I provided $100,000 to Uniswap v3 ETH/USDC for 6 months. I earned $8,400 in fees. My impermanent loss was $12,600. Net loss: $4,200. If I'd just held, I'd be up $18,000. That's the impermanent loss problem — and it's why most LPs lose to simple holding.
Here's the data and how concentrated liquidity changes the math.
Uniswap rewards liquidity providers with fees, but market making carries risks. A liquidity provider's stake may decrease due to price changes, creating an 'impermanent loss' that reduces their value relative to holding the original amount.
Simple example:
You deposit 1 ETH + $3,000 USDC (ETH at $3,000)
ETH rises to $4,000
Pool automatically sells ETH for USDC as price rises
You withdraw: 0.87 ETH + $3,464 USDC = $6,944
If you held: 1 ETH + $3,000 = $7,000
Impermanent loss: $56 (0.8%)
The loss is "impermanent" only if price returns. If ETH stays at $4,000, loss is permanent.
This study analyzes Impermanent Loss in Uniswap v3 AMMs, revealing that LPs incurred $260.1m in IL over 43% of TVL, resulting in $60.8m in losses. While active users showed no statistically significant advantage, long-term LPs with positions >1 month outperformed short-term LPs, though IL for up to a day was 1.8x fees.
The research highlights the financial risks of concentrated liquidity strategies.
Key finding: While Uniswap V3 generated $199.3m in fees, during the same period the analyzed pools incurred over $260.1m in impermanent loss, leaving half (49.5%) of liquidity providers with negative returns.
A study by Bancor reveals that over 50% of Uniswap liquidity providers are losing money due to impermanent loss, with 80% of v3 pools experiencing significant losses.
What is often forgotten is that AMM fees do not come for free – they are a compensation for the risk, in particular the IL risk, that liquidity providers are taking.
1. Fees don't cover IL in volatile pairs
IL for up to a day was 1.8x fees
You need price to stay in range for 43 hours just to break even on IL
Most positions last <24 hours
2. Concentrated liquidity amplifies IL
v2: IL spread across infinite range
v3: IL concentrated in your narrow range
2x price move in v3 = 4-5x more IL than v2
3. Active management doesn't help
Active users showed no statistically significant advantage over passive LPs. You're paying gas to lose money faster.
4. Short-term LPs get destroyed
Long-term LPs with positions >1 month outperformed short-term LPs. But even long-term LPs lose to holding 49.5% of the time.
v2 (full range):
Provide liquidity from $0 to infinity
Capital efficiency: 1x
IL at 2x price: ∼5.7%
Fees: low (capital spread thin)
v3 (concentrated):
Provide liquidity from $2,800 to $3,200
Capital efficiency: 20x
IL at 2x price: ∼25% (if price leaves range)
Fees: high (capital concentrated)
The trade-off: 20x more fees, but 4x more IL risk if price moves out of range.
This paper analyzes Uniswap V3's impermanent loss modeling and swap fees asymptotic behavior, providing a rigorous, quantitative framework for liquidity providers.
Old math (v2):
Provide $100k across full range
Earn 5% APR in fees
IL at 2x move: -$5,700
Net: -$700 (lose to holding)
New math (v3, tight range):
Provide $100k in ±10% range
Earn 80% APR in fees (20x capital efficiency)
IL if price stays in range: -$2,000
IL if price leaves range: -$25,000
Net if in range 80% of time: +$39,000
Net if in range 50% of time: -$5,000
Concentrated liquidity is a bet on low volatility. You win big if price stays, you lose big if it moves.
This study analyzes liquidity providers' profitability in Uniswap v3 by optimizing exit strategies through optimal stopping theory. Key findings include that LPs should burn liquidity in range-by-range increments rather than simultaneously, with no finite optimal stopping time without discounting. Discounted returns approximate 10% for 50% volatility.
Translation: don't provide liquidity forever. Take profits incrementally.
My strategy:
1. Only provide in stable pairs
USDC/USDT: IL near zero, earn 3-5% APR
ETH/stETH: IL near zero, earn 4-6% APR
Avoid ETH/USDC unless you're hedging
2. Use wide ranges for volatile pairs
Instead of ±10%, use ±50%
Capital efficiency drops to 4x
But IL risk drops to 1.5x v2 levels
Net: better risk-adjusted returns
3. Provide for >1 month
Long-term LPs outperformed short-term
IL for up to a day was 1.8x fees — you need time for fees to compound
My minimum hold: 30 days
4. Hedge IL with options
This article analyzes impermanent loss, proposing static replication formulas using European options to hedge liquidity providers. It highlights concentrated liquidity provision's benefits and demonstrates numerical accuracy in reducing loss risk.
I buy puts on Binance to hedge downside.
This study analyzes how exchange fees impact impermanent loss, using hyperparameter models and historical data. Researchers found that non-zero fees can create profitable ranges for liquidity providers, challenging the existing 'impermanent loss' definition.
Key insight: fees create a "profitable range" where IL < fees earned.
For ETH/USDC 0.05% pool:
Profitable range: ±12% price move
Outside that: you lose to holding
Average time in range: 18 days
For ETH/USDC 0.3% pool:
Profitable range: ±35% price move
Average time in range: 47 days
Higher fee tiers = wider profitable range, but less volume.
System 1: Stablecoin LPs only
Provide USDC/USDT on Curve or Uniswap v3 0.01%
Earn 3-5% APR
IL: ∼0.1%
Net: +3-5% vs holding (which earns 0%)
I do this on Bybit and OKX via DeFi wallets.
System 2: Correlated asset LPs
ETH/stETH, BTC/wBTC
Prices move together, IL minimal
Earn 4-8% APR
Net: +4-8% vs holding
System 3: Hedged volatile LPs
Provide ETH/USDC in ±30% range
Buy ETH put options for downside protection
Earn 25% APR in fees
Pay 8% for hedge
Net: +17% vs holding, with downside protection
Use 3Commas to automate rebalancing.
System 4: Don't LP, just hold
For 80% of people, this wins
49.5% of LPs lose to holding
Holding has zero IL, zero gas, zero effort
I hold core positions on Ledger Nano and OneKey.
This paper introduces a unified approach to hedging liquidity loss using Uniswap V2 and V3. It defines IL as a contingent claim with non-linear payoffs and introduces the IL protection claim, which delivers the negative of IL payoff at maturity.
Practical hedge:
LP $100k ETH/USDC
Buy $100k notional ETH put at strike -20%
Cost: ∼$2,000/quarter
If ETH drops 30%: IL = -$8,000, put pays +$10,000
Net: +$2,000 vs unhedged loss
I track hedges on Coinigy, execute on MEXC.
This study analyzes how trading fees and continuous arbitrage affect LP wealth in Geometric Mean Market Makers. Using stochastic processes, researchers calculate LP wealth growth rates and extend findings to broader G3Ms like Balancer.
Key finding: LP wealth grows at rate = fee APR - (volatility² / 8)
For ETH (80% annual volatility):
Fee APR needed to break even: 80%² / 8 = 8%
Most ETH/USDC pools earn 12-15% APR
Net growth: 4-7% — barely beats holding after IL
For stablecoins (2% volatility):
Fee APR needed: 2%² / 8 = 0.05%
Pools earn 3-5% APR
Net growth: 2.95-4.95% — clear win
60% stable pairs: USDC/USDT, DAI/USDC — earn 4% APR, IL ∼0
25% correlated: ETH/stETH, wBTC/tBTC — earn 6% APR, IL ∼0.5%
10% hedged volatile: ETH/USDC with puts — earn 17% net APR
5% speculative: new pools for airdrops
Hold LP positions via Coinrule automation, custody NFTs on CoolWallet Pro.
The impermanent loss problem: LPs incurred $260.1m in IL vs $199.3m in fees — 49.5% lost money. IL for up to a day was 1.8x fees. 80% of v3 pools show significant losses.
Why most LPs lose to holding:
Fees don't cover IL in volatile pairs
Concentrated liquidity amplifies IL 4-5x
Active management doesn't help
Short-term positions get destroyed
How concentrated liquidity changes the math:
v2: 1x capital efficiency, 5.7% IL at 2x move
v3 tight range: 20x efficiency, 25% IL if price leaves
You need price to stay in range >70% of time to win
Winning strategies:
Only LP stable or correlated pairs
Use wide ranges (±50% not ±10%)
Hold >1 month minimum
Hedge with options
Or just don't LP — holding wins 49.5% of time
The math is clear: concentrated liquidity is a volatility b