adaptive logic to decide when to open, close, and reposition. In practice, that can include layered confirmation such as Donchian breakout logic, ATR-based volatility contraction, breakout confirmation candles, close buffers, drawdown protection, cooldown rules, and optional momentum or trend filters.

The important difference is conceptual. The system is built to avoid bad action, not just automate more action.

Decision quality matters more than action speed

The operational side matters too. Serious LP automation isn't just strategy logic in isolation.

A practical setup includes:

  • Execution infrastructure: keeper services that handle opening, closing, and rebalancing without constant human supervision.
  • Guardrails: slippage controls, cooldowns, buffers, and emergency controls to reduce churn and poor entries.
  • Monitoring: dashboards, alerts, reporting, and runtime configuration so you can see what the strategy is doing and why.
  • Production reliability: multi-RPC failover and containerized deployment so the system behaves like operations software, not a weekend script.

That's the key shift in mindset. LPing stops being a manual craft and starts becoming a repeatable operating process.

A related debate in DeFi circles is whether predictive tooling helps. That question matters, but it's only useful if it improves decision quality rather than adding mystery. This overview of AI price prediction in crypto trading systems is relevant because good automation still needs transparent rules and measurable outcomes.

The benchmark still has to be HODL

No matter how complex the stack becomes, the scorecard can't be fee income alone.

The only honest way to evaluate an LP strategy is to track portfolio value, fees earned, gas costs, swap costs, and the divergence from a simple hold baseline. If the strategy can't explain its result against HODL, then the automation may be polished but the decision framework is incomplete.

“Did the strategy beat my passive alternative?” is the question that keeps an LP system honest.

How Uniswap V4 Hooks Enable Smarter LP Strategies

Uniswap v4 adds a powerful building block for strategy designers: hooks.

The simplest way to think about a hook is as custom logic that can run during pool actions. Instead of forcing every liquidity strategy into the same fixed behavior, hooks let developers shape how the pool responds around swaps and related events.

What hooks actually change

That matters because LP management often fails at the point where strategy meets execution. The idea may be good, but the system can't act precisely enough or cheaply enough to express it.

With v4 hooks, the strategy can do more inside the protocol's flow. According to Acheron Trading's write-up, rebalancing hooks can automatically reposition liquidity and can also be designed to capture MEV or arbitrage profits and push them back into the pool, partially offsetting impermanent loss. That kind of mechanism wasn't possible on v3 because v3 didn't have on-chain hooks (Acheron Trading on Uniswap v4 hooks and offsetting IL).

Why this matters for IL management

This opens a different design space for LP systems.

Instead of treating rebalancing as a blunt external action, a strategy can become more context-aware and more efficient in how it shifts liquidity. That doesn't remove IL. Nothing in the AMM design makes that risk disappear. But it does give builders more tools to respond intelligently.

For readers exploring how automation primitives fit into broader pool design, this piece on automatic payment pools and programmable on-chain logic is a useful companion.

The larger point is straightforward. As LP strategies become more customizable, execution quality becomes part of the strategy itself. In v4, better liquidity management won't just come from nicer dashboards. It will come from better embedded decision logic.

The Future of LPing Is Active and Data-Driven

Impermanent loss is part of the AMM bargain. It isn't a weird exception, and it isn't something serious LPs can afford to hand-wave away.

Concentrated liquidity made that reality sharper. It increased capital efficiency, but it also turned LPing into an active management discipline. Range selection, exit timing, re-entry quality, and performance measurement now matter as much as pool choice.

The LPs who last in this game usually stop thinking like yield tourists. They start thinking like operators. They compare every result against HODL. They treat fee income as compensation for risk, not free money. And they prefer systems that can filter conditions, reduce unnecessary action, and keep a consistent record of what worked.

That's where the industry is going. Not toward magical IL elimination, but toward more disciplined, data-driven management of a risk that's already baked into the structure.

If you want a more systematic way to manage concentrated liquidity on Uniswap v4, UBAMM.AI is worth a look. It's built as a non-custodial, rules-driven liquidity management system that automates opening, closing, and rebalancing while tracking performance against a HODL baseline, so you can judge the strategy by the benchmark that is most relevant.